Bug#1068946: png23d: Package Homepage is a broken link, here's the correct one
Package: png23d Version: 1.10-1.3 Severity: minor Dear Maintainer, The package home page is listed as http://kyllikki.github.com/png23d/ but that returns 404 Page not found error. The following looks to me like the page to use at this point. https://github.com/kyllikki/png23d -- David Fries -- System Information: Debian Release: 11.6 Architecture: amd64 (x86_64) Versions of packages png23d depends on: ii libc62.31-13+deb11u5 ii libpng16-16 1.6.37-3 png23d recommends no packages. png23d suggests no packages.
Bug#1004415: libassimp-dev: assimpTargets.cmake generates an invalid INTERFACE_INCLUDE_DIRECTORIES
Package: libassimp-dev Version: 5.0.1~ds0-2 Severity: normal Dear Maintainer, CMake configure of qt6.2 failed because libassimp-dev generated a path /usr/lib/include in INTERFACE_INCLUDE_DIRECTORIES which doesn't exit and cmake aborted. /usr/lib/x86_64-linux-gnu/cmake/assimp-5.0/assimpTargets.cmake When building the _IMPORT_PREFIX it removes one too few path elements. When the _IMPORT_PREFIX section is replaced with the version from 5.1.6_ds0-1 it does generate a valid path and Qt is able to finish the configure. It would be great if the Debian stable version of assimpTargets.cmake be released with the fix. -- System Information: Debian Release: 11.1 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 5.16.0+ (SMP w/4 CPU threads; PREEMPT) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15), LANGUAGE not set Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages libassimp-dev depends on: ii libassimp5 5.0.1~ds0-2 libassimp-dev recommends no packages. libassimp-dev suggests no packages. -- no debconf information -- debsums errors found: debsums: changed file /usr/lib/x86_64-linux-gnu/cmake/assimp-5.0/assimpTargets.cmake (from libassimp-dev:amd64 package)
Bug#912767: isc-dhcp-server: dhcp starts too early, named/session.key for ddns not available
Package: isc-dhcp-server Version: 4.3.5-3+deb9u1 Severity: normal Dear Maintainer, After a reboot I found isc-dhcp-server wasn't running because dhcpd was started before named wrote out session.key (included for ddns). I have ddns enabled with the following included in /etc/dhcp/dhcpd.conf include "/var/run/named/session.key"; I see /etc/init.d/isc-dhcp-server has $named listed in Should-Start, but apparently that's not sufficient with systemd, I expect dhcpd takes less time to start compared with the multiple domains named is serving on the system. # Should-Start: $local_fs slapd $named I added the following loop to the default file, I'm not sure the best way to properly handle this. /etc/default/isc-dhcp-server # delay watching for the session.key file to appear from named for ddns include="/var/run/named/session.key" i=0 while [ "$1" = "start" -a $i -lt 30 -a ! -e "$include" ] ; do echo "waiting on $include" i=$(($i+1)) sleep 1 done unset include i Here are the relevant parts of the system log file showing dhcpd not finding session.key, then a second later named starts. Oct 31 08:17:24 localhost dhcpd[1933]: Internet Systems Consortium DHCP Server 4.3.5 Oct 31 08:17:24 localhost dhcpd[1933]: Can't open /var/run/named/session.key: No such file or directory Oct 31 08:17:24 localhost dhcpd[1933]: exiting. Oct 31 08:17:24 localhost systemd[1]: isc-dhcp-server.service: Control process exited, code=exited status=1 Oct 31 08:17:24 localhost systemd[1]: Failed to start LSB: DHCP server. Oct 31 08:17:24 localhost systemd[1]: isc-dhcp-server.service: Unit entered failed state. Oct 31 08:17:24 localhost sh[1883]: Restarting isc-dhcp-server (via systemctl): isc-dhcp-server.serviceJob for isc-dhcp-server.service failed because the control process exited with error code. Oct 31 08:17:24 localhost sh[1883]: See "systemctl status isc-dhcp-server.service" and "journalctl -xe" for details. Oct 31 08:17:24 localhost systemd[1]: isc-dhcp-server.service: Failed with result 'exit-code'. Oct 31 08:17:25 localhost named[2048]: starting BIND 9.10.3-P4-Debian -f -u bind Oct 31 08:17:25 localhost named[2048]: loading configuration from '/etc/bind/named.conf' Oct 31 08:17:26 localhost named[2048]: running -- System Information: Debian Release: 9.5 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 4.19.0+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15), LANGUAGE=C (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages isc-dhcp-server depends on: ii debconf [debconf-2.0] 1.5.61 ii debianutils4.8.1.1 ii libc6 2.24-11+deb9u3 ii libdns-export162 1:9.10.3.dfsg.P4-12.3+deb9u4 ii libirs-export141 1:9.10.3.dfsg.P4-12.3+deb9u4 ii libisc-export160 1:9.10.3.dfsg.P4-12.3+deb9u4 ii lsb-base 9.20161125 Versions of packages isc-dhcp-server recommends: ii isc-dhcp-common 4.3.5-3+deb9u1 pn policycoreutils Versions of packages isc-dhcp-server suggests: pn isc-dhcp-server-ldap ii policykit-1 0.105-18 -- Configuration Files: /etc/dhcp/dhcpd.conf changed [not included] /etc/dhcp/dhcpd6.conf changed [not included] -- debconf information excluded
Bug#912387: dar doesn't backup maximum length files
Package: dar Version: 2.5.8-3 Severity: normal Tags: patch Dear Maintainer, Firefox when saving a complete web page seems to like to create very long filenames, on ext4 this means 255 byte filenames (that doesn't include the path, just the filename). dar then warns and doesn't backup the file. This can be reproduced by creating a file of the maximum length, 255 for ext4, and `dar -c /tmp/test -R /tmp/path_to_file` Filename provided by the operating system seems truncated in directory /tmp/path_to_file ignoring it: bbb The problem is that file doesn't get backed up. This modification changes the code to warn if it is the maximum length, and only skip if it is larger than the maximum length. I'm not sure I like the logic of ignoring it even in that case, if the file is that long, there is a problem with determining the maximum file length, if it isn't actually longer than the maximum file length, there's a bug someplace and I'm not sure skipping it helps. I would actually prefer that it not warn if it is the maximum length. The file exists, back it up, it isn't the backup program's job to critique the file names it is backing up. >From aaa60fc12f3b73b304a0b5451dd6d8ba78b4838a Mon Sep 17 00:00:00 2001 From: David Fries Date: Tue, 30 Oct 2018 21:02:26 -0500 Subject: [PATCH 1/2] complain >= bail > --- src/libdar/etage.cpp | 7 +-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/src/libdar/etage.cpp b/src/libdar/etage.cpp index 8d444869..8187e201 100644 --- a/src/libdar/etage.cpp +++ b/src/libdar/etage.cpp @@ -156,8 +156,11 @@ namespace libdar ret->d_name[max_alloc_filename] = '\0'; // yes, one byte is allocated for the terminal zero if(strlen(ret->d_name) >= max_alloc_filename) { - ui.warning(tools_printf(gettext("Filename provided by the operating system seems truncated in directory %s ignoring it: %s"), dirname, ret->d_name)); - continue; + bool ignore = strlen(ret->d_name) > max_alloc_filename; + ui.warning(tools_printf(gettext("Filename provided by the operating system seems truncated in directory %s%s: %s"), + dirname, ignore ? " ignoring it" : "", ret->d_name)); + if(ignore) + continue; } #else while(!is_cache_dir && (ret = readdir(tmp)) != nullptr) -- 2.11.0 -- System Information: Debian Release: 9.5 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 4.19.0+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15), LANGUAGE=C (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages dar depends on: ii libassuan0 2.4.3-2 ii libattr1 1:2.4.47-2+b2 ii libbz2-1.0 1.0.6-8.1 ii libc6 2.24-11+deb9u3 ii libcap21:2.25-1 ii libdar64-5000 2.5.8-3+b1 ii libgcc11:6.3.0-18+deb9u1 ii libgcrypt201.7.6-2+deb9u3 ii libgpg-error0 1.26-2 ii libgpgme11 1.8.0-3+b2 ii liblzma5 5.2.2-1.2+b1 ii liblzo2-2 2.08-1.2+b2 ii libstdc++6 6.3.0-18+deb9u1 ii zlib1g 1:1.2.8.dfsg-5 dar recommends no packages. Versions of packages dar suggests: pn dar-docs pn par2 -- no debconf information
Bug#591430: xserver-xorg-core: "Maximum number of clients reached" hardcoded
xserver-xorg-core 1.19 in stretch does have the option to increase the client limit to 512, any chance of have this the default option? I only stumbled on this when I was looking for a different bug. Section "ServerFlags" Option "MaxClients" "512" I also think this bug report should be closed now that it is fixed upstream and is in a Debian stable release. -- David Fries
Bug#910555: libkf5notifications5: KStatusNotifierItem No Icon set infinite loop
Package: libkf5notifications5 Version: 5.28.0-1 Severity: normal Dear Maintainer, I started X with fvwm and stalonetray. If I run korgac with the following theme, export QT_QPA_PLATFORMTHEME=kde I get the following repeated over and over, while it is eating up the CPU. QSystemTrayIcon::setVisible: No Icon set It looks like it is in an infinite recursion and it will eventually crash. #57045 0x7fffd5fdcb57 in ?? () from /usr/lib/x86_64-linux-gnu/qt5/plugins/platformthemes/KDEPlasmaPlatformTheme.so #57046 0x73b63a54 in ?? () from /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5 #57047 0x73b8217e in ?? () from /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5 #57048 0x73b636cd in QSystemTrayIcon::setVisible(bool) () from /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5 #57049 0x7645d5f6 in ?? () from /usr/lib/x86_64-linux-gnu/libKF5Notifications.so.5 #57050 0x7645f2de in ?? () from /usr/lib/x86_64-linux-gnu/libKF5Notifications.so.5 #57051 0x764607c3 in ?? () from /usr/lib/x86_64-linux-gnu/libKF5Notifications.so.5 #57052 0x76460cce in KStatusNotifierItem::KStatusNotifierItem(QObject*) () from /usr/lib/x86_64-linux-gnu/libKF5Notifications.so.5 #57053 0x7fffd5fdcb57 in ?? () from /usr/lib/x86_64-linux-gnu/qt5/plugins/platformthemes/KDEPlasmaPlatformTheme.so #57054 0x73b63a54 in ?? () from /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5 #57055 0x73b8217e in ?? () from /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5 #57056 0x73b636cd in QSystemTrayIcon::setVisible(bool) () ---Type to continue, or q to quit--- from /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5 #57057 0x7645d5f6 in ?? () from /usr/lib/x86_64-linux-gnu/libKF5Notifications.so.5 #57058 0x7645f2de in ?? () from /usr/lib/x86_64-linux-gnu/libKF5Notifications.so.5 #57059 0x764607c3 in ?? () from /usr/lib/x86_64-linux-gnu/libKF5Notifications.so.5 #57060 0x76460cce in KStatusNotifierItem::KStatusNotifierItem(QObject*) () from /usr/lib/x86_64-linux-gnu/libKF5Notifications.so.5 #57061 0x55576d04 in ?? () #57062 0x55569d35 in ?? () #57063 0x555663af in ?? () #57064 0x723942e1 in __libc_start_main (main=0x55565ee0, argc=1, argv=0x7fffe8c8, init=, fini=, rtld_fini=, stack_end=0x7fffe8b8) at ../csu/libc-start.c:291 #57065 0x555665da in _start () If I don't set the theme it will run, but other KDE based applications fail to find icons. I would expect if it can't find an icon that it would use a substitute, solid color, or something that will allow the application to run properly rather than eat up the CPU, ram, and eventually crash. I've looked at the korgac source and don't see it can do anything, should it be KDEPlasmaPlatformTheme.so instead? -- System Information: Debian Release: 9.5 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 4.18.0+ (SMP w/4 CPU cores) Locale: LANG=, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15), LANGUAGE= (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages libkf5notifications5 depends on: ii libc6 2.24-11+deb9u3 ii libdbusmenu-qt5-2 0.9.3+16.04.20160218-1 ii libkf5codecs5 5.28.0-1+b2 ii libkf5configcore5 5.28.0-2 ii libkf5coreaddons5 5.28.0-2 ii libkf5notifications-data 5.28.0-1 ii libkf5windowsystem5 5.28.0-2 ii libphonon4qt5-4 4:4.9.0-4 ii libqt5core5a 5.7.1+dfsg-3+b1 ii libqt5dbus5 5.7.1+dfsg-3+b1 ii libqt5gui55.7.1+dfsg-3+b1 ii libqt5widgets55.7.1+dfsg-3+b1 ii libqt5x11extras5 5.7.1~20161021-2 ii libstdc++66.3.0-18+deb9u1 ii phonon4qt54:4.9.0-4 libkf5notifications5 recommends no packages. libkf5notifications5 suggests no packages. -- no debconf information
Bug#910204: rxvt: fix terminating due to X error and fix busy waiting with a 5ms sleep when obscured
Package: rxvt Version: 1:2.7.10-7 Severity: normal Tags: patch Dear Maintainer, Here are two patches, see the git commit message for more details. rxvt is incorrectly using sizeof to calculate the format for XChangeProperty, getting 64 on a 64 bit system, when 8, 16, 32 are the only options. This results in rxvt terminating when it receives a request for a TIMESTAMP on the primary selection. If completely obscured and something requests a draw, rxvt correctly detects that there's no point in drawing, but needs to clear want_refresh to avoid sleeping 5ms and trying again. Both changes are effectively one line code changes. -- System Information: Debian Release: 9.5 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 4.18.0+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15), LANGUAGE=C (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages rxvt depends on: ii libc6 2.24-11+deb9u3 ii libx11-6 2:1.6.4-3 ii libxpm4 1:3.5.12-1 rxvt recommends no packages. rxvt suggests no packages. -- no debconf information >From f03a0f43a92bb06acefe29a62b50001218f4c45a Mon Sep 17 00:00:00 2001 From: David Fries Date: Sat, 29 Sep 2018 22:26:43 -0500 Subject: [PATCH 1/3] fix XA_TIMESTAMP XChangeProperty format value If rxvt owns the primary selection buffer and another program queries the TIMESTAMP as the target to XConvertSelection, rxvt will exit with the following X error (after disabling rxvt's error handler to let the built in handler decode it). X Error of failed request: BadValue (integer parameter out of range for operation) Major opcode of failed request: 18 (X_ChangeProperty) Value in failed request: 0x40 Serial number of failed request: 232 Current serial number in output stream: 233 The failed value is 0x40 which is 64, which is (8 * sizeof(Time)) on a 64 bit system, but the man page to XChangeProperty property states, "Possible values are 8, 16, and 32." This is an integer, set the value to 32. source code to a simple program to request the TIMESTAMP from the primary selection owner can be found here, https://gitlab.com/dfries64/xconvertselection_timestamp --- src/screen.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/screen.c b/src/screen.c index 5f3ed75..dd5def7 100644 --- a/src/screen.c +++ b/src/screen.c @@ -3647,7 +3647,7 @@ rxvt_selection_send(rxvt_t *r, const XSelectionRequestEvent *rq) /* TODO: Handle MULTIPLE */ } else if (rq->target == r->h->xa[XA_TIMESTAMP] && r->selection.text) { XChangeProperty(r->Xdisplay, rq->requestor, rq->property, XA_INTEGER, - (8 * sizeof(Time)), PropModeReplace, + 32, PropModeReplace, (unsigned char *)&r->h->selection_time, 1); ev.property = rq->property; } else if (rq->target == XA_STRING -- 2.11.0 >From 1f5325b8c546a859d4f00a02292656187e19f8e4 Mon Sep 17 00:00:00 2001 From: David Fries Date: Thu, 27 Sep 2018 22:47:51 -0500 Subject: [PATCH 2/3] rxvt eats cpu when needing to draw while obscured If the window is completely obscured and needs to draw then rxvt_scr_refresh will return immediately instead of drawing, (no need to draw what can't be seen), but if want_refresh is true, it will do a 5ms select timeout and try again forever. The solution is to clear want_refresh before returning from rxvt_scr_refresh. To reproduce run `strace rxvt`, then `top` in that rxvt window (or anything with a delay that will trigger a draw, `sleep 5` works), completely cover the rxvt window with another and wait for the timeout. strace will show a steady stream of select (which times out), and recvmsg. --- src/screen.c | 9 - 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/src/screen.c b/src/screen.c index dd5def7..d637930 100644 --- a/src/screen.c +++ b/src/screen.c @@ -2046,8 +2046,15 @@ rxvt_scr_refresh(rxvt_t *r, unsigned char refresh_type) int (*draw_string_func) () = draw_string; int (*draw_image_string_func) () = draw_image_string; -if (refresh_type == NO_REFRESH || !r->TermWin.mapped) +if (refresh_type == NO_REFRESH || !r->TermWin.mapped) { +/* If the window is completely obscured (NO_REFRESH) we really don't + * want to refresh, that is why we are returning instead, but first we + * must say mark the flag don't refresh. A similar argument works for + * !mapped. + */ + h->want_refresh = 0; return; +} /* * A: set up vars -- 2.11.0 >From f12ebf710079400f44e9b16fc0bf7c2a9635fd1c Mon Sep 17 00:00:00 2001 From: David Fries Date: Wed, 26 Sep 2018 21:55:47 -0500 Subject: [PATCH 3/3] change long with hang, hang,
Bug#890753: rxvt processes get stuck after the shell exits
I'm running into the same problem and with one of my systems it's seems like 10% of the terminals hang this way. Here's the fix, applied against rxvt_2.7.10-7.dsc There are two similar problems, both involving deadlocks with a mutex. seteuid() gets a mutex, is interrupted with SIGCHLD signal, calls the same exit path and seteuid() blocks waiting formutex malloc gets a mutex, is interrupted with SIGCHLD signal, rxvt_clean_exit calls XDestroyIC(), and free() blocks waiting for mutex -- David Fries >From 39c8effc203f939146658cee5d93cb8a0936244a Mon Sep 17 00:00:00 2001 From: David Fries Date: Wed, 26 Sep 2018 21:05:15 -0500 Subject: [PATCH 1/2] fix deadlock in recursive seteuid calls rxvt is a signal threaded program, but signals are causing mutex deadlocks. When the command exits two things happen, rxvt will receive a read error from the tty and it will receive a SIGCHLD signal at nearly the same time. If the read error happens first, rxvt_cmd_getc() was calling rxvt_clean_exit() directly, which calls rxvt_privileged_ttydev/utmp() which calls seteuid() which must internally get some kind of mutex. If in seteuid() the SIGCHLD is delivered, it will also call exit() and rxvt_clean_exit() and on to seteuid() and deadlock waiting for the mutex. That sequence is not theoretical, at least two people have filed bug reports and one of my systems sees 10% of rxvts hang like this. It seems the best option is to disable signals both before calling rxvt_clean_exit()/exit() in rxvt_cmd_getc() and also in rxvt_clean_exit() just in case two signals might be delivered back to back. Given exit() calls rxvt_clean_exit() there's no need for rxvt_cmd_getc() to call it before exit and run through it twice (provided registered with onexit() earlier). --- src/command.c | 18 ++ src/init.c | 17 + src/init.intpro | 1 + src/main.c | 11 +++ 4 files changed, 47 insertions(+) diff --git a/src/command.c b/src/command.c index 976d087..d559050 100644 --- a/src/command.c +++ b/src/command.c @@ -46,6 +46,7 @@ /*{{{ includes: */ #include "../config.h" /* NECESSARY */ #include "rxvt.h" /* NECESSARY */ +#include "init.intpro" /* PROTOS for internal routines */ #include "version.h" #include "command.h" @@ -751,7 +752,24 @@ rxvt_cmd_getc(rxvt_t *r) else if (n == 0 || (n < 0 && errno == EAGAIN)) break; else { + /* + * One likely reason for a read error is the child just + * exited. If that's the case it's just a matter of time + * before SIGCHLD handler runs, which also runs exit(), + * and if that signal is called in the middle of a call + * to seteuid/setegid the process can dead lock, because + * there is an internal mutex, which is then locked the + * second time through. Disable signals before calling + * exit. If the signal goes off before it is disabled, + * it calls exit() and control never returns here, if the + * signal is disabled first exit will run without + * interruption. + */ + rxvt_disable_signals(); +#if !defined (HAVE_ATEXIT) && !defined (HAVE_ON_EXIT) + /* will be called from exit if registered */ rxvt_clean_exit(); +#endif exit(EXIT_FAILURE); /* bad order of events? */ } if (count != BUFSIZ) /* some characters read in */ diff --git a/src/init.c b/src/init.c index 65f4d70..9a2a3b3 100644 --- a/src/init.c +++ b/src/init.c @@ -1382,6 +1382,23 @@ rxvt_run_command(rxvt_t *r, const char *const *argv) return cfd; } +/*--*/ +/* + * Disable signals for use during the exit path. Combines them all in one + * place. + */ +/* INTPROTO */ +void +rxvt_disable_signals(void) +{ +#ifndef __svr4__ +signal(SIGINT, SIG_IGN); +#endif +signal(SIGQUIT, SIG_IGN); +signal(SIGTERM, SIG_IGN); +signal(SIGCHLD, SIG_IGN); +} + /* - * * CHILD PROCESS OPERATIONS * * - */ diff --git a/src/init.intpro b/src/init.intpro index 8759272..8e9c983 100644 --- a/src/init.intpro +++ b/src/init.intpro @@ -3,5 +3,6 @@ void rxvt_Get_Colours __PROTO((rxvt_t *r)); void rxvt_color_aliases __PROTO((rxvt_t *r, int idx)); void rxvt_get_ourmods __PROTO((rxvt_t *r)); int rxvt_run_command __PROTO((rxvt_t *r, const char *const *argv)); +void rxvt_disable_signals __PROTO((void)); int rxvt_run_child __PROTO((rxvt_t *r, const char *const *argv)); void rxvt_get_ttymode __PROTO((ttymode_t *tio)); diff --
Bug#860296: lirc is spamming syslog with messages
Package: lirc Version: 0.9.4c-9 Followup-For: Bug #860296 Dear Maintainer, Just upgraded to stretch and got bit by this as well, it's been roughly a day and it's logged 149,336 messages to syslog, purging ASAP. Would it be accepted if I wrote a patch to write this message once on startup and again once if it finds a device and that device is closed? Sep 19 23:47:26 localhost lircd[24438]: lircd-0.9.4c[24438]: Error: Cannot glob /sys/class/rc/rc0/input[0-9]*/event[0-9]* Sep 19 23:47:26 localhost lircd-0.9.4c[24438]: Error: Cannot glob /sys/class/rc/rc0/input[0-9]*/event[0-9]* -- System Information: Debian Release: 9.5 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 4.18.0+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15), LANGUAGE=C (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages lirc depends on: ii init-system-helpers 1.48 ii libasound2 1.1.3-5 ii libc62.24-11+deb9u3 ii libftdi1-2 1.3-2+b2 ii libgcc1 1:6.3.0-18+deb9u1 ii liblirc-client0 0.9.4c-9 ii liblirc0 0.9.4c-9 ii libportaudio219.6.0-1 ii libstdc++6 6.3.0-18+deb9u1 ii libsystemd0 232-25+deb9u4 ii libudev1 232-25+deb9u4 ii libusb-0.1-4 2:0.1.12-30 ii libusb-1.0-0 2:1.0.21-1 ii lsb-base 9.20161125 ii python3 3.5.3-1 Versions of packages lirc recommends: pn gir1.2-vte ii python3-gi3.22.0-2 pn python3-yaml Versions of packages lirc suggests: pn ir-keytable pn lirc-compat-remotes pn lirc-doc pn lirc-drv-irman ii lirc-x 0.9.4c-9 ii setserial2.17-50 -- Configuration Files: /etc/lirc/lircd.conf [Errno 2] No such file or directory: '/etc/lirc/lircd.conf' -- no debconf information
Bug#889476: /usr/lib/x86_64-linux-gnu/libgtk-x11-2.0.so.0.2400.31: gtk_file_chooser auto completes when it shouldn't, can't select directory with one subdirectory
Package: libgtk2.0-0 Version: 2.24.31-2 Severity: important File: /usr/lib/x86_64-linux-gnu/libgtk-x11-2.0.so.0.2400.31 Tags: upstream Dear Maintainer, >From chromium "Load unpacked extension..." it brought up what I assume is the gtk file chooser. Enable "Type a file name" then I put in the directory to my extension (calling it /abs_path here). It added a slash at the end so now /abs_path/ , I pressed Enter, and I can see it briefly flash on /abs_path/.git before the dialog closed, chromium gives an error loading extension because it has the /abs_path/.git path. I gave it the correct path, it decided to auto complete something else and not only that but to do so after accepting the path I gave it doesn't give the user a chance to fix the program's wrong choice. It will also auto complete to an invalid filename. In chromimum select a file to upload, do the "Type a file name" option, put in the full path to a bunch of image files that all have the same first few characters. After putting in the path into the location I press enter, any other file browser will go to that directory, this will instead will provide the application with path/prefix where prefix is the first few characters common to all the name files in that directory. The result is giving the application a file name that doesn't exist. Don't auto select a file, ever! If you give the "Type a file name" a directory that has only one file in it, and press enter, in the case of uploading a browser file, it provides that file to the browser, which uploads it, and the user has no chance to stop. Browsers have required a user to select the file for years to avoid uploading security sensitive files. In this case the user would have to mindfully use a separate program to view that location to verify that yes it is the file they wanted to upload. -- System Information: Debian Release: 9.3 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Kernel: Linux 4.13.0 (SMP w/4 CPU cores) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), LANGUAGE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages libgtk2.0-0:amd64 depends on: ii adwaita-icon-theme 3.22.0-1+deb9u1 ii gnome-icon-theme 3.12.0-2 ii hicolor-icon-theme 0.15-1 ii libatk1.0-0 2.22.0-1 ii libc62.24-11+deb9u1 ii libcairo21.14.8-1 ii libcups2 2.2.1-8 ii libfontconfig1 2.11.0-6.7+b1 ii libfreetype6 2.6.3-3.2 ii libgdk-pixbuf2.0-0 2.36.5-2+deb9u2 ii libglib2.0-0 2.50.3-2 ii libgtk2.0-common 2.24.31-2 ii libpango-1.0-0 1.40.5-1 ii libpangocairo-1.0-0 1.40.5-1 ii libpangoft2-1.0-01.40.5-1 ii libx11-6 2:1.6.4-3 ii libxcomposite1 1:0.4.4-2 ii libxcursor1 1:1.1.14-1+deb9u1 ii libxdamage1 1:1.1.4-2+b3 ii libxext6 2:1.3.3-1+b2 ii libxfixes3 1:5.0.3-1 ii libxi6 2:1.7.9-1 ii libxinerama1 2:1.1.3-1+b3 ii libxrandr2 2:1.5.1-1 ii libxrender1 1:0.9.10-1 ii shared-mime-info 1.8-1 Versions of packages libgtk2.0-0:amd64 recommends: ii libgail-common 2.24.31-2 ii libgtk2.0-bin 2.24.31-2 Versions of packages libgtk2.0-0:amd64 suggests: ii gvfs 1.30.4-1 ii librsvg2-common 2.40.16-1+b1 -- no debconf information
Bug#883554: cups keeps breaking network printer with implicitclass:
y exists E [14/Dec/2017:01:45:34 -0600] [Job 1403] No destination host name supplied by cups-browsed for printer \"Canon_BJC-2100\", is cups-browsed running? > 6. Now set up this queue: > lpadmin -p 2100 -v ipp://server_ip/print/queue_name -E -m raw lpadmin -p ipps_Canon_BJC-2100 -v ipps://.local:631/printers/Canon_BJC-2100 -E -m raw > 7. You should be able to print to both queues. implicitclass is failing, ipps_Canon_BJC-2100 is printing. ipps_Canon_BJC-2100 is working. I'll have to try again tomorrow. > [Snip] > > I understand. Don't rush. If there is a bug it appears to be in > cups-browsed. But I don't experience it - so I hope you can see > my quandry. As a data point, my dad has been finding out that it prints for a little bit of time after he reboots the client. I would assume it is using the implicitclass entry, at least it only has the one queue when I look at it and he's not going to be adding/removing printers. Question, do you have the Canon BJC-2100 model printer? We're running with the replacement BC-20 ink cartridge, that's a black only cartridge. CUPS prints slow, as in the speed of printing only using the nozzles of a color cartridge. The default options in general is Color Model: Grayscale, printer Features Common, Ink Type: Black, Ink Set: Black. Is there anything I'm missing? lprng, magicfilter, and gs prints much faster, here's the line. /etc/magicfilter/bj800-filter filter /usr/bin/gs -q -dSAFER -dNOPAUSE -r360 -sDEVICE=bjc800 -dMonochromePrint=true -sOutputFile=- - -c quit -- David Fries
Bug#883554: cups keeps breaking network printer with implicitclass:
On Tue, Dec 12, 2017 at 12:10:22PM +, Brian Potkin wrote: > On Wed 06 Dec 2017 at 18:56:33 +, Brian Potkin wrote: > > On Tue 05 Dec 2017 at 19:43:54 -0600, David Fries wrote: > > > On Tue, Dec 05, 2017 at 01:52:53PM +, Brian Potkin wrote: > > > > On Mon 04 Dec 2017 at 23:47:04 -0600, David Fries wrote: > > Forget about this. It doesn't help towards a solution and, if I had > > thought on about it, I should not have sought out the information I > > was after with that command. > > > > What I wanted to find out was whether you used a PPD with the CUPS > > web interface and an ipp:// or dnssd:// URI. Because you can print > > (initially at least) it implies you didn't. I don't know how to answer your question, this is the sequence I do go get it to print, then after a period of time it reverts to something that's broken. Does this select a PPD? Through the web interface I go to modify printer, Discovered Network Printers, continue, Make: Canon, Model, "Current Driver - Remote printer: Canon BJC-2100 - CUPS+Gutenprint v5.2.11", Modify Printer. Then it lists the connection with dnssd://Canon%20BJC-2100%20%40%20... > > > The majority seems to be different page sizes, dithering, and such. > > > The header and cupsFilter might be of interest, so included here. > > > > > > --- Canon_BJC-2100.ppd 2017-12-05 09:35:07.689792328 -0600 > > > > This is the PPD on the client? > > > > > +++ Canon_BJC-2100_remote.ppd 2015-11-29 00:33:04.432331311 -0600 > > > > This is the PPD on the server? No, both were from the client, the Canon_BJC-2100_remote.ppd was from 2015. > > > lpstat -t > > > scheduler is running > > > system default destination: Canon_BJC-2100 > > > device for Canon_BJC-2100: implicitclass:Canon_BJC-2100 > > > Canon_BJC-2100 accepting requests since Tue Dec 5 19:19:12 2017 > > > printer Canon_BJC-2100 is idle. enabled since Tue Dec 5 19:19:12 2017 > > > > > > How about this sequence, I go through the web interface, change it to > > > the ipps URI that works. Verify in the browser it is the ipps URI, > > > verify with `lpstat -t` it is ipps URI, wait until > > > /etc/cups/printers.conf has the ipps DeviceURI, then. > > > systemctl stop cups-browsed.service > > > systemctl start cups-browsed.service > > > and the browser immediately lists implicitclass:Canon_BJC-2100 > > > > I can now reproduce your observations, apart from the non-printing > > aspect. Do you get implicitclass:Canon_BJC-2100 ? > > 1. Start with cups-browsed running and no queue using ipp:// or > >dnssd://. Printing takes place for me. For you I suspect it > >doesn't. The PPD in /etc/cups/ppd is obtained by cups-browsed > >from the server and modified slightly. The queue is a raw > >queue and the PPD is only there so that applications know what > >to display in their dialogs; it does not lead to any filtering > >on the client to alter the submitted job file. > > > > 2. Configure a raw queue with ipp:// having the same queue name as > >on the server. This setup method removes the existing PPD and > >overrides the cups-browsed automatic setup. Printing takes place > >for both of us. > > > > 3. At some future time cups-browsed refreshes what it knows about > >remote queues, using what is in /var/cache/cups, and reinstates > >the queue, once again getting the PPD it had in 1. I can still > >print. (With a different queue name from the server's in 2 you > >would also be able to print). > > > > > I tried printing some text with lpr, after while the state was > > > aborted. > > > > > > lpstat -t > > > scheduler is running > > > system default destination: Canon_BJC-2100 > > > device for Canon_BJC-2100: implicitclass:Canon_BJC-2100 > > > Canon_BJC-2100 accepting requests since Tue Dec 5 19:32:48 2017 > > > printer Canon_BJC-2100 is idle. enabled since Tue Dec 5 19:32:48 > > > 2017 > > > > I would question whether having two queues and two setup mehods > > managing them is for the best. I'm not saying you shouldn't be able > > to do it but my advice would be > > > > 1. Keep cups-browsed on the client. It automatically sets up print > >queues from the server's DNS-SD broadcasts. Do not manually set > >up any other queue. I'm not sure what I need to do for this. Delete all printers on the client and something is just magically supposed to work? I go to manage printers and only one is listed, from what you wrote I expect you are thinking there are two? > > 2. Stop (or purge) cups-browsed and set up a queue with an ipp:// > >URI and no PPD (raw). when you say stop you mean uninstall? This sounds preferable to me, I give it a ipps URI and it doesn't change it. > > Both methods have been tested here to work for your printer in your > > circumstances. > > > > Have a look at the server and let us know how you go on. > > Any progress on this, David? Sorry finding time is hard, thanks for being responsive. -- David Fries
Bug#883554: cups keeps breaking network printer with implicitclass:
On Tue, Dec 05, 2017 at 01:52:53PM +, Brian Potkin wrote: > On Mon 04 Dec 2017 at 23:47:04 -0600, David Fries wrote: > > > Package: cups > > Version: 2.2.1-8 > > Severity: important > > > > Dear Maintainer, > > > > It seems to be every day or so the /etc/cups/printers.conf DeviceURI is > > modified to replace the version that works with a version that doesn't. > > Knowing "...the version that works..." would be useful. Doesn't work: DeviceURI implicitclass:Canon_BJC-2100 web job state held since Tue Dec 5 19:00:56 2017 "No suitable destination host found by cups-browsed." works (after replacing in server name) was in there since 2015: DeviceURI ipps://.local:631/printers/Canon_BJC-2100 That is it works for about a day until it changes to the one that doesn't. > > This is printing to a cups system on the local network. > > Was working with the cups in Debian jessie. > > What is the cups version on the server? Can you print from the server? Both the serve and client were running Debian jessie prior to Thanksgiving, both were upgraded to stretch. On the client I see in aptitude.1.gz [UPGRADE] cups:amd64 1.7.5-11+deb8u1 -> 2.2.1-8 Both client and server lists cups as version 2.2.1-8 Yes the server was printing before and after the upgrade. > > It doesn't matter if I use a dnssd (auto detected), ipps, ipp, it gets > > replaced by, > > DeviceURI implicitclass:Canon_BJC-2100 > > and that doesn't allow it to print. I've even gone so far as deleting > > the printer, the detecting it through the ipp web configuration > > interface and after a day or so it goes back to the implicitclass. > > I do not understand the existence of the time lag. It doesn't make any sense to me either when I'm using the cups browser interface and letting it write out the file that it works for a while and then it fails.. After the upgrade I can configure the client with that ipps and it is able to print, for about a day, and then printing fails and I check printers.conf and it is replaced with, DeviceURI implicitclass:Canon_BJC-2100 and it can no longer print. This is one of the files that cups keeps modifying. I've been using the web configuration interface when modifying it, and have used the "Discovered Network Printers:" option which fills in a long dnssd://Canon%20BJC-2100%20%40%20... but while it initially prints it doesn't matter that will be replaced by implicitclass:Canon_BJC-2100 the next day. > > Any ideas? > > There will be a PPD for the BCJ-2100 in /etc/cups/ppd. Please do > /usr/sbin/cupsfilter -p -m printer/foo -e --list-filters /etc/services /usr/sbin/cupsfilter -p /etc/cups/ppd/Canon_BJC-2100.ppd -m printer/foo -e --list-filters /etc/services That gives no output at all. I have a different version from 2015 in that directory with the following output. cupsfilter: File "/usr/lib/cups/filter/rastertogutenprint.5.2" permissions OK (040755/uid=0/gid=0). texttopdf pdftopdf gstoraster rastertogutenprint.5.2 The majority seems to be different page sizes, dithering, and such. The header and cupsFilter might be of interest, so included here. --- Canon_BJC-2100.ppd 2017-12-05 09:35:07.689792328 -0600 +++ Canon_BJC-2100_remote.ppd 2015-11-29 00:33:04.432331311 -0600 @@ -15,7 +15,7 @@ *% Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. *% *FormatVersion:"4.3" -*FileVersion: "5.2.11" +*FileVersion: "5.2.10" *LanguageVersion: English *LanguageEncoding: ISOLatin1 *PCFileName: "STP00011.PPD" @@ -23,17 +23,17 @@ *Product: "(Canon BJC-2100)" *ModelName: "Canon BJC-2100" *ShortNickName: "Canon BJC-2100" -*NickName: "Remote printer: Canon BJC-2100 - CUPS+Gutenprint v5.2.11" +*NickName: "Canon BJC-2100 - CUPS+Gutenprint v5.2.10" *PSVersion:"(3010.000) 0" *LanguageLevel:"3" *ColorDevice: True *DefaultColorSpace:RGB -*cupsManualCopies: True *FileSystem: False *LandscapeOrientation: Plus90 *TTRasterizer: Type42 *cupsVersion: 1.2 -*cupsFilter: "*/* 0 -" +*cupsManualCopies: True +*cupsFilter: "application/vnd.cups-raster 100 rastertogutenprint.5.2" *1284DeviceID: "MFG:Canon;MDL:BJC-2100;DES:Canon BJC-2100;" *cupsLanguages: "ca cs da de el en_GB es fi fr gl hu it ja nb nl pl pt ru sk sl sv tr uk vi zh_CN zh_TW" > and post its output and the outputs of > > systemctl status cups-browsed cups-browsed.service - Make remote CUPS printers available locally Loaded: loaded (/lib/systemd/system/cups-browsed.service; enabled; vendor preset: enabled) Active: active (running) since Tue 2017-12-05 09:34:59 CST; 9h ago Mai
Bug#883554: cups keeps breaking network printer with implicitclass:
Package: cups Version: 2.2.1-8 Severity: important Dear Maintainer, It seems to be every day or so the /etc/cups/printers.conf DeviceURI is modified to replace the version that works with a version that doesn't. This is printing to a cups system on the local network. Was working with the cups in Debian jessie. It doesn't matter if I use a dnssd (auto detected), ipps, ipp, it gets replaced by, DeviceURI implicitclass:Canon_BJC-2100 and that doesn't allow it to print. I've even gone so far as deleting the printer, the detecting it through the ipp web configuration interface and after a day or so it goes back to the implicitclass. Any ideas? -- System Information: Debian Release: 9.1 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 4.9.0-4-amd64 (SMP w/2 CPU cores) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages cups depends on: ii cups-client2.2.1-8 ii cups-common2.2.1-8 ii cups-core-drivers 2.2.1-8 ii cups-daemon2.2.1-8 ii cups-filters 1.11.6-3 ii cups-ppdc 2.2.1-8 ii cups-server-common 2.2.1-8 ii debconf [debconf-2.0] 1.5.61 ii ghostscript9.20~dfsg-3.2+deb9u1 ii libavahi-client3 0.6.32-2 ii libavahi-common3 0.6.32-2 ii libc-bin 2.24-11+deb9u1 ii libc6 2.24-11+deb9u1 ii libcups2 2.2.1-8 ii libcupscgi12.2.1-8 ii libcupsimage2 2.2.1-8 ii libcupsmime1 2.2.1-8 ii libcupsppdc1 2.2.1-8 ii libgcc11:6.3.0-18 ii libstdc++6 6.3.0-18 ii libusb-1.0-0 2:1.0.21-1 ii poppler-utils 0.48.0-2 ii procps 2:3.3.12-3 Versions of packages cups recommends: ii avahi-daemon 0.6.32-2 ii colord 1.3.3-2 ii cups-filters [ghostscript-cups] 1.11.6-3 ii printer-driver-gutenprint5.2.11-1+b2 Versions of packages cups suggests: ii cups-bsd 2.2.1-8 pn cups-pdf ii foomatic-db-compressed-ppds [foomatic-db] 20161201-1 ii hplip 3.16.11+repack0-3 ii printer-driver-hpcups 3.16.11+repack0-3 ii smbclient 2:4.5.12+dfsg-2+deb9u1 ii udev 232-25+deb9u1 -- Configuration Files: /etc/default/cups changed: -- debconf information: cupsys/backend: lpd, socket, usb, snmp, dnssd cupsys/raw-print: true
Bug#784868: korganizer: korgac grabs focus loosing keys and missing notifications
I tried, I need to find some time to figure out reviewboard and get the patch resubmitted as requested upstream. I think it also has a window border creeping issue, that is it is offset a little bit each time it (fails) to put it back exactly where it was. https://bugs.kde.org/show_bug.cgi?id=347486 On Fri, May 27, 2016 at 10:52:54PM +0200, Maximiliano Curia wrote: > ¡Hola David! > > El 2015-05-09 a las 15:11 -0500, David Fries escribió: > >Package: korganizer Version: 4:4.14.1-1nograb Tags: patch upstream > > >While it can make it easier to dismiss (if it grabs focus through the > >activateWindow call), it also means if you are typing and a reminder comes > >in it will both get those keystrokes and a space or return will dismiss > >the reminder before you have a chance to ever see what the reminder was > >about, which assumes you are even looking at your screen to see the > >reminder box flash up, if you are typing from paper you might not even > >know. > > Sorry for having missed this bug report. The patches look fine, but the code > base has changed since the bug was reported and the patches will need to be > adapted to a newer release. Also, since this issue deals with a non Debian > specific behaviours it would be better to send the patches upstream > directly. > > In the pkg-kde team we are currently under staffed and it would really help > us if you send the updated patch series upstream. > > I can help you with these if you need guidance. > > Happy hacking,
Bug#821048: [Pkg-alsa-devel] Bug#821048: alsa-utils: arecord -d broken, apply upstream patch
On Fri, Apr 15, 2016 at 08:06:15AM +0200, Elimar Riesebieter wrote: > Please update to alsa-utils 1.1.0-2 first. In stable there will be > only security patches accepted. Bug closed herewith. Backports are > not maintained from us. 1.1.0-2 is fixed as far as this bug report goes for the -d option. This could be seen as a denial of service, security bug, unless there is a directory limit, this can create files until it runs the system out of inodes. I was more of wanting to say the next version is fixed, thanks for the quick disposition, I don't need a response to this e-mail. -- David Fries
Bug#821048: alsa-utils: arecord -d broken, apply upstream patch
Package: alsa-utils Version: 1.0.28-1 Severity: important Tags: patch Dear Maintainer, >From the upstream repository git://git.alsa-project.org/alsa-utils 85827fbb6424 fixes the problem (see that commit message for the details) and 8f361d83cfcb avoids the problem. In summary -d has a very high likelihood of continually writing over the file you are trying to record (after capturing that much audio), or continually writing out 44 byte files, and I had 3.1 million to delete. This hasn't been a problem upstream for nearly 2 years, but Debian stable isn't on that release. Either patch, or both will address the problem. >From 8f361d83cfcb39887f5fc591633e68d9448e3425 Mon Sep 17 00:00:00 2001 From: Jaroslav Kysela Date: Wed, 1 Oct 2014 15:43:57 +0200 Subject: [PATCH] Revert "aplay: fix pcm_read() return value" This reverts commit 8aa13eec80eac312e4b99423909387660fb99b8f. The semantics for pcm_read() and pcm_readv() was changed, but the callers expect the exact frame count as requested. It's possible to fix callers, but the fix is more complicated than to revert the change. Note that '-d' processing was broken in some cases. Note: The reverted commit allows that the return value might be greater than requested (see the first condition in read routines). --- aplay/aplay.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/aplay/aplay.c b/aplay/aplay.c index 30d3f31..e58e1bc 100644 --- a/aplay/aplay.c +++ b/aplay/aplay.c @@ -2039,7 +2039,7 @@ static ssize_t pcm_read(u_char *data, size_t rcount) data += r * bits_per_frame / 8; } } - return result; + return rcount; } static ssize_t pcm_readv(u_char **data, unsigned int channels, size_t rcount) @@ -2084,7 +2084,7 @@ static ssize_t pcm_readv(u_char **data, unsigned int channels, size_t rcount) count -= r; } } - return result; + return rcount; } /* -- 2.1.4 >From 85827fbb642463ab724a9471a7a88f93fa2a217d Mon Sep 17 00:00:00 2001 From: David Fries Date: Wed, 13 Apr 2016 23:32:46 -0500 Subject: [PATCH] aplay: fix lurking capture file overwrite bug If -d was given to arecord while commit 8aa13eec80eac312e4b99423909387660fb99b8f (now reverted) was in effect, the last read would be shorter than the chunk size, but pcm_read would read and return the chunk size, the samples were discarded, and capture() continued in a loop because count never reached 0. arecord opens a new file each loop iteration, if arecord is dynamically naming files, --use-strftime option or beyond the wave 2GB limit, this will generate a series of header only wave files. If the file is unique the originally recorded data is lost and it will continue overwriting the same file with a header only wave file. While the current pcm_read can't fail (it can exit), it is better to just fix this lurking bug in case it is "fixed" again. Signed-off-by: Takashi Iwai --- aplay/aplay.c | 9 ++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/aplay/aplay.c b/aplay/aplay.c index 7acaa83..2da7dda 100644 --- a/aplay/aplay.c +++ b/aplay/aplay.c @@ -3067,11 +3067,14 @@ static void capture(char *orig_name) size_t c = (rest <= (off64_t)chunk_bytes) ? (size_t)rest : chunk_bytes; size_t f = c * 8 / bits_per_frame; - if (pcm_read(audiobuf, f) != f) + if (pcm_read(audiobuf, f) != f) { + in_aborting = 1; break; + } if (write(fd, audiobuf, c) != c) { perror(name); - prg_exit(EXIT_FAILURE); + in_aborting = 1; + break; } count -= c; rest -= c; @@ -3091,7 +3094,7 @@ static void capture(char *orig_name) } if (in_aborting) - break; + prg_exit(EXIT_FAILURE); /* repeat the loop when format is raw without timelimit or * requested counts of data are recorded -- 2.1.4 -- System Information: Debian Release: 8.3 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 4.3.0-rc7+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages alsa-utils depends on: ii dialog 1.2-20140911-1 ii kmod18-3 ii libasound2 1.0.28-1 ii libc6 2.19-18+deb8u3 ii libncursesw55.9+20140913-1+b1 ii libsamplerate0 0.1.8-8 ii libtinfo5
Bug#807531: scrollbars are snapping back moving mouse away from scroll bar
On Sat, Dec 12, 2015 at 08:56:12PM -0500, Michael Gilbert wrote: > control: tag -1 confirmed, -moreinfo > > On Sat, Dec 12, 2015 at 8:40 PM, David Fries wrote: > > Which version of chromium are you running that isn't showing the > > problem? If there is some other way for the scroll bars to work > > properly that would be the only alternative to patching it seems. Or > > do you have any suggestions on how to work with upstream to make this > > under control of a theme, or setting, or anything? > > You said it was a few pixels. For my setup, I had to move the mouse > over at least a fourth of the window, which is why I couldn't > reproduce it. I'm sorry, yes I should have been more specific, I measured it now at 138 pixels, which is a tenth of a window for me at home or a twentieth at work. > If it is only a couple pixels on your setup, then that really is a > bug, and upstream shouldn't ignore you, but you do need to provide > them more information. Unless they were to make the distance be the width of the window (minus a pixel in case it is full screen), I can no longer use the pointer to point things out while dragging the scroll bar or scroll and move directly to a link. This is a regression of expected UI behavior on Linux making it inconsistent with every other program I'm aware of. Will Debian carry simple patches that help make the user experience more cohesive? -- David Fries
Bug#807531: chromium: [Regression] scrollbars are snapping back moving mouse away from scroll bar
Package: chromium Version: 46.0.2490.71-1~deb8u1unix Severity: normal Tags: upstream Dear Maintainer, Upstream chromium broke the scroll bar behavior, move a few pixels away from the scroll bar and it jumps back someplace else, this is so annoying I had to make a patch to continue to use the browser. Reported bug upstream, but they seem to be hostile/don't care about how broken this leaves chromium or how it doesn't match how any other X11 platform scroll bars I've ever used work. Suggestions? Seems like something that should be at least a setting, or in the theme, or something so Google/chrome doesn't drive Unix users away. https://code.google.com/p/chromium/issues/detail?id=567968 >From 1c2da7540a6840234ed10b99ca04581c8ff9389b Mon Sep 17 00:00:00 2001 From: David Fries Date: Sat, 5 Dec 2015 15:15:15 -0600 Subject: [PATCH 1/2] Conform to platform expectations, no snapback on Linux Besides it's ultra annoying. --- content/renderer/webscrollbarbehavior_impl_gtkoraura.cc | 9 + 1 file changed, 9 insertions(+) diff --git a/content/renderer/webscrollbarbehavior_impl_gtkoraura.cc b/content/renderer/webscrollbarbehavior_impl_gtkoraura.cc index 8dfa998..179d17f 100644 --- a/content/renderer/webscrollbarbehavior_impl_gtkoraura.cc +++ b/content/renderer/webscrollbarbehavior_impl_gtkoraura.cc @@ -25,6 +25,15 @@ bool WebScrollbarBehaviorImpl::shouldSnapBackToDragOrigin( const blink::WebPoint& eventPoint, const blink::WebRect& scrollbarRect, bool isHorizontal) { + +// no snapback at all, this is the expected behavior for everything, otherwise +// you can't use your mouse to point out things in the center of the +// document while scrolling, use escape to snap back (apparently someone +// disabled escape) +#if defined(OS_LINUX) + return false; +#endif + // Constants used to figure the drag rect outside which we should snap the // scrollbar thumb back to its origin. These calculations are based on // observing the behavior of the MSVC8 main window scrollbar + some -- 2.1.4 -- System Information: Debian Release: 8.2 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 4.3.0-rc7+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages chromium depends on: ii libasound2 1.0.28-1 ii libatk1.0-0 2.14.0-1 ii libc62.19-18+deb8u1 ii libcairo21.14.0-2.1 ii libcups2 1.7.5-11+deb8u1 ii libdbus-1-3 1.8.20-0+deb8u1 ii libexpat12.1.0-6+deb8u1 ii libfontconfig1 2.11.0-6.3 ii libfreetype6 2.5.2-3+deb8u1 ii libgcc1 1:4.9.2-10 ii libgdk-pixbuf2.0-0 2.31.1-2+deb8u3 ii libglib2.0-0 2.42.1-1 ii libgnome-keyring03.12.0-1+b1 ii libgtk2.0-0 2.24.25-3 ii libjpeg8 8d-1+deb7u1 ii libnspr4 2:4.10.7-1 ii libnspr4-0d 2:4.10.7-1 ii libnss3 2:3.17.2-1.1+deb8u2 ii libnss3-1d 2:3.17.2-1.1+deb8u2 ii libpango-1.0-0 1.36.8-3 ii libpangocairo-1.0-0 1.36.8-3 ii libpci3 1:3.2.1-3 ii libspeechd2 0.8-7 ii libsrtp0 1.4.5~20130609~dfsg-1.1 ii libstdc++6 4.9.2-10 ii libx11-6 2:1.6.2-3 ii libxcomposite1 1:0.4.4-1 ii libxcursor1 1:1.1.14-1+b1 ii libxdamage1 1:1.1.4-2+b1 ii libxext6 2:1.3.3-1 ii libxfixes3 1:5.0.1-2+b2 ii libxi6 2:1.7.4-1+b2 ii libxml2 2.9.1+dfsg1-5 ii libxrandr2 2:1.4.2-1+b1 ii libxrender1 1:0.9.8-1+b1 ii libxslt1.1 1.1.28-2+b2 ii libxss1 1:1.2.2-1 ii libxtst6 2:1.2.2-1+b1 ii x11-utils7.7+2 ii xdg-utils1.1.0~rc1+git20111210-7.4 chromium recommends no packages. Versions of packages chromium suggests: ii chromium-inspector 46.0.2490.71-1~deb8u1unix pn chromium-l10n -- no debconf information
Bug#784868: korganizer: korgac grabs focus loosing keys and missing notifications
Package: korganizer Version: 4:4.14.1-1nograb Severity: important Tags: patch upstream Dear Maintainer, *** Reporter, please consider answering these questions, where appropriate *** * What led up to the situation? * What exactly did you do (or not do) that was effective (or ineffective)? * What was the outcome of this action? * What outcome did you expect instead? *** End of the template - remove these template lines *** While it can make it easier to dismiss (if it grabs focus through the activateWindow call), it also means if you are typing and a reminder comes in it will both get those keystrokes and a space or return will dismiss the reminder before you have a chance to ever see what the reminder was about, which assumes you are even looking at your screen to see the reminder box flash up, if you are typing from paper you might not even know. This patch will call raiseWindow instead of activateWindow, but adds a dock menu item to grab focus if someone actually wants that behavior. While I was add it I added a dock menu item to show the reminder window, which was useful in debugging, and lets you get at items that have been suspended, alloing to view or dismiss them. This might satisfy "Bug 302865 - Inaccessible reminders" There is a code comment "Try to keep the dialog small and non-obtrusive." I've made it smaller by default, and set the minimum size even smaller than reasonable, but this is better than the previous which wouldn't let you size it smaller than the default. I added more points to save the position and improve the logic, now it will save and restore the position and size, and then show the window so it will no longer jump. Patch is against Debian 4.14.1 source tree. The korgac in the previous Debian release, 4.4.11 does not grab the keyboard when the notification dialog appears. Reproducible: Always Steps to Reproduce: 1. set korganizer reminder 2. start typing a paragraph 3. continue typing as the notification goes off Actual Results: watch korgac reminder appear grab keyboard focus and vanish at the next space, depending on the timing you have no chance to stop typing before it is gone Expected Results: The notifier window to appear at the last location it was dismissed at and only get focus when the window manager directs focus to it. I'm flagging the severity at important, because it was so disruptive I had stopped using it until I fixed it. >From bb8f84deeefaaf74bc285fc2534111a78a5f12ff Mon Sep 17 00:00:00 2001 From: David Fries Date: Sat, 9 May 2015 10:48:10 -0500 Subject: [PATCH 1/5] add an option to not grab keyboard focus when a reminder is displayed While it can make it easier to dismiss (if it grabs focus through the activateWindow call), it also means if you are typing and a reminder comes in it will both get those keystrokes and a space or return will dismiss the reminder before you have a chance to ever see what the reminder was about, which assumes you are even looking at your screen to see the reminder box flash up, if you are typing in a paper you might not even know. --- korgac/alarmdialog.cpp | 12 +++- korgac/alarmdialog.h | 1 + korgac/alarmdockwindow.cpp | 15 +++ korgac/alarmdockwindow.h | 7 +++ 4 files changed, 34 insertions(+), 1 deletion(-) diff --git a/korgac/alarmdialog.cpp b/korgac/alarmdialog.cpp index 6bbd867..59c5790 100644 --- a/korgac/alarmdialog.cpp +++ b/korgac/alarmdialog.cpp @@ -567,7 +567,11 @@ void AlarmDialog::show() KWindowSystem::unminimizeWindow( winId(), false ); KWindowSystem::setState( winId(), NET::KeepAbove | NET::DemandsAttention ); KWindowSystem::setOnAllDesktops( winId(), true ); - KWindowSystem::activateWindow( winId() ); + if ( grabFocus( ) ) { +KWindowSystem::activateWindow( winId() ); + } else { +KWindowSystem::raiseWindow( winId() ); + } // Audio, Procedure, and EMail alarms eventNotification(); @@ -1036,3 +1040,9 @@ void AlarmDialog::removeFromConfig( const QList &ids ) genGroup.sync(); } +bool AlarmDialog::grabFocus( ) +{ + KSharedConfig::Ptr config = KGlobal::config(); + KConfigGroup generalConfig( config, "General" ); + return generalConfig.readEntry( "GrabFocus", false ); +} diff --git a/korgac/alarmdialog.h b/korgac/alarmdialog.h index 41acee6..8f298d4 100644 --- a/korgac/alarmdialog.h +++ b/korgac/alarmdialog.h @@ -123,6 +123,7 @@ class AlarmDialog : public KDialog void updateButtons(); void toggleDetails( QTreeWidgetItem *item ); void showDetails( QTreeWidgetItem *item ); +static bool grabFocus( ); Akonadi::ETMCalendar::Ptr mCalendar; QTreeWidget *mIncidenceTree; diff --git a/korgac/alarmdockwindow.cpp b/korgac/alarmdockwindow.cpp index 35343be..8fe6ecf 100644 --- a/korgac/alarmdockwindow.cpp +++ b/korgac/alarmdockwindow.cpp @@ -45,6 +45,7 @@ AlarmDockWindow::AlarmDockWindow() KConfigGroup config( KGloba
Bug#780003: tk8.6: antialiasing in wish8.5/8.6 make gitk unreadable dark background
Package: tk8.6 Version: 8.6.2-1 Severity: normal Dear Maintainer, Debian jessie no longer packages tk8.4, and 8.5/8.6 versions of tk enable antialiasing some of the time, which is causing text to be nearly unreadable especially with a dark background. This is with gitk and a dark background using the courier font for the diff area. The font is the default font selected with gitk. I took screen shots of the same gitk script running on the same system with three versions of tk 8.4, 8.5, and 8.6 with my normal dark configuration and the stock light option. https://plus.google.com/114119938972260211379/posts/7FH11KdgDCq With the dark background the newer versions of tk take 28% more (vertical) space to display the same text, and yet for the dark background it's very very difficult to read, like it is blurred out to the point it is blurry and blends with the background color. The white background takes the same additional amount of space, but I would say it's a matter of opinion if it is more or less readable, I would say less, but not overly so. Here is my .gitk config if it matters. Note depending on the gitk version some colors are not ideal for the dark theme, this bug report is about font display, not gitk color selection. .gitk set mainfont {Helvetica 8} set textfont {Courier 8} set uifont {Helvetica 9 bold} set tabstop 8 set findmergefiles 0 set maxgraphpct 50 set maxwidth 16 set cmitmode patch set wrapcomment none set autoselect 0 set autosellen 40 set showneartags 1 set hideremotes 0 set showlocalchanges 1 set datetimeformat {%Y-%m-%d %H:%M:%S} set limitdiffs 1 set uicolor grey85 set uifgcolor black set want_ttk 1 set override_ttk 0 set bgcolor black set fgcolor white set colors {green red lightblue magenta darkgrey lightgray orange} set diffcolors {red green cyan} set markbgcolor #404060 set diffcontext 3 set selectbgcolor gray50 set filebgcolor #aa set extdifftool meld set perfile_attrs 0 set geometry(main) 700x488+0+230 set geometry(state) normal set geometry(topwidth) 696 set geometry(topheight) 160 set geometry(pwsash0) "372 2" set geometry(pwsash1) "574 2" set geometry(botwidth) 503 set geometry(botheight) 308 set permviews {} -- System Information: Debian Release: 8.0 APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 3.19.0+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Init: unable to detect Versions of packages tk8.6 depends on: ii libc6 2.19-15 ii libfontconfig1 2.11.0-6.3 ii libfreetype62.5.2-3 ii libtcl8.6 8.6.2+dfsg-1 ii libtk8.68.6.2-1 ii libx11-62:1.6.2-3 ii libxext62:1.3.3-1 ii libxft2 2.3.2-1 ii libxss1 1:1.2.2-1 ii zlib1g 1:1.2.8.dfsg-2+b1 Versions of packages tk8.6 recommends: ii konsole [x-terminal-emulator] 4:4.14.2-1 ii xterm [x-terminal-emulator]312-2 tk8.6 suggests no packages. -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#774401: avrdude direct SPI mode NULL pointer crash and USBtiny fix
Package: avrdude Version: 5.11.1-1 Severity: normal Dear Maintainer, Pick a non-bit bang avr programmer, enter terminal mode, then give the "spi" command to enter direct SPI mdoe. avrdude crashes. Direct SPI mode allows sending and receiving data over SPI while the chip is running. This patch adds a check to avoid the crash and implements what is required for USBtiny (not a bit bang device) to do direct SPI mode, provided slave select is held low. Attached are patches for 5 and 6. Here's the upstream bug report. https://savannah.nongnu.org/bugs/index.php?43912 avrdude 5.x fix --- >From 5eb948ba4fe6b8694ce52c842efd2a00c81c2cf5 Mon Sep 17 00:00:00 2001 From: David Fries Date: Wed, 31 Dec 2014 20:55:57 -0600 Subject: [PATCH 1/3] fix NULL pointer crash entering direct spi mode If setpin isn't defined entering direct spi mode will crash, check before entering that mode. --- term.c |8 1 file changed, 8 insertions(+) diff --git a/term.c b/term.c index 9df8c49..be0afa3 100644 --- a/term.c +++ b/term.c @@ -747,6 +747,11 @@ static int cmd_help(PROGRAMMER * pgm, struct avrpart * p, static int cmd_spi(PROGRAMMER * pgm, struct avrpart * p, int argc, char * argv[]) { + if(!pgm->setpin || !pgm->pinno[PIN_AVR_RESET]) { +fprintf(stdout, "The %s programmer does not support direct SPI mode.\n", + pgm->type); +return -1; + } pgm->setpin(pgm, pgm->pinno[PIN_AVR_RESET], 1); spi_mode = 1; return 0; @@ -755,6 +760,9 @@ static int cmd_spi(PROGRAMMER * pgm, struct avrpart * p, static int cmd_pgm(PROGRAMMER * pgm, struct avrpart * p, int argc, char * argv[]) { + if(!pgm->setpin || !pgm->pinno[PIN_AVR_RESET]) { +return -1; + } pgm->setpin(pgm, pgm->pinno[PIN_AVR_RESET], 0); spi_mode = 0; pgm->initialize(pgm, p); -- 1.7.10.4 >From 417b8405319937570cce07f2faea4deb0eb7e1dc Mon Sep 17 00:00:00 2001 From: David Fries Date: Thu, 1 Jan 2015 18:12:29 -0600 Subject: [PATCH 2/3] set back to pgm mode in quit Otherwise it can't read the fuses and the safe fuse code complains. --- term.c |4 1 file changed, 4 insertions(+) diff --git a/term.c b/term.c index be0afa3..1ea3784 100644 --- a/term.c +++ b/term.c @@ -554,6 +554,10 @@ static int cmd_sig(PROGRAMMER * pgm, struct avrpart * p, static int cmd_quit(PROGRAMMER * pgm, struct avrpart * p, int argc, char * argv[]) { + /* FUSE bit verify will fail if left in SPI mode */ + if(spi_mode == 1) { +cmd_pgm(pgm, p, 0, NULL); + } return 1; } -- 1.7.10.4 >From 9c989ef324b82dc8f77ac1f460f08cccbbb50d18 Mon Sep 17 00:00:00 2001 From: David Fries Date: Wed, 31 Dec 2014 21:18:53 -0600 Subject: [PATCH 3/3] add direct spi mode to USBtiny --- usbtiny.c | 39 +++ 1 file changed, 39 insertions(+) diff --git a/usbtiny.c b/usbtiny.c index 5646757..886c53e 100644 --- a/usbtiny.c +++ b/usbtiny.c @@ -353,6 +353,20 @@ static int usbtiny_initialize (PROGRAMMER *pgm, AVRPART *p ) return 0; } +static int usbtiny_setpin(struct programmer_t * pgm, int pin, int value) +{ + /* USBtiny is not a bit bang device, but it can set RESET */ + if(pin == PIN_AVR_RESET) { +if (usb_control(pgm, USBTINY_POWERUP, + PDATA(pgm)->sck_period, value ? RESET_HIGH : RESET_LOW) < 0) { + return -1; +} +usleep(5); +return 0; + } + return -1; +} + /* Tell the USBtiny to release the output pins, etc */ static void usbtiny_powerdown(PROGRAMMER * pgm) { @@ -388,6 +402,26 @@ static int usbtiny_cmd(PROGRAMMER * pgm, unsigned char cmd[4], unsigned char res res[2] == cmd[1]); // AVR's do a delayed-echo thing } +static int usbtiny_spi(struct programmer_t * pgm, unsigned char cmd[], unsigned char res[], int count) +{ + int i; + + // Make sure its empty so we don't read previous calls if it fails + memset(res, '\0', count); + + if(count % 4) { +fprintf(stderr, "direct SPI write must be a multiple of 4 bytes for %s\n", + pgm->type); +return -1; + } + for(i=0; itype, "USBtiny"); + /* Just need to identify reset pin is requested */ + pgm->pinno[PIN_AVR_RESET] = PIN_AVR_RESET; + /* Mandatory Functions */ pgm->initialize = usbtiny_initialize; pgm->enable = usbtiny_enable; @@ -557,6 +594,8 @@ extern void usbtiny_initpgm ( PROGRAMMER* pgm ) pgm->set_sck_period = usbtiny_set_sck_period; pgm->setup= usbtiny_setup; pgm->teardown = usbtiny_teardown; + pgm->setpin = usbtiny_setpin; + pgm->spi = usbtiny_spi; } #else /* !HAVE_LIBUSB */ -- 1.7.10.4 avrdude 6 fix --- >From a68a0d190fc3acdc85600d41ca94f133fcb5eb75 Mon Sep 17 00:00:00 2001 From: David Fries Date: Thu, 1 Jan 2015 19:06:55 -0600 Subject: [PA
Bug#653320: pulseaudio: Assertion 'length % bq->base == 0' failed, kills daemon
On Sat, Apr 12, 2014 at 09:02:34PM -0300, Felipe Sateler wrote: > Hi David, > > On Mon, Dec 26, 2011 at 04:01:21PM -0600, David Fries wrote: > > Package: pulseaudio > > Version: 0.9.21-3+squeeze1 > > Severity: normal > > > > > > I found that pa_stream_write with an offset that's not a multiple of > > the base sample size causes pulseaudio to assert and abort. This is > > not the case for the nbytes parameter of how many bytes to write, that > > is allowed to not be a multiple of the sample size, but the seek > > offset parameter must be a multiple of the sample size or it asserts. > > > > E: memblockq.c: Assertion 'length % bq->base == 0' failed at > > pulsecore/memblockq.c:519, function pa_memblockq_drop(). Aborting. > > Aborted > > > > pulseaudio 1.0-4 gives the following, > > E: [alsa-sink] memblockq.c: Assertion 'length % bq->base == 0' failed > > at pulsecore/memblockq.c:600, function pa_memblockq_drop(). Aborting. > > > > I've fixed the program causing it, I just wanted to make others aware > > that any program that can play audio can take down the audio daemon. > > You reported this issue a long time ago, and the problem might have been > fixed since then. > > Can you still reproduce this problem? If so, please reply (preferably > with a short offending program) so we may debug this. Otherwise I'd like > to close this bug. > > Thanks > > > -- > Saludos, > Felipe Sateler Thanks for asking about this bug, I modified a test (attached) to demonstrate it is still a problem. There might be a shorter setup sequence, but it comes down to giving SEEK as something other than a multiple of the sample size. pa_stream_write(p, buffer, nbytes, NULL, SEEK, PA_SEEK_RELATIVE); pulseaudio 2.0-6.1 cc -g -Wall -g PulseSeekBug.c -lpulse -o PulseSeekBug && ./PulseSeekBug wrote 65472, offset 1 Stream state 3, suspended -15, corked -15, writable size -1 Stream state 3, suspended -15, corked -15, writable size -1 Stream state 3, suspended -15, corked -15, writable size -1 Stream state 3, suspended -15, corked -15, writable size -1 /usr/bin/pulseaudio W: [pulseaudio] pid.c: Stale PID file, overwriting. E: [alsa-sink] memblockq.c: Assertion 'length % bq->base == 0' failed at pulsecore/memblockq.c:613, function pa_memblockq_drop(). Aborting. Aborted -- David Fries PGP pub CB1EE8F0 http://fries.net/~david/ /* Test program using the asynchronous API. * This is demonstrating a problem where the playback under-runs, but * yet the stream lists a writable size of 0. Those are contracidtory, * if it is under-running it should have space available, if there isn't * any space available it shouldn't be under-running. Adding some * debugging to src/pulse/stream.c shows that the s->requested_bytes * is negative, but returns 0 on negative. * * At least this is the case as of PulseAudio 0.9.21, seems to be fixed * in the git code grabbed on 09-18-2010, I didn't look into when it * was fixed. */ #include #include #include #include // If it isn't a multiple of the sample size, 4, 8, 12, etc the pulseaudio // daemon asserts and dies. #define SEEK 1 void writable(pa_stream *p, size_t nbytes, void *userdata) { char *buffer; /* pa_stream_begin_write "If you place (size_t) -1 in *nbytes on * invocation the memory size will be chosen automatically (which is * recommended to do)." (from the OggPulseAudio.h) */ nbytes=(size_t)-1; pa_stream_begin_write(p, (void**)&buffer, &nbytes); // fill the buffer with audio here... memset(buffer, 0, nbytes); pa_stream_write(p, buffer, nbytes, NULL, SEEK, PA_SEEK_RELATIVE); printf("wrote %ld, offset %d\n", 1L*nbytes, SEEK); } int main(int argc, char **argv) { pa_threaded_mainloop *mainloop=NULL; pa_mainloop_api *api=NULL; pa_context *context=NULL; pa_stream *stream=NULL; pa_sample_spec spec; pa_context_state_t cstate; mainloop=pa_threaded_mainloop_new(); if(!mainloop) { fprintf(stderr, "Failed to create PulseAudio main loop\n"); return 1; } api=pa_threaded_mainloop_get_api(mainloop); pa_threaded_mainloop_start(mainloop); pa_threaded_mainloop_lock(mainloop); context=pa_context_new(api, __PRETTY_FUNCTION__); pa_context_connect(context, NULL, (pa_context_flags_t)0, NULL); pa_threaded_mainloop_unlock(mainloop); spec.format=PA_SAMPLE_S16LE; spec.rate=48000; spec.channels=2; pa_channel_map map; pa_channel_map_init_auto(&map, spec.channels, PA_CHANNEL_MAP_DEFAULT); for(;;) { pa_threaded_mainloop_lock(mainloop); cstate=pa_context_get_state(context); pa_threaded_mainloop_unlock(mainloop); if(cstate>=PA_CONTEXT_READY) break; usleep(1000); } pa_threaded_mainloop_lock(mainloop); stream=pa_stream_new(context, __fun
Bug#727633: tightvncserver: config allows arbitrary code execution
Package: tightvncserver Version: 1.3.9-6.4 Severity: normal Dear Maintainer, examples/vnc.conf.gz lists "It is perl syntax, but only variable assignment is allowed" Actually arbitrary code execution is allowed with a classic SQL ';' exploitation technique, which I figured out how to take advantage of to add a new option when starting the server. $tmp_local="ignore";push(@ARGV,'-dpi 96');print "after $Config_file, args; @ARGV\n"; Or something like this works. $tmp_local="ignore";exit(1); I'm not sure how much this matters because the file will either be an administrator owned file or the user owned file. Or the documentation could be updated. "The perl interpreter 'eval' is used and must be contain a variable assignment." Still there should be a better way to pass additional options to the vnc server from the config file. -- System Information: Debian Release: 7.1 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 3.9.0+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages tightvncserver depends on: ii libc6 2.13-38 ii libjpeg62 6b1-3 ii libx11-62:1.5.0-1+deb7u1 ii libxext62:1.3.1-2+deb7u1 ii perl5.14.2-21 ii x11-common 1:7.7+3~deb7u1 ii xbase-clients 1:7.7+3~deb7u1 ii xserver-common 2:1.12.4-6 ii zlib1g 1:1.2.7.dfsg-13 Versions of packages tightvncserver recommends: ii xfonts-base 1:1.0.3 Versions of packages tightvncserver suggests: pn tightvnc-java -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#727631: tightvncserver: need a way to add options for Xtightvnc from config file
Package: tightvncserver Version: 1.3.9-6.4 Severity: wishlist Dear Maintainer, The config startup script /usr/bin/vncserver seems to be setup for the config file ~/.vncrc to only replace existing options. It doesn't have a straight forward way to add a new option and the Xtightvnc takes many more options than the script has defaults for. I wanted to add "-dpi 96" without having to specify it on the command line every time and since there is a config file that would be a good place to put it. This was the best way I had to add another option, that is fine an existing item, and append to it. $authType.=" -dpi 96"; -- System Information: Debian Release: 7.1 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 3.9.0+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages tightvncserver depends on: ii libc6 2.13-38 ii libjpeg62 6b1-3 ii libx11-62:1.5.0-1+deb7u1 ii libxext62:1.3.1-2+deb7u1 ii perl5.14.2-21 ii x11-common 1:7.7+3~deb7u1 ii xbase-clients 1:7.7+3~deb7u1 ii xserver-common 2:1.12.4-6 ii zlib1g 1:1.2.7.dfsg-13 Versions of packages tightvncserver recommends: ii xfonts-base 1:1.0.3 Versions of packages tightvncserver suggests: pn tightvnc-java -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
debian-bugs-dist@lists.debian.org
On Mon, Mar 11, 2013 at 09:48:43AM +0100, Adam Reichold wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > Hello, > > This will have to wait until after the release of 0.4.1 since we are > currently in a string freeze. I've patched my system, so I'm not in a rush. Thanks for the quick response. I've been looking for a replacement for kpdf, and my the top missing features of qpdfview are being able to select text to copy out of the document, and having a search limiting the thumbnails to the pages with that term on it. > Could you also give a reference why you consider this standard? Maybe > an HIG or something like that? > > Best regards, Adam. It was an enlightening easter egg hunt to go on, but I haven't found anything standard sounding using E&xit. The closest I found where, Qt shortcut documentation using E&xit, but not specifically in a file menu, and given how inconsistent the official Qt programs are in using E&xit and &Quit, that isn't used as much of a standard. http://harmattan-dev.nokia.com/docs/library/html/qt4/qshortcut.html GNOME Human Interface Guidelines, Standard Menus, which uses &Quit https://developer.gnome.org/hig-book/3.4/menus-standard.html.en Ubuntu, &Quit, but it wasn't specifically file menu, https://wiki.ubuntu.com/CustomStatusMenuDesignGuidelines Doing a Google search on both E&xit and &Exit gives most &Exit result with using & instead of and, which isn't using it as a mnemonic. I found a few programming examples using E&xit, none using &Exit. The first ten pages of Google's search using &Exit gave one link using &Exit for the file menu, and the requester clearly didn't know what they were doing. http://support.realsoftware.com/listarchives/gettingstarted/2006-05/msg00174.html There was another reference using &Exit but it wasn't clear if it was a file menu, the most common &Exit result was my bug report. In another search there was this entry with &Exit for a button, but for the file menu it was E&xit, and I think &Exit was just to not conflict the mnemonic, of I assume the file menu, but it didn't go into details. http://root.cern.ch/download/doc/ROOTUsersGuideHTML/ch25s08.html I took a survey of all the programs in the Debian menu on my system that have file menus. I was surprised at how uncommon E&xit was, but I didn't find any other program using &Exit. programs with exit in their menu and &x for the mnemonic chromium, E&xit gvim E&xit xfig, E&xit Licq, E&xit Qt Linguist, E&xit ddd, E&xit Audacity, E&xit Qt Configuration, E&xit programs with exit in their menu that don't use &x stopwatch, Exit (no mnemonic) BOINC Manager, Exit (no mnemonic) programs that use something different Firefox, &quit konqueror, &quit plan, &quit pidgin, &quit k3b, &quit GOPchop, &quit hugin, &quit Image Magick, quit inkscape, &quit gimp, &quit Xcircuit, quit kdenlive, &Quit PTBatcherGUI, &Quit Xchat, &Quit Ekiga, &Quit KTorrent, &Quit wireshark, &Quit gnucash, &Quit Qt Assistant, &Quit Meld, &Quit Timidity, &Quit Sweep, Quit grecord, &Quit editres, Quit Zapping, &Quit xawtv, Quit VLC media player, Quit xine, Quit gxine, &Quit GV, Quit digikam, &Quit xdvik, Quit xditview, Quit > Am 11.03.2013 01:29, schrieb David Fries: > > Package: qpdfview Version: 0.3.1-1 Severity: normal Tags: patch > > > > Dear Maintainer, > > > > Please apply the following patch to correct the expected exit key > > sequence to the standard sequence for program exit which uses the x > > not e for the hotkey mnemonic. > > > >> From d05b57456ca48b7817e32bfa39a7eca0073decd4 Mon Sep 17 00:00:00 > >> 2001 > > From: David Fries Date: Sun, 10 Mar 2013 19:04:51 > > -0500 Subject: [PATCH] Fix non-standard mnemonic. > > > > For exiting the mnemonic is Alt-F-X not Alt-F-E, updated. --- > > sources/mainwindow.cpp |2 +- translations/qpdfview_de.ts | > > 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) > > > > diff --git a/sources/mainwindow.cpp b/sources/mainwindow.cpp index > > 4dc809b..633ddb9 100644 --- a/sources/mainwindow.cpp +++ > > b/sources/mainwindow.cpp @@ -1149,7 +1149,7 @@ void > > MainWindow::createActions() > > > > // exit > > > > -m_exitAction = new QAction(tr("&Exit"), this); + > > m_exitAction = new QAction(tr("E&xit"), this); > > m_exitAction->setShortcut(QKeySequence::Quit); > > m_exitAction->setIcon(QIcon::fromTheme("application-exit")); > > m_exitAction->setIconVisibleInMenu(true); diff --git > &
debian-bugs-dist@lists.debian.org
Package: qpdfview Version: 0.3.1-1 Severity: normal Tags: patch Dear Maintainer, Please apply the following patch to correct the expected exit key sequence to the standard sequence for program exit which uses the x not e for the hotkey mnemonic. >From d05b57456ca48b7817e32bfa39a7eca0073decd4 Mon Sep 17 00:00:00 2001 From: David Fries Date: Sun, 10 Mar 2013 19:04:51 -0500 Subject: [PATCH] Fix non-standard mnemonic. For exiting the mnemonic is Alt-F-X not Alt-F-E, updated. --- sources/mainwindow.cpp |2 +- translations/qpdfview_de.ts |2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/sources/mainwindow.cpp b/sources/mainwindow.cpp index 4dc809b..633ddb9 100644 --- a/sources/mainwindow.cpp +++ b/sources/mainwindow.cpp @@ -1149,7 +1149,7 @@ void MainWindow::createActions() // exit -m_exitAction = new QAction(tr("&Exit"), this); +m_exitAction = new QAction(tr("E&xit"), this); m_exitAction->setShortcut(QKeySequence::Quit); m_exitAction->setIcon(QIcon::fromTheme("application-exit")); m_exitAction->setIconVisibleInMenu(true); diff --git a/translations/qpdfview_de.ts b/translations/qpdfview_de.ts index e1c8505..0b1b461 100644 --- a/translations/qpdfview_de.ts +++ b/translations/qpdfview_de.ts @@ -260,7 +260,7 @@ -&Exit +E&xit &Beenden -- 1.7.10.4 -- System Information: Debian Release: 7.0 APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 3.8.0+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages qpdfview depends on: ii libc6 2.13-38 ii libcomerr21.42.5-1 ii libcups2 1.5.3-2.14 ii libgcc1 1:4.7.2-5 ii libgcrypt11 1.5.0-3 ii libgnutls26 2.12.20-4 ii libgssapi-krb5-2 1.10.1+dfsg-4 ii libk5crypto3 1.10.1+dfsg-4 ii libkrb5-3 1.10.1+dfsg-4 ii libpoppler-qt4-3 0.18.4-5 ii libqt4-dbus 4:4.8.2+dfsg-11 ii libqt4-svg4:4.8.2+dfsg-11 ii libqt4-xml4:4.8.2+dfsg-11 ii libqtcore44:4.8.2+dfsg-11 ii libqtgui4 4:4.8.2+dfsg-11 ii libstdc++64.7.2-5 ii zlib1g1:1.2.7.dfsg-13 qpdfview recommends no packages. qpdfview suggests no packages. -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#702747: korganizer: date parsing way too strict, 6:50PM should work
Package: korganizer Version: 4:4.4.11.1+l10n-3+b1 Severity: normal Dear Maintainer, korganizer date and time input prasing is too strict. Enter 6:50PM and it complains and suggests '06:44 PM', a leading 0 is not normally written by humans writing a time, so computers need to accept that, having a space before PM should not be required. This is with the default timeformat, "PH:MM:SS AMPM". -- System Information: Debian Release: 7.0 APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 3.8.0+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages korganizer depends on: ii kde-runtime 4:4.8.4-2 ii kdepim-runtime4:4.4.11.1-6 ii libakonadi-contact4 4:4.8.4-2 ii libc6 2.13-38 ii libkabc4 4:4.8.4-2 ii libkcal4 4:4.8.4-2 ii libkcmutils4 4:4.8.4-4 ii libkde3support4 4:4.8.4-4 ii libkdecore5 4:4.8.4-4 ii libkdepim44:4.4.11.1+l10n-3+b1 ii libkdeui5 4:4.8.4-4 ii libkholidays4 4:4.8.4-2 ii libkio5 4:4.8.4-4 ii libkmime4 4:4.8.4-2 ii libknewstuff2-4 4:4.8.4-4 ii libkontactinterface4 4:4.8.4-2 ii libkparts44:4.8.4-4 ii libkpimidentities44:4.8.4-2 ii libkpimutils4 4:4.8.4-2 ii libkprintutils4 4:4.8.4-4 ii libkresources44:4.8.4-2 ii libphonon44:4.6.0.0-3 ii libqt4-dbus 4:4.8.2+dfsg-11 ii libqt4-qt3support 4:4.8.2+dfsg-11 ii libqt4-xml4:4.8.2+dfsg-11 ii libqtcore44:4.8.2+dfsg-11 ii libqtgui4 4:4.8.2+dfsg-11 ii libstdc++64.7.2-5 ii perl 5.14.2-18 ii phonon4:4.6.0.0-3 ii zlib1g1:1.2.7.dfsg-13 korganizer recommends no packages. Versions of packages korganizer suggests: pn kdepim-groupware ii kdepim-kresources 4:4.4.11.1+l10n-3+b1 -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#683723: please include in Wheezy
Package: xserver-xorg-input-evdev Version: 1:2.7.0-1+b1 Followup-For: Bug #683723 Dear Maintainer, > Source: xserver-xorg-input-evdev > Source-Version: 1:2.7.1-1 > > We believe that the bug you reported is fixed in the latest version of > xserver-xorg-input-evdev, which is due to be installed in the Debian > FTP archive. That was August 2012, and now six months later it hasn't been promoted to wheezy or sid. I'm concerned if it isn't promoted it will miss Wheezy entirely and people needing multitouch will need to go to a different distribution entirely. -- System Information: Debian Release: 7.0 APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 3.8.0+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages xserver-xorg-input-evdev depends on: ii libc6 2.13-38 ii xserver-xorg-core [xorg-input-abi-16] 2:1.12.4-5 xserver-xorg-input-evdev recommends no packages. xserver-xorg-input-evdev suggests no packages. -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#702662: xserver-xorg-core: multitouch stuck button, apply upstream patch
Package: xserver-xorg-core Version: 2:1.12.4-5 Severity: important Dear Maintainer, Please apply the following two multitouch bug fixes to xserver-xorg-core for Wheezy. Currently an event from a multitouch device will be translated into a pointer event including simulating pressing the first mouse button, the problem is the button down state is never released, meaning from then on all mouse events will have a state of button one held, causing programs to misbehave in important ways. This persists until X is restarted, even unplugging the touchscreen device doesn't clear it. 3e6358ee6c33979329b78fe2097a1fdf76fb69cd 'Xi: Don't check for TOUCH_END, it's never set' 863f32c930d71073ee5f78452b78bd459d024867 'Xi: Update the device after delivering the emulated pointer event(#56558)' Interestingly enough when the synaptics driver gets an input from a touch pad it will reset the held button state, otherwise the only option is to restart X and not touch the touch pad. -- System Information: Debian Release: 7.0 APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 3.8.0+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages xserver-xorg-core depends on: ii keyboard-configuration 1.88 ii libaudit0 1:1.7.18-1.1 ii libc6 2.13-38 ii libdrm2 2.4.40-1~deb7u2 ii libgcrypt11 1.5.0-3 ii libpciaccess0 0.13.1-2 ii libpixman-1-0 0.26.0-4 ii libselinux1 2.1.9-5 ii libudev0175-7.1 ii libxau6 1:1.0.7-1 ii libxdmcp6 1:1.1.1-1 ii libxfont1 1:1.4.5-2 ii udev175-7.1 ii xserver-common 2:1.12.4-5 Versions of packages xserver-xorg-core recommends: ii libgl1-mesa-dri 8.0.5-3 Versions of packages xserver-xorg-core suggests: ii xfonts-100dpi1:1.0.3 ii xfonts-75dpi 1:1.0.3 ii xfonts-scalable 1:1.0.3-1 -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#701220: libdbus-1-dev: c++11 syntax error, trivial fix, add whitespace
Package: libdbus-1-dev Version: 1.2.24-4+squeeze1 Severity: important Tags: patch There is a syntax error when including dbus-protocol.h when compiling with a c++11 compiler such as gcc 4.7 or newer. see "User-defined literals and whitespace" http://gcc.gnu.org/gcc-4.7/porting_to.html "To fix, just add some whitespace between the string literal and the macro name." I made the following patch and now it works. --- /root/tmp/dbus-protocol.h 2013-02-22 19:16:30.267044120 -0600 +++ dbus-protocol.h 2013-02-22 19:17:33.276643990 -0600 @@ -425,7 +425,7 @@ extern "C" { /** XML system identifier of the introspection format version 1.0 */ #define DBUS_INTROSPECT_1_0_XML_SYSTEM_IDENTIFIER "http://www.freedesktop.org/standards/dbus/1.0/introspect.dtd"; /** XML document type declaration of the introspection format version 1.0 */ -#define DBUS_INTROSPECT_1_0_XML_DOCTYPE_DECL_NODE "\n" +#define DBUS_INTROSPECT_1_0_XML_DOCTYPE_DECL_NODE "\n" /** @} */ -- System Information: Debian Release: 6.0.6 APT prefers stable APT policy: (500, 'stable') Architecture: i386 (x86_64) Kernel: Linux 3.8.0+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages libdbus-1-dev depends on: ii libdbus-1-31.2.24-4+squeeze1 simple interprocess messaging syst ii pkg-config 0.25-1.1 manage compile and link flags for libdbus-1-dev recommends no packages. libdbus-1-dev suggests no packages. -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#700097: obnam should not pass excessive options to ssh
> i2013-02-08 17:59:56 DEBUG executing openssh: ['ssh', > '-oForwardX11=no', '-oForwardAgent=no', '-oClearAllForwardings=yes', > '-oProtocol=2', '-p', '22', '-l', 'root', '-s', '-o', > 'UserKnownHostsFile=/root/.ssh/known_hosts', 'mahler', 'sftp'] > > The problem with this is that it makes it impossible to set such > things as the username and port via ~/.ssh/config. I consider this a > bug because it is highly unexpected behaviour. Michael, I also thought it was strange when I saw it doing that as well. I think it is reasonable and a good idea to set/disable the first options -oForwardX11=no -oForwardAgent=no -oClearAllForwardings=yes -oProtocol=2 any reason you can think of these being enabled? The user and port I agree with, and here's a patch to let ssh pick the default if none was given. What about the known hosts file? Should that also fall back to ssh for the default? --- NEWS|2 ++ obnamlib/plugins/sftp_plugin.py | 24 +++- 2 files changed, 17 insertions(+), 9 deletions(-) diff --git a/NEWS b/NEWS index 24f79da..2252841 100644 --- a/NEWS +++ b/NEWS @@ -24,6 +24,8 @@ Version X.Y, released UNRELEASED * SSHChannelAdapter now reports the number of bytes send and received (before ssh encryption), from David Fries * new option 'ssh-options' to pass additional options to ssh from David Fries +* Only set ssh user/port if explicitily requested, otherwise let ssh + select them. Reported by Michael Goetze, fixed by David Fries. Bug fixes: diff --git a/obnamlib/plugins/sftp_plugin.py b/obnamlib/plugins/sftp_plugin.py index f57c577..e715937 100644 --- a/obnamlib/plugins/sftp_plugin.py +++ b/obnamlib/plugins/sftp_plugin.py @@ -41,9 +41,6 @@ with warnings.catch_warnings(): import obnamlib -DEFAULT_SSH_PORT = 22 - - def ioerror_to_oserror(method): '''Decorator to convert an IOError exception to OSError. @@ -169,9 +166,12 @@ class SftpFS(obnamlib.VirtualFileSystem): args = ['ssh', '-oForwardX11=no', '-oForwardAgent=no', '-oClearAllForwardings=yes', '-oProtocol=2', -'-p', str(self.port), -'-l', self.user, '-s'] +# default user/port from ssh (could be a per host configuration) +if self.port: +args += ['-p', str(self.port)] +if self.user: +args += ['-l', self.user] if self.settings: if self.settings['ssh-key']: args += ['-i', self.settings['ssh-key']] @@ -202,7 +202,11 @@ class SftpFS(obnamlib.VirtualFileSystem): def _connect_paramiko(self): logging.debug('connect_paramiko: host=%s port=%s' % (self.host, self.port)) -self.transport = paramiko.Transport((self.host, self.port)) +if self.port: +remote = (self.host, self.port) +else: +remote = (self.host) +self.transport = paramiko.Transport(remote) self.transport.connect() logging.debug('connect_paramiko: connected') try: @@ -249,6 +253,8 @@ class SftpFS(obnamlib.VirtualFileSystem): logging.debug('Host key for %s OK' % hostname) def _authenticate(self, username): +if not username: +username = self._get_username() for key in self._find_auth_keys(): try: self.transport.auth_publickey(username, key) @@ -298,12 +304,12 @@ class SftpFS(obnamlib.VirtualFileSystem): if '@' in netloc: user, netloc = netloc.split('@', 1) else: -user = self._get_username() +user = None if ':' in netloc: host, port = netloc.split(':', 1) if port == '': -port = DEFAULT_SSH_PORT +port = None else: try: port = int(port) @@ -314,7 +320,7 @@ class SftpFS(obnamlib.VirtualFileSystem): raise morphlib.Error(msg) else: host = netloc -port = DEFAULT_SSH_PORT +port = None if path.startswith('/~/'): path = path[3:] -- 1.7.10.4 -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#678826: [PATCH] add option 'ssh-options' to add ssh command line options
Reported-By: David Fries --- Joey Hess [bug reporter], Here's a patch to the current obnam bzr trunk for being able to pass additional options to ssh such as -4 or -6, without obnam having to add those options specifically. I don't know how easily it would apply to the current Debian 1.1 version. --ssh-options='-v -4' I added this because I've been using rsync and it supports limiting the bandwidth. For over Internet backups I don't want to saturation the link degrading performance when it's a backup that can run for hours. sftp also has a similar option, but I didn't realize until I added it that since obnam is calling ssh directly that option isn't available. NEWS|1 + obnamlib/plugins/sftp_plugin.py | 25 + 2 files changed, 18 insertions(+), 8 deletions(-) diff --git a/NEWS b/NEWS index 00e5506..24f79da 100644 --- a/NEWS +++ b/NEWS @@ -23,6 +23,7 @@ Version X.Y, released UNRELEASED to what `obnam backup` does. Suggested by "S. B.". * SSHChannelAdapter now reports the number of bytes send and received (before ssh encryption), from David Fries +* new option 'ssh-options' to pass additional options to ssh from David Fries Bug fixes: diff --git a/obnamlib/plugins/sftp_plugin.py b/obnamlib/plugins/sftp_plugin.py index 2049b22..d144105 100644 --- a/obnamlib/plugins/sftp_plugin.py +++ b/obnamlib/plugins/sftp_plugin.py @@ -172,14 +172,18 @@ class SftpFS(obnamlib.VirtualFileSystem): '-p', str(self.port), '-l', self.user, '-s'] -if self.settings and self.settings['ssh-key']: -args += ['-i', self.settings['ssh-key']] -if self.settings and self.settings['strict-ssh-host-keys']: -args += ['-o', 'StrictHostKeyChecking=yes'] -if self.settings and self.settings['ssh-known-hosts']: -args += ['-o', - 'UserKnownHostsFile=%s' % -self.settings['ssh-known-hosts']] +if self.settings: +if self.settings['ssh-key']: +args += ['-i', self.settings['ssh-key']] +if self.settings['strict-ssh-host-keys']: +args += ['-o', 'StrictHostKeyChecking=yes'] +if self.settings['ssh-known-hosts']: +args += ['-o', + 'UserKnownHostsFile=%s' % +self.settings['ssh-known-hosts']] +if self.settings['ssh-options']: +for option in self.settings['ssh-options'].split(): +args += [option] args += [self.host, 'sftp'] logging.debug('executing openssh: %s' % args) @@ -615,6 +619,11 @@ class SftpPlugin(obnamlib.ObnamPlugin): os.path.expanduser('~/.ssh/known_hosts'), group=ssh_group) +self.app.settings.string(['ssh-options'], + 'Options added to the ssh command line, such ' + 'as -4 -6 -o .', + group=ssh_group) + self.app.settings.boolean(['pure-paramiko'], 'do not use openssh even if available, ' 'use paramiko only instead', -- 1.7.10.4 -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#671646: qemu-kvm: [PATCH] better parse usb alt_interfaces
On Tue, Jul 03, 2012 at 06:21:53PM +0400, Michael Tokarev wrote: > (Replying to somewhat old bugreport) > > 12.05.2012 04:04, David Fries wrote: > > On Fri, May 11, 2012 at 10:59:38AM +0400, Michael Tokarev wrote: > >> On 05.05.2012 19:07, David Fries wrote: > >>> Package: qemu-kvm > >>> Version: 1.0+dfsg-12 > >>> Severity: normal > >>> Tags: patch > >>> > >>> Dear Maintainer, > >>> > >>> qemu-kvm stops parsing the USB descriptor table when it sees an > >>> alternative interface entry that doesn't match the currently set guest > >>> request preventing the guest from communicating to any device > >>> using an alternative interface other than zero. More details in the > >>> quilt patch header. > >> > >> You're attaching an upstream commit > >> 96dd9aac37d30f3425088f81523942e67b2d03ac > >> from Gerd Hoffmann which rewrites usb_linux_update_endp_table function. > >> It has been included into current qemu git tree, and will be in the > >> next qemu release, 1.1, which is about to be out. So I'll hold it > >> for the time being, it will be fixed in the next release. > > > > The patch I included was mine, I didn't mean to imply anything > > different. Gerd Hoffmann does have an independent fix, but there were > > enough changes between the git qemu tree and the current Debian > > qemu-kvm that I wasn't able to back port it, so I wrote my own. > > Oh. I misunderstand at the time. Please excuse me for that, and > thank you for your work! > > Now, as I described previously, my main attention was to get 1.1 > release for wheezy. That happened (well, sort of anyway). > > > Just the same if a kvm-qemu release for Debian will be coming out > > shortly and also have his fix then that works too. > > Does current version in sid (1.1-z0+dfsg-1) has the issue you > mentioned? If no, can I close this bugreport now? Package: qemu-kvm Version: 1.1.0+dfsg-1 Depends: libaio1 (>= 0.3.93), kvm: symbol lookup error: kvm: undefined symbol: rbd_aio_discard Preparing to replace librbd1 0.43-1 (using .../librbd1_0.47.2-1_amd64.deb) I don't know which version first included the rbd_aio_discard symbol, but kvm is now running. The updated USB is working with my device. You can close the bugreport. -- David Fries PGP pub CB1EE8F0 http://fries.net/~david/ -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#671646: qemu-kvm: [PATCH] better parse usb alt_interfaces
On Fri, May 11, 2012 at 10:59:38AM +0400, Michael Tokarev wrote: > On 05.05.2012 19:07, David Fries wrote: > > Package: qemu-kvm > > Version: 1.0+dfsg-12 > > Severity: normal > > Tags: patch > > > > Dear Maintainer, > > > > qemu-kvm stops parsing the USB descriptor table when it sees an > > alternative interface entry that doesn't match the currently set guest > > request preventing the guest from communicating to any device > > using an alternative interface other than zero. More details in the > > quilt patch header. > > You're attaching an upstream commit 96dd9aac37d30f3425088f81523942e67b2d03ac > from Gerd Hoffmann which rewrites usb_linux_update_endp_table function. > It has been included into current qemu git tree, and will be in the > next qemu release, 1.1, which is about to be out. So I'll hold it > for the time being, it will be fixed in the next release. The patch I included was mine, I didn't mean to imply anything different. Gerd Hoffmann does have an independent fix, but there were enough changes between the git qemu tree and the current Debian qemu-kvm that I wasn't able to back port it, so I wrote my own. Just the same if a kvm-qemu release for Debian will be coming out shortly and also have his fix then that works too. -- David Fries PGP pub CB1EE8F0 http://fries.net/~david/ -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#671646: qemu-kvm: [PATCH] better parse usb alt_interfaces
Package: qemu-kvm Version: 1.0+dfsg-12 Severity: normal Tags: patch Dear Maintainer, qemu-kvm stops parsing the USB descriptor table when it sees an alternative interface entry that doesn't match the currently set guest request preventing the guest from communicating to any device using an alternative interface other than zero. More details in the quilt patch header. >From f17082f0d29b02310af09ed81b6af5d2a3a914f9 Mon Sep 17 00:00:00 2001 From: David Fries Date: Thu, 26 Apr 2012 19:50:57 -0500 Subject: [PATCH] patch to better parse usb alt_interfaces This is needed until the upstream qemu fixes trickle down. --- debian/patches/series |1 + debian/patches/usb_alt_int_fixup.patch | 108 2 files changed, 109 insertions(+), 0 deletions(-) create mode 100644 debian/patches/usb_alt_int_fixup.patch diff --git a/debian/patches/series b/debian/patches/series index 5c0b5b4..dddc264 100644 --- a/debian/patches/series +++ b/debian/patches/series @@ -9,3 +9,4 @@ run-pod2man-with-utf8-option-to-enable-utf8-in-manpa.patch slirp:-signal-free-input-buffer-space-to-io-thread.patch pci-fix-corrupted-pci-conf-index-register-by-unaligned-write.patch disable-hpet-for-tcg.patch +usb_alt_int_fixup.patch diff --git a/debian/patches/usb_alt_int_fixup.patch b/debian/patches/usb_alt_int_fixup.patch new file mode 100644 index 000..919a761 --- /dev/null +++ b/debian/patches/usb_alt_int_fixup.patch @@ -0,0 +1,108 @@ +patch to better parse usb alt_interfaces + +If a device has multiple alternative interfaces updated_endp_table +breaks with "invalid descriptor data", because if it sees an +interface that doesn't match alt_interface it continues to the +top of the loop which is expecting a config descritor not an +endpoint descriptor. If the guest tries to talk to any +descriptor that wasn't parsed (or tries to set alt_interface +greater than 0), is_valid fails and it is denied. + +My question is, why does qemu need to validate the endpoint? Why not +just pass it to the host to validate? + +This is fixed upstream in qemu with the following commit, but that +commit isn't yet in qemu-kvm, and enough of the usb system changed +that it isn't trivial to backport. + +commit 96dd9aac37d30f3425088f81523942e67b2d03ac +Author: Gerd Hoffmann +Date: Thu Mar 29 16:06:28 2012 +0200 +usb-host: rewrite usb_linux_update_endp_table + +-- +David Fries + +--- qemu-kvm.orig/usb-linux.c qemu-kvm/usb-linux.c +@@ -1153,48 +1153,29 @@ + i = 0; + + while (i < length) { +-if (descriptors[i + 1] != USB_DT_CONFIG) { +-fprintf(stderr, "invalid descriptor data\n"); +-return 1; +-} else if (descriptors[i + 5] != s->configuration) { +-DPRINTF("not requested configuration %d\n", s->configuration); +-i += (descriptors[i + 3] << 8) + descriptors[i + 2]; +-continue; +-} +- +-i += descriptors[i]; +- +-if (descriptors[i + 1] != USB_DT_INTERFACE || +-(descriptors[i + 1] == USB_DT_INTERFACE && +- descriptors[i + 4] == 0)) { +-i += descriptors[i]; +-continue; +-} +- +-interface = descriptors[i + 2]; +-alt_interface = usb_linux_get_alt_setting(s, s->configuration, +- interface); +- +-/* the current interface descriptor is the active interface +- * and has endpoints */ +-if (descriptors[i + 3] != alt_interface) { +-i += descriptors[i]; +-continue; +-} +- +-/* advance to the endpoints */ +-while (i < length && descriptors[i +1] != USB_DT_ENDPOINT) { +-i += descriptors[i]; +-} +- +-if (i >= length) +-break; +- +-while (i < length) { +-if (descriptors[i + 1] != USB_DT_ENDPOINT) { +-break; ++if (descriptors[i + 1] == USB_DT_CONFIG) { ++if (descriptors[i + 5] != s->configuration) { ++DPRINTF("not requested configuration %d\n", s->configuration); ++/* advance to the next configuration */ ++i += (descriptors[i + 3] << 8) + descriptors[i + 2]; ++} else { ++i += descriptors[i]; + } +- ++} else if (descriptors[i + 1] == USB_DT_INTERFACE) { ++interface = descriptors[i + 2]; ++alt_interface = usb_linux_get_alt_setting(s, s->configuration, ++ interface); ++if (descriptors[i + 3] != alt_interface) { ++/* it isn't the current alt setting, skip it */ ++do { ++i += descriptors[i]; ++} while (i < length && ++
Bug#409349: usbhid: control queue full; hung apcupsd task
On Sat, 22 Jan 2011 19:39:24 + Steven Chamberlain wrote: > Usually I see the stream of the "control queue full" errors first, but > this time apcupsd hung very soon after boot: > INFO: task apcupsd:21604 blocked for more than 120 seconds. > [] ? ohci_urb_dequeue+0xd9/0xe9 [ohci_hcd] > [] ? usb_kill_urb+0x9d/0xbb [usbcore] > [] ? autoremove_wake_function+0x0/0x2e > [] ? usbhid_init_reports+0x8c/0xee > [] ? hiddev_ioctl+0x2bf/0x632 [usbhid] > This was despite raising apcupsd's POLLTIME from 10 to 60 seconds. My usbmon dump showed that requests were being made that the UPS wasn't returning or failing, sounds like a task for a bus analyzer and a read through the USB standard. usbhid_init_report is called from usbhid_start and from hid.c "called on probe to start the device", and as such wouldn't be called each time apcupsd polls the UPS. So increasing the POLLTIME wouldn't have helped. I don't know how many reports it is requesting, but the timeout in usbhid_wait_io is 10 seconds, so 12 could add up to 120 seconds. Even so I would expect it to eventually timeout. Now if it was sitting in ohci_urb_dequeue for 120 seconds that is a problem. On Wed, Apr 04, 2012 at 10:43:42PM +0100, Steven Chamberlain wrote: > found 409349 linux-2.6/2.6.32-41 > thanks > > On 04/04/12 22:23, Jonathan Nieder wrote: > > Ping. Is nobody trying to use an APC UPS connected by USB any more? :) I have three. I only had problems with one OHCI SiS rev 07 PCI 1039:7001 controller, and once I added the patch for "introduce timeout for stuck ctrl/out URBs" I've not had the issue anymore. > Yep! This week on 2.6.32-41 I noticed apcupsd hang after about a day, > followed by the 'control queue full' errors an hour later. That was > with the patches suggested in #631287 applied, just in case they made > any difference to this. My guess is the control queue message full problem just enlarges the window of opportunity fro 'BUG during access to hiddev' bug report #631287. I think it would help avoid crashing when it gets into this situation, but from the patch titles I don't think it would help avoid it. The queue is 256 long, and with apcupsd querying every 10 seconds, 256*10/60=42 minutes once an entry gets blocked. With the timeout set to 5 seconds in the "introduce timeout" patch and apcups polling every 10 seconds it no matter if every request failed to complete it shouldn't be able to fill up. But that's only if apcupsd is only submitting one control request every 10 seconds, but looking at my old logs that isn't the case. For my UPS it seemed to only ignore a transfer rarely, maybe there are some that are frequent enough to be a problem, in which case the timeout patch wouldn't be able to keep up. It would be nice to have a usbmon dump of it happening to see if that is the case or not. > > What kernels work and don't work? > > I already tagged affected versions in the BTS; the first report of this > was on 2.6.17 and I myself recall seeing it as far back as 2.6.26. > > > I hadn't been able to test the UPS on a 3.y kernel until this week. > Right now I have the UPS hooked up to a new server running 3.2.0-2-amd64 > (Debian 3.2.12-1) to see if it recurs. What UPS do you have that's the problem? Mine was, BackUPS Pro 500 firmware 16.3.D USB FW:4 > Then in a few weeks, I'll be taking the original hardware (on which the > issue was easily reproducible) out of production use, so I can try 3.y > kernels on that too. (Until now, I couldn't try 3.y kernels on that box > as it needs to run OpenVZ containers). What kind of timeframe is easily reproducible? Can you run information gathering patches on the box, if so I'll put something together. Then again this has been going on for years, if you prefer to wait a few weeks until it is not in production use that's fine. Back when I was first debugging this I added some printk messages for when the queue length increased. If you'll give them a try I'll look them up and also printout that jiffies value as well. -- David Fries PGP pub CB1EE8F0 http://fries.net/~david/ -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#651370: libgl1-mesa-glx: need close on exec for dri device
Updating the Debian bug report log: The patch is now in master git://anongit.freedesktop.org/git/mesa/mesa b60120608f6ddf4098bc324363197c979ee04cb7 Author: David Fries 2011-12-10 11:28:45 Committer: Dave Airlie 2012-03-11 14:05:50 -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#650626: bash tabcompletion broken in 4.2
bash tabcompletion broken in 4.2, I know this is an upstream problem, but I think it deserves a higher importance as you can no longer tab complete anything with an environment variable in it, if you do that path is broken and requires manual editing to fix, and it used to work. ls $HOME/tm 4.1.5 gives ls /home/david/tmp/ and the directory listing on enter, 4.2.20 gives ls \$HOME/tmp ls: cannot access $HOME/tmp: No such file or directory What I do more frequently is get the absolute path of a file in the current directory to give to another program. ls $PWD/.bashrc was ls /home/david/.bashrc now ls \$PWD/.bashrc I'm confused on this post and proposed direxpand option fix. The way I'm reading this bash 4.2 will be broken by default hitting tab on a valid path (if you press enter), $HOME/tmp/ will be replaced by an invalid path \$HOME/tmp/ and that is a bug, adding an option to cover that up doesn't change it. http://lists.gnu.org/archive/html/bug-bash/2011-09/msg7.html -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#653320: pulseaudio: Assertion 'length % bq->base == 0' failed, kills daemon
Package: pulseaudio Version: 0.9.21-3+squeeze1 Severity: normal I found that pa_stream_write with an offset that's not a multiple of the base sample size causes pulseaudio to assert and abort. This is not the case for the nbytes parameter of how many bytes to write, that is allowed to not be a multiple of the sample size, but the seek offset parameter must be a multiple of the sample size or it asserts. E: memblockq.c: Assertion 'length % bq->base == 0' failed at pulsecore/memblockq.c:519, function pa_memblockq_drop(). Aborting. Aborted pulseaudio 1.0-4 gives the following, E: [alsa-sink] memblockq.c: Assertion 'length % bq->base == 0' failed at pulsecore/memblockq.c:600, function pa_memblockq_drop(). Aborting. I've fixed the program causing it, I just wanted to make others aware that any program that can play audio can take down the audio daemon. -- System Information: Debian Release: 6.0.3 APT prefers stable APT policy: (500, 'stable') Architecture: i386 (x86_64) Kernel: Linux 3.2.0-rc5+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages pulseaudio depends on: ii adduser3.112+nmu2add and remove users and groups ii consolekit 0.4.1-4 framework for defining and trackin ii libasound2 1.0.23-2.1shared library for ALSA applicatio ii libc6 2.11.2-10 Embedded GNU C Library: Shared lib ii libcap21:2.19-3 support for getting/setting POSIX. ii libdbus-1-31.2.24-4+squeeze1 simple interprocess messaging syst ii libgdbm3 1.8.3-9 GNU dbm database routines (runtime ii libice62:1.0.6-2 X11 Inter-Client Exchange library ii libltdl7 2.2.6b-2 A system independent dlopen wrappe ii libpulse0 0.9.21-3+squeeze1 PulseAudio client libraries ii libsamplerate0 0.1.7-3 Audio sample rate conversion libra ii libsm6 2:1.1.1-1 X11 Session Management library ii libsndfile11.0.21-3+squeeze1 Library for reading/writing audio ii libspeexdsp1 1.2~rc1-1 The Speex extended runtime library ii libudev0 164-3 libudev shared library ii libx11-6 2:1.3.3-4 X11 client-side library ii libxtst6 2:1.1.0-3 X11 Testing -- Record extension li ii lsb-base 3.2-23.2squeeze1 Linux Standard Base 3.2 init scrip ii udev 164-3 /dev/ and hotplug management daemo Versions of packages pulseaudio recommends: pn gstreamer0.10-pulseaudio (no description available) pn libasound2-plugins (no description available) pn pulseaudio-esound-compat (no description available) pn pulseaudio-module-x11 (no description available) Versions of packages pulseaudio suggests: pn paman (no description available) pn paprefs(no description available) ii pavucontrol0.9.9-1 PulseAudio Volume Control pn pavumeter (no description available) ii pulseaudio-utils 0.9.21-3+squeeze1 Command line tools for the PulseAu -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#651652: bridge-utils: Enable ipv6 when finished (complement of #607741)
Package: bridge-utils Version: 1.5-2 Severity: normal Tags: patch This is in response to the following bug report, Disable ipv6 on the interfaces of the bridge. Closes: #607741. It sounds like a good idea, but it left me really confused as to what happened to my IPv6 link local address and autoconfiguration on eth0 after I took down the bridge and connected to a different network. Here's a patch to enable IPv6 again when the bridge is taken down. --- ifupdown.sh.orig2011-12-10 15:23:05.0 -0600 +++ ifupdown.sh 2011-12-10 15:22:47.0 -0600 @@ -81,6 +81,10 @@ do if [ -x /etc/network/if-post-down.d/vlan ]; then env IFACE=$port /etc/network/if-post-down.d/vlan fi + if [ -f /proc/sys/net/ipv6/conf/$port/disable_ipv6 ] + then +echo 0 > /proc/sys/net/ipv6/conf/$port/disable_ipv6 + fi fi done done -- System Information: Debian Release: wheezy/sid APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 3.2.0-rc4+ (SMP w/2 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages bridge-utils depends on: ii libc6 2.13-21 bridge-utils recommends no packages. Versions of packages bridge-utils suggests: ii ifupdown 0.7~alpha5+really0.6.16 -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#651370: libgl1-mesa-glx: need close on exec for dri device
;fd == -1 && errno == EINVAL) +#endif + { + drmdpy->fd = open(drmdpy->device_name, O_RDWR); + if (drmdpy->fd != -1) + fcntl(drmdpy->fd, F_SETFD, fcntl(drmdpy->fd, F_GETFD) | FD_CLOEXEC); + } if (drmdpy->fd == -1) { _eglLog(_EGL_WARNING, "wayland-egl: could not open %s (%s)", drmdpy->device_name, strerror(errno)); diff --git a/src/gallium/state_trackers/egl/x11/x11_screen.c b/src/gallium/state_trackers/egl/x11/x11_screen.c index 6155b4d..140668f 100644 --- a/src/gallium/state_trackers/egl/x11/x11_screen.c +++ b/src/gallium/state_trackers/egl/x11/x11_screen.c @@ -265,7 +265,15 @@ x11_screen_enable_dri2(struct x11_screen *xscr, if (!x11_screen_probe_dri2(xscr, NULL, NULL)) return -1; - fd = open(xscr->dri_device, O_RDWR); +#ifdef O_CLOEXEC + fd = open(xscr->dri_device, O_RDWR | O_CLOEXEC); + if (fd == -1 && errno == EINVAL) +#endif + { + fd = open(xscr->dri_device, O_RDWR); + if (fd != -1) +fcntl(fd, F_SETFD, fcntl(fd, F_GETFD) | FD_CLOEXEC); + } if (fd < 0) { _eglLog(_EGL_WARNING, "failed to open %s", xscr->dri_device); return -1; diff --git a/src/glx/dri2_glx.c b/src/glx/dri2_glx.c index 553869a..b931a7b 100644 --- a/src/glx/dri2_glx.c +++ b/src/glx/dri2_glx.c @@ -914,7 +914,15 @@ dri2CreateScreen(int screen, struct glx_display * priv) goto handle_error; } - psc->fd = open(deviceName, O_RDWR); +#ifdef O_CLOEXEC + psc->fd = open(deviceName, O_RDWR | O_CLOEXEC); + if (psc->fd == -1 && errno == EINVAL) +#endif + { + psc->fd = open(deviceName, O_RDWR); + if (psc->fd != -1) + fcntl(psc->fd, F_SETFD, fcntl(psc->fd, F_GETFD) | FD_CLOEXEC); + } if (psc->fd < 0) { ErrorMessageF("failed to open drm device: %s\n", strerror(errno)); goto handle_error; -- 1.7.7.1 On Thu, Dec 08, 2011 at 08:37:46PM +0100, Julien Cristau wrote: > On Wed, Dec 7, 2011 at 23:43:13 -0600, David Fries wrote: > > > diff -upr /tmp/mesa-7.11/src/egl/drivers/dri2/platform_wayland.c > > mesa-7.11/src/egl/drivers/dri2/platform_wayland.c > > --- /tmp/mesa-7.11/src/egl/drivers/dri2/platform_wayland.c 2011-07-08 > > 20:37:09.0 -0500 > > +++ mesa-7.11/src/egl/drivers/dri2/platform_wayland.c 2011-12-07 > > 21:28:10.0 -0600 > > @@ -734,17 +734,25 @@ drm_handle_device(void *data, struct wl_ > > { > > struct dri2_egl_display *dri2_dpy = data; > > drm_magic_t magic; > > +#ifdef O_CLOEXEC > > + int flags = O_RDWR | O_CLOEXEC; > > +#else > > + int flags = O_RDWR; > > +#endif > > > > dri2_dpy->device_name = strdup(device); > > if (!dri2_dpy->device_name) > >return; > > > > - dri2_dpy->fd = open(dri2_dpy->device_name, O_RDWR); > > + dri2_dpy->fd = open(dri2_dpy->device_name, flags); > > if (dri2_dpy->fd == -1) { > >_eglLog(_EGL_WARNING, "wayland-egl: could not open %s (%s)", > > dri2_dpy->device_name, strerror(errno)); > >return; > > } > > +#ifndef O_CLOEXEC > > + fcntl(dri2_dpy->fd, F_SETFD, fcntl(dri2_dpy->fd, F_GETFD) | FD_CLOEXEC); > > +#endif > > > > I'd do something like > > #ifdef O_CLOEXEC >dri2_dpy->fd = open(dri2_dpy->device_name, O_RDWR | O_CLOEXEC); >if (dri2_dpy->fd == -1 && errno == EINVAL) > #endif >{ > dri2_dpy->fd = open(dri2_dpy->device_name, O_RDWR); > if (dri2_dpy->fd >= 0) > fcntl(dri2_dpy->fd, F_SETFD, fcntl(dri2_dpy->fd, F_GETFD) | > FD_CLOEXEC); >} >if (dri2_dpy->fd == -1) { >_eglLog(...); >return; > > so it still works if the build environment has O_CLOEXEC but the kernel > doesn't support it at runtime. > > Cheers, > Julien mesa_cloexec_retry.bundle Description: Binary data
Bug#647577: ticgit: use + for truncated lines (or detect UTF-8)
Package: ticgit Version: 1.0.2.9-1 Severity: normal *** Please type your report below this line *** ticgit-ng/cli.rb is blindly outputting characters that don't work on all terminals sub_value = "#{value[0, size - 1]}\xe2\x80\xa6" replace it with a single character that does, sub_value = "#{value[0, size - 1]}+" If it's going to use the UTF-8 escape sequence it needs to first check $LANG or something to verify UTF-8 is supported otherwise it shows up as garbage. --- /tmp/cli.rb 2011-11-03 10:08:08.927671832 -0500 +++ /usr/lib/ruby/vendor_ruby/ticgit-ng/cli.rb 2011-11-03 10:06:04.431915459 -0500 @@ -201,7 +201,7 @@ module TicGitNG value = value.to_s if value.bytesize > size - sub_value = "#{value[0, size - 1]}\xe2\x80\xa6" + sub_value = "#{value[0, size - 1]}+" else sub_value = value[0, size] end @@ -213,7 +213,7 @@ module TicGitNG chars = value.to_s.scan(/./um) if chars.size > size - sub_value = "#{chars[0, size-1]}\xe2\x80\xa6" + sub_value = "#{chars[0, size-1]}+" else sub_value = chars.join end -- System Information: Debian Release: 6.0 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Kernel: Linux 2.6.32-5-amd64 (SMP w/8 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages ticgit depends on: ii ruby-git 1.2.5-2 Ruby interface to the Git revision ii ruby1.8 1.8.7.302-2 Interpreter of object-oriented scr Versions of packages ticgit recommends: ii git [git-core] 1:1.7.2.3-2.2 fast, scalable, distributed revisi ii git-core 1:1.7.2.3-2.2 fast, scalable, distributed revisi Versions of packages ticgit suggests: pn ticgitweb (no description available) -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#620557: procps: vmstat truncate so 32 bit programs get meaningful data
Package: procps Version: 1:3.2.8-9 Severity: normal Tags: upstream patch >From 64822112bb271a2195f3df3f9a54498d04005597 Mon Sep 17 00:00:00 2001 From: David Fries Date: Sat, 2 Apr 2011 12:39:11 -0500 Subject: [PATCH 2/2] vmstat truncate so 32 bit programs get meaningful data Truncate 64 bit to 32 bit when reading /proc/vmstat in 32 bit programs. 32 bit kernels would have truncated the values already, but 64 bit kernels don't, so once a variable goes over 32 bits existing programs only get a max unsigned long value. Truncate it so they would get the same values they always used to on 32 bit kernels. 64 bit programs continue to get the full value. It was either this, modify the API to always use 64 bit values, or add another function that would pass the 64 bit values in 32 bit programs. --- proc/sysinfo.c | 11 +++ 1 files changed, 11 insertions(+), 0 deletions(-) diff --git a/proc/sysinfo.c b/proc/sysinfo.c index 6b40825..744dc47 100644 --- a/proc/sysinfo.c +++ b/proc/sysinfo.c @@ -745,6 +745,9 @@ void vminfo(void){ {"slabs_scanned", &vm_slabs_scanned}, }; const int vm_table_count = sizeof(vm_table)/sizeof(vm_table_struct); +#if __SIZEOF_LONG__ == 4 + unsigned long long slotll; +#endif vm_pgalloc = 0; vm_pgrefill = 0; @@ -768,7 +771,15 @@ void vminfo(void){ ); head = tail+1; if(!found) goto nextline; +#if __SIZEOF_LONG__ == 4 +// A 32 bit kernel would have already truncated the value, a 64 bit kernel +// doesn't need to. Truncate here to let 32 bit programs to continue to get +// truncated values. It's that or change the API for a larger data type. +slotll = strtoull(head,&tail,10); +*(found->slot) = (unsigned long)slotll; +#else *(found->slot) = strtoul(head,&tail,10); +#endif nextline: //if(found) fprintf(stderr,"%s=%d\n",found->name,*(found->slot)); -- 1.7.2.3 >From c5a8218729754095cd6c8d9573f6e16123da190d Mon Sep 17 00:00:00 2001 From: David Fries Date: Sat, 2 Apr 2011 12:05:52 -0500 Subject: [PATCH 1/2] sysinfo.c correct sscanf argument Was taking the address of the array, but should have just been the array pointer without taking the address. --- proc/sysinfo.c |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/proc/sysinfo.c b/proc/sysinfo.c index 65d94d7..6b40825 100644 --- a/proc/sysinfo.c +++ b/proc/sysinfo.c @@ -838,7 +838,7 @@ unsigned int getdiskstat(struct disk_stat **disks, struct partition_stat **parti break; } fields = sscanf(buff, " %*d %*d %15s %*u %*u %*u %*u %*u %*u %*u %*u %*u %*u %u", -&devname, &dummy); +devname, &dummy); if (fields == 2 && is_disk(devname)){ (*disks) = realloc(*disks, (cDisk+1)*sizeof(struct disk_stat)); sscanf(buff, " %*d%*d %15s %u %u %llu %u %u %u %llu %u %u %u %u", -- 1.7.2.3 -- System Information: Debian Release: 6.0 APT prefers stable APT policy: (500, 'stable') Architecture: i386 (x86_64) Kernel: Linux 2.6.36-rc7+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages procps depends on: ii initscripts 2.88dsf-13.1 scripts for initializing and shutt ii libc6 2.11.2-10Embedded GNU C Library: Shared lib ii libncurses5 5.7+20100313-5 shared libraries for terminal hand ii libncursesw55.7+20100313-5 shared libraries for terminal hand ii lsb-base3.2-23.2squeeze1 Linux Standard Base 3.2 init scrip Versions of packages procps recommends: ii psmisc22.11-1utilities that use the proc file s procps suggests no packages. -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#619398: libpulse0: license issue, compile client LGPL
Package: libpulse0 Version: 0.9.21-3 Severity: normal Tags: upstream patch As PulseAudio is trying to be the Linux desktop audio system, it is important that the client library support programs which aren't GPL compatible as well. "Hence you should treat the client library ('libpulse') of PulseAudio as being LGPL licensed" but pulseaudio by default uses libgdbm-dev, and that is GPL licensed. Compile with libtdb-dev which is LGPL to keep the PulseAudio client library as LGPL. --- debian/control |2 +- debian/rules |3 +++ 2 files changed, 4 insertions(+), 1 deletions(-) diff --git a/debian/control b/debian/control index 61c0f50..fde2d88 100644 --- a/debian/control +++ b/debian/control @@ -15,7 +15,7 @@ Build-Depends: debhelper (>= 5), cdbs, quilt, m4, libltdl-dev (>= 2.2.6a-2), libasyncns-dev, libatomic-ops-dev, libspeexdsp-dev (>= 1.2~rc1), libbluetooth-dev (>= 4.40) [!kfreebsd-i386 !kfreebsd-amd64 !hurd-i386], -libgdbm-dev, intltool, libgtk2.0-dev, libxtst-dev, +libtdb-dev, intltool, libgtk2.0-dev, libxtst-dev, libssl-dev Standards-Version: 3.8.1 Vcs-Git: git://git.debian.org/git/pkg-pulseaudio/pulseaudio.git diff --git a/debian/rules b/debian/rules index 4c0387f..2f33a41 100755 --- a/debian/rules +++ b/debian/rules @@ -54,6 +54,9 @@ ifneq (,$(findstring $(DEB_HOST_ARCH_OS), "linux")) DEB_CONFIGURE_EXTRA_FLAGS = --enable-hal-compat --disable-hal endif +# Use Trivial Database (tdb) to keep the client library GNU LGPL +DEB_CONFIGURE_EXTRA_FLAGS += --with-database=tdb +DEB_CONFIGURE_EXTRA_FLAGS += --enable-hal-compat --disable-hal DEB_DH_MAKESHLIBS_ARGS_ALL = --exclude=usr/lib/pulse-0.9.21/modules DEB_DH_SHLIBDEPS_ARGS_libpulse0 = -u -L/dev/null DEB_DH_SHLIBDEPS_ARGS_pulseaudio = -u -Ldebian/shlibs_pulseaudio.local -- 1.7.2.3 -- System Information: Debian Release: 6.0 APT prefers stable APT policy: (500, 'stable') Architecture: i386 (x86_64) Kernel: Linux 2.6.36-rc7+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages libpulse0 depends on: ii libasyncns0 0.3-1.1Asyncronous name service query lib ii libc6 2.11.2-10 Embedded GNU C Library: Shared lib ii libcap2 1:2.19-3 support for getting/setting POSIX. ii libdbus-1-3 1.2.24-4 simple interprocess messaging syst ii libgdbm3 1.8.3-9GNU dbm database routines (runtime ii libice6 2:1.0.6-2 X11 Inter-Client Exchange library ii libsm62:1.1.1-1 X11 Session Management library ii libsndfile1 1.0.21-3 Library for reading/writing audio ii libwrap0 7.6.q-19 Wietse Venema's TCP wrappers libra ii libx11-6 2:1.3.3-4 X11 client-side library ii libxtst6 2:1.1.0-3 X11 Testing -- Record extension li libpulse0 recommends no packages. Versions of packages libpulse0 suggests: ii pulseaudio0.9.21-3 PulseAudio sound server -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#607835: kmail compose, spell check language
On Sun, Dec 26, 2010 at 02:42:23PM +0200, Modestas Vainius wrote: > Hello, > > On ketvirtadienis 23 Gruodis 2010 03:32:48 David Fries wrote: > > Now which package should I file a bug report for defaulting that value > > to the first alphabetically installed language instead of the system > > settings language? > > Sorry but "first alphabetically installed language" as default makes > absolutely no sense. System language is at least some kind of criteria which > is based on user preferences expressed while configuring the system > previously. But what is alphabet? It says nothing about user preferences. I agree. Should I open a separate bug report on kmail or where should I go from here? Before I explicitely set the kmail identities dictionary setting it was on the first of the available dictionaries which was 'Basque (Spain)', then after removing the package hunspell-eu-es it went to the next availale language which was English, but not the system default 'English (United States of America)', but Philippines I think. Reinstall hunspell-eu-es and it went back to 'Basque (Spain)'. -- David Fries http://fries.net/~david/ (PGP encryption key available) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#607835: kmail compose, spell check language
Package: kmail Version: 4:4.4.6-1 Severity: normal Click new message, then Settings, Spellchecker It displays, Default language: Basque (Spain) Change to 'English (United States of America)' Apply, but the setting isn't saved and it goes back to Spain on the next e-mail. This is with hunspell-eu-es and myspell-en-us installed. Removing hunspell-eu-es changes it to default to 'English (Philippines)' The System Settings spell check has Default language: English (United States of America) So, why does kmail compose not use that? -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: i386 (x86_64) Kernel: Linux 2.6.36-rc7+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages kmail depends on: ii kdebase-runtime 4:4.4.5-1runtime components from the offici ii kdepim-runtime 4:4.4.6-1Runtime components for akonadi-kde ii kdepimlibs-kio-plug 4:4.4.5-1kio slaves used by KDE PIM applica ii libakonadi-contact4 4:4.4.5-1library for using the Akonadi PIM ii libakonadi-kde4 4:4.4.5-1library for using the Akonadi PIM ii libc6 2.11.2-6+squeeze1Embedded GNU C Library: Shared lib ii libgcc1 1:4.4.5-4GCC support library ii libgpgme++2 4:4.4.5-1c++ wrapper library for gpgme ii libkabc44:4.4.5-1library for handling address book ii libkcal44:4.4.5-1library for handling calendar data ii libkde3support4 4:4.4.5-1the KDE 3 Support Library for the ii libkdecore5 4:4.4.5-1the KDE Platform Core Library ii libkdepim4 4:4.4.6-1KDE PIM library ii libkdeui5 4:4.4.5-1the KDE Platform User Interface Li ii libkhtml5 4:4.4.5-1the KHTML Web Content Rendering En ii libkimap4 4:4.4.5-1library for handling IMAP data ii libkio5 4:4.4.5-1the Network-enabled File Managemen ii libkldap4 4:4.4.5-1library for accessing LDAP ii libkleo44:4.4.6-1certificate based crypto library ii libkmime4 4:4.4.5-1library for handling MIME data ii libknotifyconfig4 4:4.4.5-1library for configuring KDE Notifi ii libkontactinterface 4:4.4.5-1Kontact interface library ii libkparts4 4:4.4.5-1the Framework for the KDE Platform ii libkpgp44:4.4.6-1gpg based crypto library ii libkpimidentities4 4:4.4.5-1library for managing user identiti ii libkpimtextedit44:4.4.5-1library that provides a textedit w ii libkpimutils4 4:4.4.5-1library for dealing with email add ii libkresources4 4:4.4.5-1the KDE Resource framework library ii libksieve4 4:4.4.6-1mail/news message filtering librar ii libktnef4 4:4.4.5-1library for handling TNEF data ii libkutils4 4:4.4.5-1various utility classes for the KD ii libmailtransport4 4:4.4.5-1mail transport service library ii libmessagecore4 4:4.4.6-1message core library ii libmessagelist4 4:4.4.6-1message list library ii libmimelib4 4:4.4.6-1MIME library ii libnepomuk4 4:4.4.5-1the Nepomuk Meta Data Library ii libphonon4 4:4.6.0really4.4.2-1 the core library of the Phonon mul ii libqt4-dbus 4:4.6.3-4Qt 4 D-Bus module ii libqt4-network 4:4.6.3-4Qt 4 network module ii libqt4-qt3support 4:4.6.3-4Qt 3 compatibility library for Qt ii libqt4-xml 4:4.6.3-4Qt 4 XML module ii libqtcore4 4:4.6.3-4Qt 4 core module ii libqtgui4 4:4.6.3-4Qt 4 GUI module ii libstdc++6 4.4.5-4 The GNU Standard C++ Library v3 ii libthreadweaver44:4.4.5-1the ThreadWeaver Library for the K ii perl5.10.1-15Larry Wall's Practical Extraction ii phonon 4:4.6.0really4.4.2-1 metapackage for the Phonon multime Versions of packages kmail recommends: pn gnupg-agent(no description available) pn gnupg2 (no description available) pn pinentry-qt4 | pinentry-x11(no description available) Versions of packages kmail suggests: ii bogofilter1.2.2-2a fast Bayesian spam filter (dummy pn clamav | f-prot-installer (no description available) ii kaddressbook 4:4.4.6-1 address book and contact data mana pn kleopatra (no description available) ii procmail 3.22-19Versat
Bug#603129: git-email: git-send-email why bother prompting on Cc adds? (need edit option)
Package: git-email Version: 1:1.7.2.3-2 Severity: normal From: David Fries To: g...@vger.kernel.org Cc: David Fries , David Fries ... The Cc list above has been expanded by additional addresses found in the patch commit message. By default send-email prompts before sending whenever this occurs. This behavior is controlled by the sendemail.confirm configuration setting. ... Send this email? ([y]es|[n]o|[q]uit|[a]ll): It's nice that it stops to let me know that it added to the Cc list, but it would be better if it would also let me do something about it. The options come down to send or don't send the message, where's the option to remove the addional Cc's or event edit the Cc list? As a side note I don't think I need to be included in the Cc list since I'm the one sending, and I'm sure I don't need to be listed twice. -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: i386 (x86_64) Kernel: Linux 2.6.36-rc7+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages git-email depends on: ii git 1:1.7.2.3-2 fast, scalable, distributed revisi Versions of packages git-email recommends: ii libauthen-sasl-perl 2.1500-1 Authen::SASL - SASL Authentication ii libemail-valid-perl 0.184-1Perl module for checking the valid ii libnet-smtp-ssl-perl 1.01-2 SSL support for Net::SMTP Versions of packages git-email suggests: ii git-doc 1:1.7.2.3-2 fast, scalable, distributed revisi -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#602772: ekiga, remote can unmute audio and unsuspend video
Package: ekiga Version: 3.2.7-2 Severity: normal Start an ekiga video conference, tell ekiga to suspend audio and video output (other side gets silence and no more video as expected), but when the remote suspends or resumes their video, the local side starts to send audio and video without any indication to the user. This can allow the remote side to eavsdrop thinks they are muted and not sending video. In my case the remote client is the N900 Telepathy-SofiaSIP/0.6.2 client. >From ekiga's point of view when only the video was suspended, received v=0 o=- s=- t=0 0 m=audio 7078 RTP/AVP 18 96 8 0 99 c=IN IP4 X a=rtpmap:18 G729/8000 a=rtpmap:96 ILBC/8000 a=fmtp:96 mode=30 a=rtpmap:8 PCMA/8000 a=rtpmap:0 PCMU/8000 a=rtpmap:99 telephone-event/8000 a=fmtp:99 0-15 m=video 9078 RTP/AVP 97 c=IN IP4 X a=rtpmap:97 h264/9 a=fmtp:97 sprop-parameter-sets==XX==;packetization-mode=1;profile-level-id= a=recvonly Is ekiga seeing the "a=recvonly" and doing what it was instructed to do, that is send video? -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: i386 (x86_64) Kernel: Linux 2.6.36-rc7+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages ekiga depends on: ii evolution-data-server 2.30.3-2 evolution database backend server ii gconf2 2.28.1-5 GNOME configuration database syste ii libatk1.0-01.30.0-1 The ATK accessibility toolkit ii libavahi-client3 0.6.27-2 Avahi client library ii libavahi-common3 0.6.27-2 Avahi common library ii libavahi-glib1 0.6.27-2 Avahi glib integration library ii libc6 2.11.2-6+squeeze1 Embedded GNU C Library: Shared lib ii libdbus-glib-1-2 0.88-2simple interprocess messaging syst ii libebook1.2-9 2.30.3-2 Client library for evolution addre ii libedataserver1.2-13 2.30.3-2 Utility library for evolution data ii libgcc11:4.4.5-4 GCC support library ii libgconf2-42.28.1-5 GNOME configuration database syste ii libglib2.0-0 2.24.2-1 The GLib library of C routines ii libgtk2.0-02.20.1-2 The GTK+ graphical user interface ii libldap-2.4-2 2.4.23-6 OpenLDAP libraries ii libnotify1 [libnotify1 0.5.0-2 sends desktop notifications to a n ii libopal3.6.8 3.6.8~dfsg-2 Open Phone Abstraction Library - s ii libpango1.0-0 1.28.3-1 Layout and rendering of internatio ii libpt2.6.7 2.6.7-1 Portable Tools Library ii libsasl2-2 2.1.23.dfsg1-6Cyrus SASL - authentication abstra ii libsigc++-2.0-0c2a 2.2.4.2-1 type-safe Signal Framework for C++ ii libstdc++6 4.4.5-4 The GNU Standard C++ Library v3 ii libx11-6 2:1.3.3-3 X11 client-side library ii libxext6 2:1.1.2-1 X11 miscellaneous extension librar ii libxml22.7.7.dfsg-4 GNOME XML library ii libxv1 2:1.0.5-1 X11 Video extension library Versions of packages ekiga recommends: ii gvfs 1.6.4-2 userspace virtual filesystem - ser ii yelp 2.30.1+webkit-1 Help browser for GNOME Versions of packages ekiga suggests: pn asterisk (no description available) pn gnugk (no description available) pn mediaproxy (no description available) pn rtpproxy (no description available) pn ser(no description available) pn siproxd(no description available) pn yate (no description available) -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#602764: ekiga: listens to wrong ip
Package: ekiga Version: 3.2.7-2 Severity: important Tags: patch Ekiga/opal initially listens to each individual ip:port on the system. When a call is started it is picking the first non-localhost ip address and binding to that one, which is a very bad guess as to which IP the remote system is going to send RPT packets to. This means if the system has multiple ip addresses it either works, or all the incoming UDP packets get rejected with an ICMP error because ekiga is listening to a different ip and port. Capture ekiga -d 4, and search for "Initial interface list:" for the interface order. Also, on startup not only does ekiga bind to each ip address it sends out the initial subscribe SIP requests from each to the remote server. In my case only one ip address is on the network which has a default route, the other networks are in the 192.168.0.0 and 172.20.0.0 ranges, which is causing packets to be routed to internet with local non-routable ip addresses. If it isn't going to look at the routing tables to figure out which local ip addresses can get to the route system, it should just call connect, let the kernel assign the local ip address, and read what that local ip address is. Until some more sane algorithm is picked, here is a hack listen to all of them. I'm not going to call this the right solution, but it does let me hear the other end. diff --git a/src/opal/rtpconn.cxx b/src/opal/rtpconn.cxx index 499f58b..ae007d9 100644 --- a/src/opal/rtpconn.cxx +++ b/src/opal/rtpconn.cxx @@ -138,6 +138,11 @@ RTP_Session * OpalRTPConnection::CreateSession(const OpalTransport & transport, WORD firstPort = manager.GetRtpIpPortPair(); WORD nextPort = firstPort; + // Ekiga/opal is picking the first non-localhost ip address and binding to + // that one, which is a very bad guess as to which ip the remote system is + // going to connect with. Until some more sane algorithm is picked, listen + // to all of them. + localAddress=PIPSocket::GetDefaultIpAny(); while (!rtpSession->Open(localAddress, nextPort, nextPort, manager.GetRtpIpTypeofService(), natMethod, rtpqos)) { nextPort = manager.GetRtpIpPortPair(); if (nextPort == firstPort) { -- 1.7.2.3 -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: i386 (x86_64) Kernel: Linux 2.6.36-rc7+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages ekiga depends on: ii evolution-data-server 2.30.3-2 evolution database backend server ii gconf2 2.28.1-5 GNOME configuration database syste ii libatk1.0-01.30.0-1 The ATK accessibility toolkit ii libavahi-client3 0.6.27-2 Avahi client library ii libavahi-common3 0.6.27-2 Avahi common library ii libavahi-glib1 0.6.27-2 Avahi glib integration library ii libc6 2.11.2-6+squeeze1 Embedded GNU C Library: Shared lib ii libdbus-glib-1-2 0.88-2simple interprocess messaging syst ii libebook1.2-9 2.30.3-2 Client library for evolution addre ii libedataserver1.2-13 2.30.3-2 Utility library for evolution data ii libgcc11:4.4.5-4 GCC support library ii libgconf2-42.28.1-5 GNOME configuration database syste ii libglib2.0-0 2.24.2-1 The GLib library of C routines ii libgtk2.0-02.20.1-2 The GTK+ graphical user interface ii libldap-2.4-2 2.4.23-6 OpenLDAP libraries ii libnotify1 [libnotify1 0.5.0-2 sends desktop notifications to a n ii libopal3.6.8 3.6.8~dfsg-2 Open Phone Abstraction Library - s ii libpango1.0-0 1.28.3-1 Layout and rendering of internatio ii libpt2.6.7 2.6.7-1 Portable Tools Library ii libsasl2-2 2.1.23.dfsg1-6Cyrus SASL - authentication abstra ii libsigc++-2.0-0c2a 2.2.4.2-1 type-safe Signal Framework for C++ ii libstdc++6 4.4.5-4 The GNU Standard C++ Library v3 ii libx11-6 2:1.3.3-3 X11 client-side library ii libxext6 2:1.1.2-1 X11 miscellaneous extension librar ii libxml22.7.7.dfsg-4 GNOME XML library ii libxv1 2:1.0.5-1 X11 Video extension library Versions of packages ekiga recommends: ii gvfs 1.6.4-2 userspace virtual filesystem - ser ii yelp 2.30.1+webkit-1 Help browser for GNOME Versions of packages ekiga suggests: pn asterisk (no description available) pn gnugk (no description available) pn mediaproxy (no description available) pn rtpproxy (no description available) pn ser
Bug#602140: konqueror: Cookie policy editor's Delete button wipes out most of the policy entries
Any plans to include some version of the fix in squeeze? With 8 bug reports and 20 votes there are multiple people wanting it fixed. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#602140: konqueror: Cookie policy editor's Delete button wipes out most of the policy entries
I have been bitten by the same bug, and I've tracked it down after restoring the polcy file from backup. Konqueror is saving only the entries currently selected by the site policy search instead of all of them. Type something in the search like google, modify something in the dialog so apply is enabled, click apply, look at kcookiejarrc and only the current list is saved. Clear the site policy search (as long as you haven't exited the dialog all the entries are still there), press apply and all the entries are saved. Here's a patch. >From c1604cb669e16a8f08894a79935510351f68f47a Mon Sep 17 00:00:00 2001 From: David Fries Date: Tue, 2 Nov 2010 23:48:59 -0500 Subject: [PATCH] kcookiespolicies, save all cookie polcy entries Without this patch konqueror would only save the cookie polcy entries that were visible (the first was always saved). This is because QTreeWidget->itemBelow() "Returns the item visually below the given item." Hidden items are not visible, but clearly all items should be saved, so get the item count and save all top level items. --- apps/konqueror/settings/kio/kcookiespolicies.cpp |5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) diff --git a/apps/konqueror/settings/kio/kcookiespolicies.cpp b/apps/konqueror/settings/kio/kcookiespolicies.cpp index 17dfe10..d314af4 100644 --- a/apps/konqueror/settings/kio/kcookiespolicies.cpp +++ b/apps/konqueror/settings/kio/kcookiespolicies.cpp @@ -397,10 +397,11 @@ void KCookiesPolicies::save() QStringList domainConfig; QTreeWidgetItem *at = dlg->lvDomainPolicy->topLevelItem(0); - while( at ) + // QTreeWidget->itemBelow() skips hidden items, save all of them + for(int i=0, stop=dlg->lvDomainPolicy->topLevelItemCount(); ilvDomainPolicy->topLevelItem(i); domainConfig.append(QString::fromLatin1("%1:%2").arg(QString(tolerantToAce(at->text(0.arg(m_pDomainPolicy[at])); -at = dlg->lvDomainPolicy->itemBelow(at); } group.writeEntry("CookieDomainAdvice", domainConfig); -- 1.7.1 -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#601147: iputils-ping: Only emit an audible ping when requested (including flooding).
Package: iputils-ping Version: 3:20100418-2 Severity: normal Tags: patch upstream Previously flood pinging wouldn't beep at all. The upstream source changed in iputils-s20100214 to always beep without any way to disable it. A normal ping uses -a to enable beeping, -A is already used for an unrelated option so that can't be used as a diable beeping, so it makes sense to change the flood behavior to no beeping by default and -a will cause normal and flood pings to beep as the user did request it. That also makes both options be consistent. Patch included. I've already e-mailed the patch to YOSHIFUJI Hideaki, adding a bug report here for tracking. >From ef19fb7a7367748e19fa1634631e6a60620bb3d0 Mon Sep 17 00:00:00 2001 From: David Fries Date: Sat, 23 Oct 2010 13:30:39 -0500 Subject: [PATCH] Only emit an audible ping when requested (including flooding). The changeset 59de58d0f92a2c9b90174e51070aca32e7959e28 "iputils-s20100214" changed flooding to always make an audible ping when flooding, but didn't provide anyway to disable the audible ping when flooding. That's a problem when you want to see the results of the flood ping, but not hear it. This patch changes the behavior to match the normal ping case, no audible ping by default, and -a enables the ping. Signed-off-by: David Fries --- ping.c | 10 +- ping6.c | 10 +- 2 files changed, 10 insertions(+), 10 deletions(-) diff --git a/ping.c b/ping.c index 6b9a231..502d1e5 100644 --- a/ping.c +++ b/ping.c @@ -806,16 +806,16 @@ parse_reply(struct msghdr *msg, int cc, void *addr, struct timeval *tv) return 0; } + if (options & F_AUDIBLE) { + putchar('\a'); + if(options & F_FLOOD) + fflush(stdout); + } if (!(options & F_FLOOD)) { pr_options(buf + sizeof(struct iphdr), hlen); - if (options & F_AUDIBLE) - putchar('\a'); putchar('\n'); fflush(stdout); - } else { - putchar('\a'); - fflush(stdout); } return 0; } diff --git a/ping6.c b/ping6.c index c5ff881..8e6b991 100644 --- a/ping6.c +++ b/ping6.c @@ -1411,14 +1411,14 @@ parse_reply(struct msghdr *msg, int cc, void *addr, struct timeval *tv) pr_icmph(icmph->icmp6_type, icmph->icmp6_code, ntohl(icmph->icmp6_mtu)); } + if (options & F_AUDIBLE) { + putchar('\a'); + if(options & F_FLOOD) + fflush(stdout); + } if (!(options & F_FLOOD)) { - if (options & F_AUDIBLE) - putchar('\a'); putchar('\n'); fflush(stdout); - } else { - putchar('\a'); - fflush(stdout); } return 0; } -- 1.7.1 -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: i386 (x86_64) Kernel: Linux 2.6.36-rc7+ (SMP w/4 CPU cores) Locale: LANG=en_US.ISO-8859-15, LC_CTYPE=en_US.ISO-8859-15 (charmap=ISO-8859-15) Shell: /bin/sh linked to /bin/dash Versions of packages iputils-ping depends on: ii libc6 2.11.2-6 Embedded GNU C Library: Shared lib ii libssl0.9.8 0.9.8o-2 SSL shared libraries iputils-ping recommends no packages. iputils-ping suggests no packages. -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#600153: kword: can't paste selected text
Package: kword Version: 1:2.2.1-3 Severity: normal kword is no longer able to paste from the selection buffer with the middle mouse click, even when the selected text comes from the same document. A middle click will paste from the clipboard. Selecting text in kword will populate the selection buffer, so you can select something in kword and paste it to an xterm, but you no longer can select in an xterm and paste to kword. -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: i386 (x86_64) Kernel: Linux 2.6.36-rc7+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/dash Versions of packages kword depends on: ii kdebase-runtime 4:4.4.5-1runtime components from the offici ii koffice-libs1:2.2.1-3common libraries and binaries for ii kword-data 1:2.2.1-3data files for KWord word processo ii libc6 2.11.2-6 Embedded GNU C Library: Shared lib ii libgcc1 1:4.4.5-2GCC support library ii libglib2.0-02.24.2-1 The GLib library of C routines ii libgsf-1-1141.14.18-1Structured File Library - runtime ii libkdecore5 4:4.4.5-1the KDE Platform Core Library ii libkdeui5 4:4.4.5-1the KDE Platform User Interface Li ii libkhtml5 4:4.4.5-1the KHTML Web Content Rendering En ii libkio5 4:4.4.5-1the Network-enabled File Managemen ii libkparts4 4:4.4.5-1the Framework for the KDE Platform ii libqt4-qt3support 4:4.6.3-1+b1 Qt 3 compatibility library for Qt ii libqt4-svg 4:4.6.3-1+b1 Qt 4 SVG module ii libqt4-xml 4:4.6.3-1+b1 Qt 4 XML module ii libqtcore4 4:4.6.3-1+b1 Qt 4 core module ii libqtgui4 4:4.6.3-1+b1 Qt 4 GUI module ii libstdc++6 4.4.5-2 The GNU Standard C++ Library v3 ii libwpd8c2a 0.8.14-1 Library for handling WordPerfect d ii zlib1g 1:1.2.3.4.dfsg-3 compression library - runtime kword recommends no packages. Versions of packages kword suggests: ii khelpcenter4 4:4.4.5-1 help center ii koffice-doc-html 1:2.2.1-3 KDE Office Suite documentation in -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#600152: kspread: can't paste numbers
Package: kspread Version: 1:2.2.1-3 Severity: important A common use of spread sheets is to calculate numbers that come from some other program. Select some text from another program like $2.34 and middle click to paste into a kspread cell. The cell is left justified indicating that kspread is classifying it as a text string, not a number. That is the same result as doing a copy and paste operation with the same text (instead of the selection buffer). Type $2.34 into another cell and it is right justified as kspread interprits it as a number. As long as kspread classifies it as text none of the formulas will operate on it, such as sum, and unless you realize that it is left justified it isn't immediately obvious why all your forumlas are broken. Editing the cell in kspread and it will properly revert to a number. -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: i386 (x86_64) Kernel: Linux 2.6.36-rc7+ (SMP w/4 CPU cores) Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/dash Versions of packages kspread depends on: ii kdebase-runtime 4:4.4.5-1runtime components from the offici ii koffice-libs1:2.2.1-3common libraries and binaries for ii libc6 2.11.2-6 Embedded GNU C Library: Shared lib ii libgcc1 1:4.4.5-2GCC support library ii libgsl0ldbl 1.14+dfsg-1 GNU Scientific Library (GSL) -- li ii libkde3support4 4:4.4.5-1the KDE 3 Support Library for the ii libkdecore5 4:4.4.5-1the KDE Platform Core Library ii libkdeui5 4:4.4.5-1the KDE Platform User Interface Li ii libkhtml5 4:4.4.5-1the KHTML Web Content Rendering En ii libkio5 4:4.4.5-1the Network-enabled File Managemen ii libknotifyconfig4 4:4.4.5-1library for configuring KDE Notifi ii libkparts4 4:4.4.5-1the Framework for the KDE Platform ii libkrosscore4 4:4.4.5-1the Kross Core Library ii libkutils4 4:4.4.5-1various utility classes for the KD ii libqt4-dbus 4:4.6.3-1+b1 Qt 4 D-Bus module ii libqt4-qt3support 4:4.6.3-1+b1 Qt 3 compatibility library for Qt ii libqt4-sql 4:4.6.3-1+b1 Qt 4 SQL module ii libqt4-xml 4:4.6.3-1+b1 Qt 4 XML module ii libqtcore4 4:4.6.3-1+b1 Qt 4 core module ii libqtgui4 4:4.6.3-1+b1 Qt 4 GUI module ii libstdc++6 4.4.5-2 The GNU Standard C++ Library v3 kspread recommends no packages. Versions of packages kspread suggests: ii khelpcenter4 4:4.4.5-1 help center ii koffice-doc-html 1:2.2.1-3 KDE Office Suite documentation in -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#596452: /usr/bin/ucf: $PAGER ignored
Package: ucf Version: 3.0025 Severity: important File: /usr/bin/ucf [More] is displayed at the end of each page requring enter to be pressed. This is dangerous because you can't just hold down return to get to the end of a very long diff because that starts answering who knows what questions that were coming up. This makes the diff output almost as unusable as putting it in a text dialog where you can't copy and paste because of the line wraps and side characters. The problem is the $PAGER environment variable is ignored and should not be, I have libterm-readline-gnu-perl installed and the following in /var/cache/debconf/config.dat Name: debconf/frontend Template: debconf/frontend Value: Readline Owners: debconf, debconf-tiny Flags: seen I specifically ran this example after being ignored by the upgrade. /usr/bin/ucf --three-way /etc/apcupsd/apcupsd.conf /etc/apcupsd/apcupsd.conf.dpkg-dist -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i586) Kernel: Linux 2.6.34-rc2 Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/dash Versions of packages ucf depends on: ii coreutils 8.5-1 GNU core utilities ii debconf 1.5.35 Debian configuration management sy ucf recommends no packages. ucf suggests no packages. -- debconf information: * ucf/show_diff: ucf/changeprompt_threeway: keep_current ucf/title: * ucf/changeprompt: keep_current ucf/conflicts_found: -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#593308: xscreensaver: select instead of sleep in waiting for the reply
On Wed, Aug 18, 2010 at 09:01:31PM -0700, Jamie Zawinski wrote: > This looks good, thanks! But I think you do need to go around the > loop multiple times, because there's technically no guarantee that > the response event we're polling for in XCheckIfEvent will be the > first event to come in. There might be some event we don't care > about that causes select() to return, and then the event we need > comes in some fraction of a second later. I saw about three ways to easily implement the select. 1. select with the existing counter (the first patch I sent in) If select times out (and returns zero) it will check one last time (which shouldn't be necessary, but but no harm either). If select found fd readable it will loop to check the event, and select again up to 10 times if none of them timeout. To address your comment, it does go around the loop multiple times in both cases. I liked this solution because it was the most compact change and I felt that it was extremely highly likely that even if there were other messages going to xscreensaver-command before the expected event there wouldn't be ten sets (one to trigger each select return), before the event of interest. On the other hand I do grant that if the xscreensaver isn't responding it could happen, and it would return before the current ten second timeout when the xscreensaver might have responded before ten seconds. The other problem is if xscreensaver really isn't responding and there are other messages coming at a very low rate, say 9.9 seconds apart, select always restarts with ten seconds and so it could wait 99 seconds before returning, where the previous code would be fixed at ten seconds. 2. select and gettimeofday (patch follows) This wasn't my first choice because it is just more code, and I thought it was more complicated. By adding gettimeofday it will wait ten seconds no matter how many messages come in that it isn't looking for. It's about ten seconds because I didn't bother calculating the fractions of a second. 3. one second selects and variations Instead of a ten second select, use ten one second selects. This is kind of like #1 only it can't go over ten seconds, but then it's calling select ten times. Like #1 it can also go way under ten seconds if other messages are coming in. I didn't code this one up. Feel free to take your pick. Thanks for your quick reply. >From 873bb17999c7c3384c36611f35171ce02f889853 Mon Sep 17 00:00:00 2001 From: David Fries Date: Mon, 16 Aug 2010 21:59:31 -0500 Subject: [PATCH] Call select to wait for the reply instead of sleeping. This version will wait 10 seconds even when receiving messages that it isn't interested in. The previous version would execute the loop 10 times, so if there were 11 messages before the event of interest it wouldn't get it even if it in total only waited a fraction of a second. --- driver/remote.c | 41 +++-- 1 files changed, 39 insertions(+), 2 deletions(-) diff --git a/driver/remote.c b/driver/remote.c index 1e67c97..c706f6c 100644 --- a/driver/remote.c +++ b/driver/remote.c @@ -371,17 +371,54 @@ xscreensaver_command_response (Display *dpy, Window window, Bool verbose_p, Bool exiting_p, char **error_ret) { - int sleep_count = 0; char err[2048]; XEvent event; Bool got_event = False; + const int timeout = 10; + +#if defined(HAVE_SELECT) + struct timeval start, now; + struct timeval tv; + int fd = XConnectionNumber(dpy); + fd_set rset, rcopy; + +#ifdef GETTIMEOFDAY_TWO_ARGS + gettimeofday(&start, NULL); +#else + gettimeofday(&start); +#endif + now=start; + FD_ZERO(&rset); + FD_SET(fd, &rset); + /* Wait for the response for ten seconds (fractions aren't important). */ while (!(got_event = XCheckIfEvent(dpy, &event, &xscreensaver_command_event_p, 0)) && -sleep_count++ < 10) +now.tv_sec - start.tv_sec < timeout) +{ + tv.tv_sec = timeout - (now.tv_sec - start.tv_sec); + tv.tv_usec = 0; + rcopy = rset; + printf("sleep %d\n", (int)tv.tv_sec); + (void) select (fd+1, &rcopy, NULL, NULL, &tv); + +#ifdef GETTIMEOFDAY_TWO_ARGS + gettimeofday(&now, NULL); +#else + gettimeofday(&now); +#endif +} + +#else + int sleep_count = 0; + + while (!(got_event = XCheckIfEvent(dpy, &event, + &xscreensaver_command_event_p, 0)) && +sleep_count++ < timeout) { sleep(1); } +#endif if (!got_event) { -- 1.7.1 -- David Fries http://fries.net/~david/ (PGP encryption key available) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#593309: pm-utils: No video quirks required for Acer Aspire 5100
On Tue, Aug 17, 2010 at 09:48:47AM +0200, Michael Biebl wrote: > On 17.08.2010 05:51, David Fries wrote: > > Package: pm-utils > > Version: 1.3.0-2 > > Severity: normal > > > > > > I'm trying to increase the suspend/resume time. This Acer Aspire 5100 > > suspends and resumes just fine as far as video goes, with just the > > 2.6.35 kernel > > `echo mem > /sys/power/state` and also switching to --quirk-none. > > > > Hi David, > > have you also tested the currrent default Debian kernel 2.6.32? installing, linux-image-2.6.32-5-amd64 2.6.32-18 suspend to memory and resumed without any problems X reported, Kernel modesetting enabled. it is using the graphical frame buffer for console log: Initial commandline parameters: Wed Aug 18 11:21:28 CDT 2010: Running hooks for suspend. /usr/lib/pm-utils/sleep.d/000kernel-change suspend suspend:success. /usr/lib/pm-utils/sleep.d/00logging suspend suspend:Linux Pulsar 2.6.32-5-amd64 #1 SMP Sat Jul 24 01:47:24 UTC 2010 x86_64 GNU/Linux total used free sharedbuffers cached Mem: 1928880 2342161694664 0 12100 115088 -/+ buffers/cache: 1070281821852 Swap: 4008208 04008208 success. /usr/lib/pm-utils/sleep.d/00powersave suspend suspend:success. /usr/lib/pm-utils/sleep.d/49bluetooth suspend suspend:not applicable. /usr/lib/pm-utils/sleep.d/55NetworkManager suspend suspend:not applicable. /usr/lib/pm-utils/sleep.d/70action_wpa suspend suspend:success. /usr/lib/pm-utils/sleep.d/75modules suspend suspend:not applicable. /usr/lib/pm-utils/sleep.d/90clock suspend suspend:not applicable. /usr/lib/pm-utils/sleep.d/94cpufreq suspend suspend:success. /usr/lib/pm-utils/sleep.d/95anacron suspend suspend:success. /usr/lib/pm-utils/sleep.d/95led suspend suspend:not applicable. /usr/lib/pm-utils/sleep.d/98video-quirk-db-handler suspend suspend:success. /etc/pm/sleep.d/99hdparm suspend suspend:not applicable. /usr/lib/pm-utils/sleep.d/99video suspend suspend:kernel.acpi_video_flags = 0 success. Wed Aug 18 11:21:29 CDT 2010: performing suspend Wed Aug 18 11:21:38 CDT 2010: Awake. Wed Aug 18 11:21:38 CDT 2010: Running hooks for resume /usr/lib/pm-utils/sleep.d/99video resume suspend:success. /etc/pm/sleep.d/99hdparm resume suspend:Setting parameters of disc: /dev/sda . success. /usr/lib/pm-utils/sleep.d/98video-quirk-db-handler resume suspend:success. /usr/lib/pm-utils/sleep.d/95led resume suspend:not applicable. /usr/lib/pm-utils/sleep.d/95anacron resume suspend:success. /usr/lib/pm-utils/sleep.d/94cpufreq resume suspend:success. /usr/lib/pm-utils/sleep.d/90clock resume suspend:not applicable. /usr/lib/pm-utils/sleep.d/75modules resume suspend:success. /usr/lib/pm-utils/sleep.d/70action_wpa resume suspend:success. /usr/lib/pm-utils/sleep.d/55NetworkManager resume suspend:not applicable. /usr/lib/pm-utils/sleep.d/49bluetooth resume suspend:not applicable. /usr/lib/pm-utils/sleep.d/00powersave resume suspend:success. /usr/lib/pm-utils/sleep.d/00logging resume suspend:success. /usr/lib/pm-utils/sleep.d/000kernel-change resume suspend:success. Wed Aug 18 11:21:39 CDT 2010: Finished. hibernate to disk and resume without any problems X reported, Kernel modesetting enabled. it is using the graphical frame buffer for console log: Initial commandline parameters: Wed Aug 18 11:26:51 CDT 2010: Running hooks for hibernate. /usr/lib/pm-utils/sleep.d/000kernel-change hibernate hibernate:success. /usr/lib/pm-utils/sleep.d/00logging hibernate hibernate:Linux Pulsar 2.6.32-5-amd64 #1 SMP Sat Jul 24 01:47:24 UTC 2010 x86_64 GNU/Linux total used free sharedbuffers cached Mem: 1928880 2394481689432 0 12528 115704 -/+ buffers/cache: 1112161817664 Swap: 4008208 04008208 success. /usr/lib/pm-utils/sleep.d/00powersave hibernate hibernate:success. /usr/lib/pm-utils/sleep.d/49bluetooth hibernate hibernate:not applicable. /usr/lib/pm-utils/sleep.d/55NetworkManager hibernate hibernate:not applicable. /usr/lib/pm-utils/sleep.d/70action_wpa hibernate hibernate:success. /usr/lib/pm-utils/sleep.d/75modules hibernate hibernate:not applicable. /usr/lib/pm-utils/sleep.d/90clock hibernate hibernate:not applicable. /usr/lib/pm-utils/sleep.d/94cpufreq hibernate hibernate:success. /usr/lib/pm-utils/sleep.d/95anacron hibernate hibernate:success. /usr/lib/pm-utils/sleep.d/95led hibernate hibernate:not applicable. /usr/lib/pm-utils/sleep.d/98video-quirk-db-handler hibernate hibernate:success. /etc/pm/sleep.d/99hdparm hibernate hibernate:not applicable. /usr/lib/pm-utils/sleep.d/99video hibernate hibernate:success. Wed Aug 18 11:26:52 CDT 2010: performing hibernate Wed Aug 18 11:27:40 CDT 2010: Awake. Wed Aug 18 11:27:40 CDT 2010: Running hooks for thaw /usr/lib/pm-utils/sleep.d/99video thaw hibernate:success. /etc/pm/sleep.d/99hdparm thaw hibernate:Setting
Bug#593309: pm-utils: No video quirks required for Acer Aspire 5100
Package: pm-utils Version: 1.3.0-2 Severity: normal I'm trying to increase the suspend/resume time. This Acer Aspire 5100 suspends and resumes just fine as far as video goes, with just the 2.6.35 kernel `echo mem > /sys/power/state` and also switching to --quirk-none. Index: usr/lib/pm-utils/video-quirks/20-video-quirk-pm-acer.quirkdb === --- usr/lib/pm-utils/video-quirks/20-video-quirk-pm-acer.quirkdb (revision 112) +++ usr/lib/pm-utils/video-quirks/20-video-quirk-pm-acer.quirkdb (revision 113) @@ -94,7 +94,10 @@ addquirk --quirk-vbe-post addquirk --quirk-vbestate-restore endmatch - match system.hardware.product regex 1520|1650|5100|5110|5570|5920 + match system.hardware.product regex 5100 +addquirk --quirk-none + endmatch + match system.hardware.product regex 1520|1650|5110|5570|5920 addquirk --quirk-vbe-post addquirk --quirk-vbemode-restore endmatch -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 2.6.35 (SMP w/2 CPU cores) Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/dash Versions of packages pm-utils depends on: ii console-tools 1:0.2.3dbs-69 Linux console and font utilities ii kbd-compat [kbd] 1:0.2.3dbs-69 Wrappers around console-tools for ii powermgmt-base 1.31 Common utils and configs for power Versions of packages pm-utils recommends: ii procps1:3.2.8-9 /proc file system utilities pn radeontool (no description available) ii vbetool 1.1-2 run real-mode video BIOS code to a Versions of packages pm-utils suggests: ii cpufrequtils 007-1 utilities to deal with the cpufreq pn uswsusp(no description available) -- Configuration Files: /etc/kernel/postinst.d/pm-utils changed [not included] -- no debconf information -- debsums errors found: debsums: changed file /usr/lib/pm-utils/sleep.d/00logging (from pm-utils package) debsums: changed file /usr/lib/pm-utils/video-quirks/20-video-quirk-pm-acer.quirkdb (from pm-utils package) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#593308: xscreensaver: select instead of sleep in waiting for the reply
Package: xscreensaver Version: 5.11-1 Severity: normal Tags: patch When xscreensaver-command sends a command to the X server it will sleep for one second up to ten times waiting for the response from xscreensaver. In my experience it will normally check for events before the event has arrived and require a one second sleep, and generally the one second delay doesn't matter. For 2.6.35 on this system `echo mem > /sys/power/state` is taking 1.4 seconds to complete. The current acpi sleep.sh is calling xscreensaver-command twice, once with --throttle and another with --lock (which has an extra second delay), which is adding an extra three seconds to the suspend. This patch will make most of two seconds go away by using a select to process any responses as soon as they are received. --- driver/remote.c | 13 + 1 files changed, 13 insertions(+), 0 deletions(-) diff --git a/driver/remote.c b/driver/remote.c index 1e67c97..dba8171 100644 --- a/driver/remote.c +++ b/driver/remote.c @@ -380,7 +380,20 @@ xscreensaver_command_response (Display *dpy, Window window, &xscreensaver_command_event_p, 0)) && sleep_count++ < 10) { +#if defined(HAVE_SELECT) + struct timeval tv = {10, 0}; + int fd=XConnectionNumber(dpy); + fd_set rset; + FD_ZERO(&rset); + FD_SET(fd, &rset); + if (select (fd+1, &rset, NULL, NULL, &tv) == 0) +{ + /* timeout, no need to wait any more */ + sleep_count += 10; +} +#else sleep(1); +#endif } if (!got_event) -- 1.7.1 -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 2.6.35 (SMP w/2 CPU cores) Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/dash Versions of packages xscreensaver depends on: ii libatk1.0-0 1.30.0-1 The ATK accessibility toolkit ii libc6 2.11.2-2 Embedded GNU C Library: Shared lib ii libcairo2 1.8.10-4 The Cairo 2D vector graphics libra ii libfontconfig1 2.8.0-2.1generic font configuration library ii libfreetype62.4.0-2 FreeType 2 font engine, shared lib ii libglade2-0 1:2.6.4-1library to load .glade files at ru ii libglib2.0-02.24.1-1 The GLib library of C routines ii libgtk2.0-0 2.20.1-1 The GTK+ graphical user interface ii libice6 2:1.0.6-1X11 Inter-Client Exchange library ii libpam0g1.1.1-3 Pluggable Authentication Modules l ii libpango1.0-0 1.28.1-1 Layout and rendering of internatio ii libsm6 2:1.1.1-1X11 Session Management library ii libx11-62:1.3.3-3X11 client-side library ii libxext62:1.1.2-1X11 miscellaneous extension librar ii libxinerama12:1.1-3 X11 Xinerama extension library ii libxml2 2.7.7.dfsg-4 GNOME XML library ii libxmu6 2:1.0.5-1X11 miscellaneous utility library ii libxpm4 1:3.5.8-1X11 pixmap library ii libxrandr2 2:1.3.0-3X11 RandR extension library ii libxrender1 1:0.9.6-1X Rendering Extension client libra ii libxt6 1:1.0.7-1X11 toolkit intrinsics library ii libxxf86vm1 1:1.1.0-2X11 XFree86 video mode extension l ii xscreensaver-data 5.11-1 data files to be shared among scre Versions of packages xscreensaver recommends: ii libjpeg-progs 8b-1 Programs for manipulating JPEG fil ii perl [perl5] 5.10.1-13 Larry Wall's Practical Extraction ii wamerican [wordlist] 6-3 American English dictionary words ii xli1.17.0+20061110-3 command line tool for viewing imag ii xloadimage 4.1-16.1 Graphics file viewer under X11 Versions of packages xscreensaver suggests: pn fortune(no description available) ii iceape-browser [www-browser] 2.0.6-1Iceape Navigator (Internet browser ii iceweasel [www-browser] 3.5.10-1 Web browser based on Firefox ii konqueror [www-browser] 4:4.4.5-1 advanced file manager, web browser pn qcam | streamer(no description available) ii w3m [www-browser] 0.5.2-7WWW browsable pager with excellent pn xdaliclock (no description available) pn xfishtank (no description available) pn xscreensaver-gl(no description available) -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#593303: acpi-support: suspend in half the time - background xscreensaver
Package: acpi-support Version: 0.137-3 Severity: normal Tags: patch I was excited when the 2.6.35 kernel suspended to ram `echo mem > /sys/power/state` in 1.4 seconds, but then seeing Fn Zz acpi suspend hot key taking 6.6 seconds was disapointing, especially since /etc/acpi/sleep.sh was also using the same kernel suspend to ram method. Granted it was going to be slower because it has the whole acpi and pm-suspend set of shell scripts to process through, but 4.7 times slower? I tracked down 3+ seconds of that being xscreensaver-command sleeping three times for a second each. Here's a patch to execute those sleeps in the background and half the time to get the system suspended using the hot keys. It could be executed in another half the time, but I haven't tracked that down. As a side note, throttling the screen saver and forcing dpms to off is pointless when the cpu and graphics card are going to be powered off. Index: usr/share/acpi-support/screenblank === --- usr/share/acpi-support/screenblank (revision 106) +++ usr/share/acpi-support/screenblank (revision 107) @@ -1,7 +1,7 @@ if [ `pidof xscreensaver` ]; then - su $user -s /bin/sh -c "(xscreensaver-command -throttle)" + su $user -s /bin/sh -c "(xscreensaver-command -throttle&)" if [ x$LOCK_SCREEN = xtrue ]; then - su $user -s /bin/sh -c "(xscreensaver-command -lock)" + su $user -s /bin/sh -c "(xscreensaver-command -lock&)" fi elif [ `pidof dcopserver` ]; then if [ x$LOCK_SCREEN = xtrue ]; then @@ -16,13 +16,13 @@ fi case "$DISPLAY_DPMS" in xset) -su $user -s /bin/sh -c "xset dpms force off" +su $user -s /bin/sh -c "xset dpms force off&" ;; xrandr) - su $user -s /bin/sh -c "xrandr --output LVDS --off" + su $user -s /bin/sh -c "xrandr --output LVDS --off&" ;; vbetool) - /usr/sbin/vbetool dpms off + /usr/sbin/vbetool dpms off & ;; esac -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 2.6.35 (SMP w/2 CPU cores) Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/dash Versions of packages acpi-support depends on: ii acpi-fakekey 0.137-3tool to generate fake key events ii acpi-support-base 0.137-3scripts for handling base ACPI eve ii acpid 1:2.0.6-1 Advanced Configuration and Power I ii lsb-base 3.2-23.1 Linux Standard Base 3.2 init scrip ii pm-utils 1.3.0-2utilities and scripts for power ma ii x11-xserver-utils 7.5+2 X server utilities Versions of packages acpi-support recommends: ii dbus 1.2.24-3 simple interprocess messaging syst pn radeontool (no description available) ii vbetool 1.1-2 run real-mode video BIOS code to a ii xscreensaver 5.11-1 Automatic screensaver for X Versions of packages acpi-support suggests: pn rfkill (no description available) ii xinput1.5.2-1Runtime configuration and test of -- no debconf information -- debsums errors found: debsums: changed file /usr/share/acpi-support/screenblank (from acpi-support package) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#586166: libqt4-dev: moc autoinclude fix ignore dot in directories
Package: libqt4-dev Version: 4:4.6.2-4 Severity: normal Tags: patch *** Please type your report below this line *** A period in the path of a file given to the moc generator should not be used to determine if it is a header file or a source file, currently it can. This causes problem with build generators like cmake that use absolute paths and projects that use extension-less header files, because if there is a period anyplace in the path it will fail. This is a problem because directories outside of the source repository can influence the build, even though it might be working on another system or mysteriously fail after the source was moved. See the bug report for more details. http://bugreports.qt.nokia.com/browse/QTBUG-11369 It has been merged upstream, and here is the commit entry in the 4.6 qt git tree. http://qt.gitorious.org/qt/qt/commit/78e74249014b152b93dae07bf194b16769016ad0 and a copy of the patch. --- src/tools/moc/main.cpp.orig 2010-05-13 19:29:29.645663823 -0500 +++ src/tools/moc/main.cpp 2010-05-13 19:30:37.950162304 -0500 @@ -359,11 +359,9 @@ int runMoc(int _argc, char **_argv) if (autoInclude) { +int spos = filename.lastIndexOf(QDir::separator().toLatin1()); int ppos = filename.lastIndexOf('.'); -moc.noInclude = (ppos >= 0 - && tolower(filename[ppos + 1]) != 'h' - && tolower(filename[ppos + 1]) != QDir::separator().toLatin1() -); +moc.noInclude = ppos > spos && ppos >= 0 && tolower(filename[ppos + 1]) != 'h'; } if (moc.includeFiles.isEmpty()) { if (moc.includePath.isEmpty()) { -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 2.6.35-rc1+ (SMP w/2 CPU cores) Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/dash Versions of packages libqt4-dev depends on: ii libc6 2.10.2-9 Embedded GNU C Library: Shared lib ii libgcc1 1:4.4.4-1GCC support library ii libqt4-assistant4:4.6.2-4Qt 4 assistant module ii libqt4-dbus 4:4.6.2-4Qt 4 D-Bus module ii libqt4-designer 4:4.6.2-4Qt 4 designer module ii libqt4-help 4:4.6.2-4Qt 4 help module ii libqt4-multimedia 4:4.6.2-4Qt 4 Multimedia module ii libqt4-network 4:4.6.2-4Qt 4 network module ii libqt4-qt3support 4:4.6.2-4Qt 3 compatibility library for Qt ii libqt4-script 4:4.6.2-4Qt 4 script module ii libqt4-scripttools 4:4.6.2-4Qt 4 script tools module ii libqt4-sql 4:4.6.2-4Qt 4 SQL module ii libqt4-svg 4:4.6.2-4Qt 4 SVG module ii libqt4-test 4:4.6.2-4Qt 4 test module ii libqt4-webkit 4:4.6.2-4Qt 4 WebKit module ii libqt4-xml 4:4.6.2-4Qt 4 XML module ii libqt4-xmlpatterns 4:4.6.2-4Qt 4 XML patterns module ii libqtcore4 4:4.6.2-4Qt 4 core module ii libqtgui4 4:4.6.2-4Qt 4 GUI module ii libstdc++6 4.4.4-1 The GNU Standard C++ Library v3 ii qt4-qmake 4:4.6.2-4Qt 4 qmake Makefile generator tool ii zlib1g 1:1.2.3.4.dfsg-3 compression library - runtime Versions of packages libqt4-dev recommends: ii libqt4-opengl-dev 4:4.6.2-4 Qt 4 OpenGL library development fi Versions of packages libqt4-dev suggests: pn firebird2.1-dev(no description available) ii libmysqlclient15-dev [ 5.0.51a-24+lenny1 MySQL database development files ii libpq-dev 8.4.4-1 header files for libpq5 (PostgreSQ pn libsqlite0-dev (no description available) ii libsqlite3-dev 3.6.23.1-2SQLite 3 development files ii qt4-dev-tools 4:4.6.2-4 Qt 4 development tools ii qt4-doc4:4.6.2-4 Qt 4 API documentation pn unixodbc-dev (no description available) -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#584686: rdesktop: fix clipboard fails due to missing timestamp problem
Package: rdesktop Version: 1.6.0-3 Severity: important Tags: patch *** Please type your report below this line *** Summary: Don't abort a clipboard operation on a failed timestamp transfer. X clients that register selections for clipboard and primary selections are supposed to support the TIMESTAMP query to find out the time that they acquired the selection. Not all clients support the timestamp query. rdesktop will pull from either the primary or clipboard selection when a Unix to Windows clipboard operation is requested, if only one of the two is currently owned that one is selected for a transfer. If both are owned rdesktop sends both selection owners a timestamp request and selects the newer of the two for the clipboard operation. Without this patch if both primary and clipboard are owned, rdesktop aborts the transfer on the first failed (not supported) timestamp query, this is very non-obvious and difficult to clear. The user just sees clipboard transfers fail randomly and until the application that doesn't support timestamps looses its selection clipboard operations Unix to Windows fail. Currently if a client returns a timestamp of 0 it is set to 1 for the purpose of comparing the to selection timestamp values. rdesktop uses 0 to mean it has not yet received a timestamp for that application. A returned 0 timestamp value indicates the client supports timestamp, but probably used 0 when acquiring the selection in the first place even though clients are not supposed to. This patch will set the timestamp to 1 for any client that doesn't support timestamps so at least a transfer can take place even though it can't know for sure which selection is the newest and will select the clipboard over the primary selection in the case of a tie. -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 2.6.35-rc1+ (SMP w/2 CPU cores) Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/dash Versions of packages rdesktop depends on: ii libasound21.0.22-2 shared library for ALSA applicatio ii libc6 2.10.2-9 Embedded GNU C Library: Shared lib ii libssl0.9.8 0.9.8n-1 SSL shared libraries ii libx11-6 2:1.3.3-3 X11 client-side library rdesktop recommends no packages. rdesktop suggests no packages. -- no debconf information --- xclip.c | 47 --- 1 files changed, 24 insertions(+), 23 deletions(-) diff --git a/xclip.c b/xclip.c index 37290fc..5de88dd 100644 --- a/xclip.c +++ b/xclip.c @@ -508,58 +508,59 @@ xclip_handle_SelectionNotify(XSelectionEvent * event) int res, i, format; uint8 *data = NULL; - if (event->property == None) + if (event->target != timestamp_atom && event->property == None) goto fail; DEBUG_CLIPBOARD(("xclip_handle_SelectionNotify: selection=%s, target=%s, property=%s\n", XGetAtomName(g_display, event->selection), XGetAtomName(g_display, event->target), - XGetAtomName(g_display, event->property))); + event->property ? XGetAtomName(g_display, event->property) : "None")); if (event->target == timestamp_atom) { - if (event->selection == primary_atom) + Time timestamp; + Atom query = event->selection == primary_atom ? + rdesktop_primary_timestamp_target_atom : + rdesktop_clipboard_timestamp_target_atom; + if (event->property == None) { - res = XGetWindowProperty(g_display, g_wnd, - rdesktop_primary_timestamp_target_atom, 0, - XMaxRequestSize(g_display), False, AnyPropertyType, - &type, &format, &nitems, &bytes_left, &data); + /* If a client doesn't support timestamp, treat it the same as ones that + * return zero. + */ + timestamp = 1; } else { - res = XGetWindowProperty(g_display, g_wnd, - rdesktop_clipboard_timestamp_target_atom, 0, + res = XGetWindowProperty(g_display, g_wnd, query, 0, XMaxRequestSize(g_display), False, AnyPropertyType, &type, &format, &nitems, &bytes_left, &data); - } + if ((res != Success) || (nitems != 1) || (format != 32)) + { +DEBUG_CLIPBOARD(("XGetWindowProperty failed!\n")); +goto fail; + } - if ((res != Success) || (nitems != 1) || (format != 32)) - { - DEBUG_CLIPBOARD(("XGetWindowProperty failed!\n")); - goto fail; + timestamp = *(Time *) data; + if (timestamp == 0) +timestamp++; + XDeleteProperty(g_display, g_wnd, query); + XFree(data); } if (event->selection == primary_atom) { - primary_timestamp = *(Time *) data; - if (primary_timestamp == 0) -primary_timestamp++; - XDeleteProperty(g_display, g_wnd, rdesktop_primary_timestamp_target_atom); + primary_timestamp = timestamp; DEBUG_CLIPBOARD(("Got PRIMARY timestamp: %u\n", (unsigned) primary_timestamp)); } else { - clipboard_timestamp = *(Time *) data; - if (clipboard
Bug#584684: rdesktop: make right Alt usable (AltGr killed it)
Package: rdesktop Version: 1.6.0-3 Severity: normal Tags: patch *** Please type your report below this line *** If AltGr is down, leave it down, the translators list when it needs to be added not removed. It's either already held for the key (if the keyboard mappings are the same), or it isn't and isn't required for that keysym. Currently for en-us keyboard layouts the right alt is nearly useless. It can't be used with any other key in chording such as Alt-F4, Alt-F, Alt-Left, Alt-Right, trying will send an Alt_R up before and Alt_R down after causing windows to see it as if the Alt_R wasn't held. I think this is a good compromise, those who are using Alt_R for AltGr composition will be used to not holding it when pressing other keys, and for the rest of us it will finally work. Note, I don't have an AltGr client or server to test this with. -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 2.6.35-rc1+ (SMP w/2 CPU cores) Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/dash Versions of packages rdesktop depends on: ii libasound21.0.22-2 shared library for ALSA applicatio ii libc6 2.10.2-9 Embedded GNU C Library: Shared lib ii libssl0.9.8 0.9.8n-1 SSL shared libraries ii libx11-6 2:1.3.3-3 X11 client-side library rdesktop recommends no packages. rdesktop suggests no packages. -- no debconf information --- xkeymap.c |4 1 files changed, 4 insertions(+), 0 deletions(-) diff --git a/xkeymap.c b/xkeymap.c index f729b4e..937e9f1 100644 --- a/xkeymap.c +++ b/xkeymap.c @@ -714,6 +714,10 @@ xkeymap_send_keys(uint32 keysym, unsigned int keycode, unsigned int state, uint3 if (tr.scancode == 0) return; + /* If AltGr is down, leave it down, the translators list + * when it needs to be added not removed. + */ + tr.modifiers |= remote_modifier_state & MapAltGrMask; save_remote_modifiers(tr.scancode); ensure_remote_modifiers(ev_time, tr); rdp_send_scancode(ev_time, pressed ? RDP_KEYPRESS : RDP_KEYRELEASE, tr.scancode); -- 1.7.0
Bug#577036: rxvt: add selection TIMESTAMP support
Package: rxvt Version: 1:2.6.4-14 Severity: normal Tags: patch Some clients expect TIMESTAMP to be provided and misbehave otherwise. Especially those that try to read the TIMESTAMP to determine if the primary selection or clipboard should be read. diff --git a/src/screen.c b/src/screen.c index 6c78e08..193b33f 100644 --- a/src/screen.c +++ b/src/screen.c @@ -2894,6 +2894,7 @@ selection_clear(void) FREE(selection.text); selection.text = NULL; selection.len = 0; +selection.own_time = 0; CLEAR_SELECTION; } @@ -2983,6 +2984,8 @@ selection_make(Time tm) XSetSelectionOwner(Xdisplay, XA_PRIMARY, TermWin.vt, tm); if (XGetSelectionOwner(Xdisplay, XA_PRIMARY) != TermWin.vt) print_error("can't get primary selection"); +else + selection.own_time = tm; XChangeProperty(Xdisplay, Xroot, XA_CUT_BUFFER0, XA_STRING, 8, PropModeReplace, selection.text, selection.len); D_SELECT((stderr, "selection_make(): selection.len=%d", selection.len)); @@ -3423,6 +3426,7 @@ selection_send(const XSelectionRequestEvent * rq) static Atom xa_targets = None; static Atom xa_compound_text = None; static Atom xa_text = None; +static Atom xa_timestamp = None; XTextProperty ct; XICCEncodingStyle style; char *cl[4]; @@ -3433,6 +3437,8 @@ selection_send(const XSelectionRequestEvent * rq) xa_compound_text = XInternAtom(Xdisplay, "COMPOUND_TEXT", False); if (xa_targets == None) xa_targets = XInternAtom(Xdisplay, "TARGETS", False); +if (xa_timestamp == None) + xa_timestamp = XInternAtom(Xdisplay, "TIMESTAMP", False); ev.xselection.type = SelectionNotify; ev.xselection.property = None; @@ -3442,7 +3448,12 @@ selection_send(const XSelectionRequestEvent * rq) ev.xselection.target = rq->target; ev.xselection.time = rq->time; -if (rq->target == xa_targets) { +if (rq->target == xa_timestamp && selection.own_time) { + XChangeProperty(Xdisplay, rq->requestor, rq->property, XA_INTEGER, + 32, PropModeReplace, + (unsigned char *)&selection.own_time, 1); + ev.xselection.property = rq->property; +} else if (rq->target == xa_targets) { target_list[0] = xa_targets; target_list[1] = XA_STRING; target_list[2] = xa_text; diff --git a/src/screen.h b/src/screen.h index 5c2fca5..13afa4c 100644 --- a/src/screen.h +++ b/src/screen.h @@ -83,6 +83,7 @@ typedef struct { short screen;/* screen being used */ short clicks;/* number of clicks */ row_col_t beg, mark, end; +Timeown_time; /* time of selection ownership */ } selection_t; typedef enum { -- 1.7.0 -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 2.6.34-rc2 (SMP w/2 CPU cores) Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/dash Versions of packages rxvt depends on: ii libc6 2.10.2-6 Embedded GNU C Library: Shared lib ii libx11-6 2:1.3.3-2 X11 client-side library ii libxpm4 1:3.5.8-1 X11 pixmap library rxvt recommends no packages. rxvt suggests no packages. -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#577035: vim-gtk: timestamp is incorrectly 0 for selection transfers
Package: vim-gtk Version: 2:7.2.330-1 Severity: normal Tags: patch >From fe54576a85eb1f27bf02cd397d84f8a2379b2bca Mon Sep 17 00:00:00 2001 From: David Fries Date: Fri, 12 Mar 2010 09:17:11 -0600 Subject: [PATCH] Seed XtOwnSelection with non-zero time. XtOwnSelection is given CurrentTime for the selection time value. intrinsics.PS lists "the value CurrentTime is not acceptable." CurrentTime is 0 and that will be returned to any client requesting the TIMESTAMP conversion. Some clients request a timestamp from both primary and clipboard selection owners and use the most recent of the two, always returning 0 for a timestamp causes problems with these programs. --- src/ui.c | 14 +- 1 files changed, 13 insertions(+), 1 deletions(-) diff --git a/src/ui.c b/src/ui.c index df8008c..61f33f0 100644 --- a/src/ui.c +++ b/src/ui.c @@ -2341,7 +2341,19 @@ clip_x11_own_selection(myShell, cbd) Widget myShell; VimClipboard *cbd; { -if (XtOwnSelection(myShell, cbd->sel_atom, CurrentTime, +XEvent event; + +/* Get the time by a zero-length append, then process events until the + * notify comes back. */ +XChangeProperty(XtDisplay(myShell), XtWindow(myShell), cbd->sel_atom, +XA_ATOM, 32, PropModeAppend, NULL, 0); +while (XNextEvent(XtDisplay(myShell), &event)) +{ + XtDispatchEvent(&event); + if (event.type == PropertyNotify) + break; +} +if (XtOwnSelection(myShell, cbd->sel_atom, event.xproperty.time, clip_x11_convert_selection_cb, clip_x11_lose_ownership_cb, NULL) == False) return FAIL; -- 1.7.0 -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 2.6.34-rc2 (SMP w/2 CPU cores) Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/dash Versions of packages vim-gtk depends on: ii libacl1 2.2.49-2 Access control list shared library ii libc6 2.10.2-6 Embedded GNU C Library: Shared lib ii libglib2.0-0 2.22.4-1 The GLib library of C routines ii libgpm2 1.20.4-3.3 General Purpose Mouse - shared lib ii libgtk2.0-0 2.18.9-2 The GTK+ graphical user interface ii libice6 2:1.0.6-1 X11 Inter-Client Exchange library ii libncurses5 5.7+20100313-1 shared libraries for terminal hand ii libpango1.0-0 1.26.2-2 Layout and rendering of internatio ii libperl5.10 5.10.1-11 shared Perl library ii libruby1.81.8.7.249-2Libraries necessary to run Ruby 1. ii libselinux1 2.0.89-4 SELinux runtime shared libraries ii libsm62:1.1.1-1 X11 Session Management library ii libx11-6 2:1.3.3-2 X11 client-side library ii libxt61:1.0.7-1 X11 toolkit intrinsics library ii python2.5 2.5.5-2An interactive high-level object-o ii tcl8.48.4.19-4 Tcl (the Tool Command Language) v8 ii vim-common2:7.2.330-1Vi IMproved - Common files ii vim-gui-common2:7.2.330-1Vi IMproved - Common GUI files ii vim-runtime 2:7.2.330-1Vi IMproved - Runtime files vim-gtk recommends no packages. Versions of packages vim-gtk suggests: ii cscope 15.7a-2 Interactively examine a C program pn gnome-icon-theme (no description available) ii ttf-dejavu 2.30-2 Metapackage to pull in ttf-dejavu- ii vim-doc 2:7.2.330-1 Vi IMproved - HTML documentation -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#573395: rdesktop: [PATCH] don't duplicate scroll wheel events
Package: rdesktop Version: 1.6.0-2 Severity: normal >From 096a39b35773e8d58358f2176b9b96cc6edfb16d Mon Sep 17 00:00:00 2001 From: David Fries Date: Wed, 10 Mar 2010 19:00:23 -0600 Subject: [PATCH] skip scroll wheel up event (generates double events) Unix has a mouse scroll event button press followed by release for that button number. For windows just send button press and skip button release. If the release is sent, it is treated as a second scroll action on the windows side. Tested with Windows XP. It's easiest to see when scrolling through a combo box, previously it would skip entries. Now it doesn't. --- xwin.c | 10 ++ 1 files changed, 10 insertions(+), 0 deletions(-) diff --git a/xwin.c b/xwin.c index 1691d00..235b4e8 100644 --- a/xwin.c +++ b/xwin.c @@ -2254,6 +2254,16 @@ handle_button_event(XEvent xevent, RD_BOOL down) } } + /* Unix has a mouse scroll event button press followed by release +* for that button number. For windows just send button press and +* skip button release. If the release is sent, it is treated as +* a second scroll action on the windows side. Tested with Windows XP. +*/ + if (!down && + (button == MOUSE_FLAG_BUTTON4 || button == MOUSE_FLAG_BUTTON5)) + { + return; + } if (xevent.xmotion.window == g_wnd) { rdp_send_input(time(NULL), RDP_INPUT_MOUSE, -- 1.7.0 -- System Information: Debian Release: 5.0.2 APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 2.6.33-rc4 (SMP w/2 CPU cores) Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/bash Versions of packages rdesktop depends on: ii libc6 2.9-25 GNU C Library: Shared libraries ii libssl0.9.8 0.9.8k-8 SSL shared libraries ii libx11-6 2:1.2.2-1 X11 client-side library rdesktop recommends no packages. rdesktop suggests no packages. -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#409349: [PATCH] control queue full fix
I tracked down this issue. Using usbmon, the control and interrupt transfers would go on for days without problem, but once and a while, d1c301e0 2316273046 S Ci:1:002:0 s a1 01 0314 0003 8 < de7cfde0 233494 C Ii:1:002:1 0:64 6 = 0c64 3c00 de7cfde0 2334333468 S Ii:1:002:1 -115:64 6 < de7cfde0 2337308683 C Ii:1:002:1 0:64 4 = 0608 there would be a control transfer submitted followed by only interrupt transfers. The control transfer would not complete and the usbhid queue would built up until it was full and "generic-usb 0003:051D:0002.0001: control queue full" would be printed from then on. The following patch has been merged into the git kernel tree after 2.6.33, and will be in 2.6.34. I verified it fixes this bug on my system. Note the 10 second delay and error -104 in one request. c2fede40 161671434 S Ci:1:002:0 s a1 01 0326 0004 8 < c2fede40 161747009 C Ci:1:002:0 0 4 = 26343130 c2fede40 161847590 S Ci:1:002:0 s a1 01 0314 0003 8 < c2fede40 161848982 C Ci:1:002:0 0 3 = 14 c2fede40 161908315 S Ci:1:002:0 s a1 01 0314 0003 8 < c2fede40 171928585 C Ci:1:002:0 -104 3 = 14 c2fede40 171928605 S Ci:1:002:0 s a1 01 0306 0004 8 < c2fede40 171930579 C Ci:1:002:0 0 4 = 0608 c2fede40 171968852 S Ci:1:002:0 s a1 01 030c 0006 8 < c2fede40 171970572 C Ci:1:002:0 0 6 = 0c64 3f00 c2fede40 171989721 S Ci:1:002:0 s a1 01 030c 0006 8 < c2fede40 171991562 C Ci:1:002:0 0 6 = 0c64 3f00 -- David Fries http://fries.net/~david/ (PGP encryption key available) commit 858155fbcc0cd713f6382c527bb1c3abc0ed6d00 Author: Oliver Neukum Date: Fri Feb 12 13:02:28 2010 +0100 HID: usbhid: introduce timeout for stuck ctrl/out URBs Some devices do not react to a control request (seen on APC UPS's) resulting in a slow stream of messages, "generic-usb ... control queue full". Therefore request needs a timeout. Cc: sta...@kernel.org Signed-off-by: Oliver Neukum Signed-off-by: David Fries Signed-off-by: Jiri Kosina diff --git a/drivers/hid/usbhid/hid-core.c b/drivers/hid/usbhid/hid-core.c index 5406074..74bd3ca 100644 --- a/drivers/hid/usbhid/hid-core.c +++ b/drivers/hid/usbhid/hid-core.c @@ -316,6 +316,7 @@ static int hid_submit_out(struct hid_device *hid) err_hid("usb_submit_urb(out) failed"); return -1; } + usbhid->last_out = jiffies; } else { /* * queue work to wake up the device. @@ -377,6 +378,7 @@ static int hid_submit_ctrl(struct hid_device *hid) err_hid("usb_submit_urb(ctrl) failed"); return -1; } + usbhid->last_ctrl = jiffies; } else { /* * queue work to wake up the device. @@ -512,9 +514,20 @@ static void __usbhid_submit_report(struct hid_device *hid, struct hid_report *re usbhid->out[usbhid->outhead].report = report; usbhid->outhead = head; - if (!test_and_set_bit(HID_OUT_RUNNING, &usbhid->iofl)) + if (!test_and_set_bit(HID_OUT_RUNNING, &usbhid->iofl)) { if (hid_submit_out(hid)) clear_bit(HID_OUT_RUNNING, &usbhid->iofl); + } else { + /* +* the queue is known to run +* but an earlier request may be stuck +* we may need to time out +* no race because this is called under +* spinlock +*/ + if (time_after(jiffies, usbhid->last_out + HZ * 5)) + usb_unlink_urb(usbhid->urbout); + } return; } @@ -535,9 +548,20 @@ static void __usbhid_submit_report(struct hid_device *hid, struct hid_report *re usbhid->ctrl[usbhid->ctrlhead].dir = dir; usbhid->ctrlhead = head; - if (!test_and_set_bit(HID_CTRL_RUNNING, &usbhid->iofl)) + if (!test_and_set_bit(HID_CTRL_RUNNING, &usbhid->iofl)) { if (hid_submit_ctrl(hid)) clear_bit(HID_CTRL_RUNNING, &usbhid->iofl); + } else { + /* +* the queue is known to run +* but an earlier request may be stuck +* we may need to time out +* no race because this is called under +* spinlock +*/ + if (time_after(jiffies, usbhid->last_ctrl + HZ * 5)) + usb_unlink_urb(usbhid->urbctrl); + } } void usbhid_submit_report(struct hid_device *hid, struct hid_report *report, unsigned char dir) diff --git a/drivers/
Bug#539699: xscreensaver: unlocked because killed, infinite loop with small screen
Package: xscreensaver Version: 5.05-3 Severity: grave Tags: security patch Justification: user security hole Reproduce by setting a narrow X resolution, in this case I was running Xnest at 410x384, type something to get the password dialog, but it never comes up. xscreensaver goes into an infinite loop eating memory, and the one time I let it run, the Linux kernel out of memory detector killed xscreensaver unlocking the X server. That's why I marked this with the security tags. Do any hardware and X software automatically pick up and use newly plugged in displays? If so, it might be exploited by, plugging in a custom display device with a small screen, and use this exploit to kill the screen saver get access to the system, then restart the xscreensaver to make detection more difficult. The problem is when mlstring_wrap detects a space will copy the whitespace on to a new line and then truncate the original string for the current line. If in the next iteration the line is still too long, and the current whitespace character is the first one encountered it is in an infinite loop finding the same whitespace character. In my case line_length is 5, and the string is "Please enter your password." This patch will leave the whitespace in the previous line so it can't be found the next iteration. It will also make the previous line too wide, but only the whitespace would be over the border. On a site note it might be a good idea to always put the logo on the right side, or put it on the right side when the width is too small. With the patch at 410 pixels wide the logo takes up almost all of the screen and the text is just visible at the right side. diff --git a/driver/mlstring.c b/driver/mlstring.c index d6df844..a850890 100644 --- a/driver/mlstring.c +++ b/driver/mlstring.c @@ -153,6 +153,8 @@ mlstring_wrap(mlstring *mstring, XFontStruct *font, Dimension width) if (wrap_at == -1) /* No space found, hard wrap */ wrap_at = line_length; + else + wrap_at++; /* Leave the space at the end of the line. */ newml = calloc(1, sizeof(*newml)); if (!newml) /* OOM, don't bother trying to wrap */ -- System Information: Debian Release: 5.0.2 APT prefers stable APT policy: (500, 'stable') Architecture: i386 (i586) Kernel: Linux 2.6.29-rc3 Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/bash Versions of packages xscreensaver depends on: ii libatk1.0-0 1.22.0-1 The ATK accessibility toolkit ii libc6 2.7-18 GNU C Library: Shared libraries ii libcairo2 1.6.4-7 The Cairo 2D vector graphics libra ii libglade2-0 1:2.6.2-1library to load .glade files at ru ii libglib2.0-02.16.6-2 The GLib library of C routines ii libgtk2.0-0 2.12.12-1~lenny1 The GTK+ graphical user interface ii libice6 2:1.0.4-1X11 Inter-Client Exchange library ii libpam0g1.0.1-5+lenny1 Pluggable Authentication Modules l ii libpango1.0-0 1.20.5-5 Layout and rendering of internatio ii libsm6 2:1.0.3-2X11 Session Management library ii libx11-62:1.1.5-2X11 client-side library ii libxext62:1.0.4-1X11 miscellaneous extension librar ii libxinerama12:1.0.3-2X11 Xinerama extension library ii libxml2 2.6.32.dfsg-5GNOME XML library ii libxmu6 2:1.0.4-1X11 miscellaneous utility library ii libxpm4 1:3.5.7-1X11 pixmap library ii libxrandr2 2:1.2.3-1X11 RandR extension library ii libxrender1 1:0.9.4-2X Rendering Extension client libra ii libxt6 1:1.0.5-3X11 toolkit intrinsics library ii libxxf86misc1 1:1.0.1-3X11 XFree86 miscellaneous extensio ii libxxf86vm1 1:1.0.2-1X11 XFree86 video mode extension l ii xscreensaver-data 5.05-3 data files to be shared among scre Versions of packages xscreensaver recommends: ii libjpeg-progs 6b-14 Programs for manipulating JPEG fil ii perl [perl5] 5.10.0-19 Larry Wall's Practical Extraction ii wamerican [wordlist] 6-2.3 American English dictionary words pn xli | xloadimage (no description available) Versions of packages xscreensaver suggests: ii fortune-mod [fortune] 1:1.99.1-3.1 provides fortune cookies on demand ii iceape-browser [www-bro 1.1.14-1 Iceape Navigator (Internet browser ii iceweasel [www-browser] 3.0.6-1 lightweight web browser based on M ii konqueror [www-browser] 4:3.5.9.dfsg.1-6 KDE's advanced file manager, web b ii lynx-cur [www-browser] 2.8.7dev9-2.1Text-mode WWW Browser with NLS sup ii streamer3.95.df
Bug#537632: vim-gtk: receive clipboard INCR transfer fails
Package: vim-gtk Version: 1:7.1.314-3+lenny2 Severity: normal vim fails when trying to receive incremental clipboard selection transfers. Specifically large 300KB selection transfers fail from vim to vim, but work from vim to kword. In vim select some text V"+ motion y and try to paste in another vim session "+p if the selection size was small it uses one transfer and succeed, if it is large, like 300KB worth, the clipboard owner will initiate an incremental selection transfer, but the requestor will fail, and try at least four times before giving up. It will then fall back to reading the CUT_BUFFER0 and if the desired text happens to be there it will appear to work, or if it is some other random text it can be really confusing what happened. vim seems to populate CUT_BUFFER0 on exit. Use xprop -root -remove CUT_BUFFER0 to make sure the cut buffer is empty. If it is empty vim will report, E353: Nothing in register + and no transfer takes place. I also checked vim-gtk_7.2.148-2_i386.deb and it has the same problem. -- System Information: Debian Release: 5.0.2 APT prefers stable APT policy: (500, 'stable') Architecture: i386 (i586) Kernel: Linux 2.6.29-rc3 Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Shell: /bin/sh linked to /bin/bash Versions of packages vim-gtk depends on: ii libacl1 2.2.47-2 Access control list shared library ii libc6 2.7-18 GNU C Library: Shared libraries ii libglib2.0-0 2.16.6-2 The GLib library of C routines ii libgpm2 1.20.4-3.1 General Purpose Mouse - shared lib ii libgtk2.0-0 2.12.12-1~lenny1 The GTK+ graphical user interface ii libice6 2:1.0.4-1 X11 Inter-Client Exchange library ii libncurses5 5.7+20081213-1 shared libraries for terminal hand ii libpango1.0-0 1.20.5-5 Layout and rendering of internatio ii libperl5.10 5.10.0-19 Shared Perl library ii libruby1.81.8.7.72-3 Libraries necessary to run Ruby 1. ii libselinux1 2.0.65-5 SELinux shared libraries ii libsm62:1.0.3-2 X11 Session Management library ii libx11-6 2:1.1.5-2 X11 client-side library ii libxt61:1.0.5-3 X11 toolkit intrinsics library ii python2.5 2.5.2-15 An interactive high-level object-o ii tcl8.48.4.19-2 Tcl (the Tool Command Language) v8 ii vim-common1:7.1.314-3+lenny2 Vi IMproved - Common files ii vim-gui-common1:7.1.314-3+lenny2 Vi IMproved - Common GUI files ii vim-runtime 1:7.1.314-3+lenny2 Vi IMproved - Runtime files vim-gtk recommends no packages. Versions of packages vim-gtk suggests: ii cscope15.6-6+lenny1 Interactively examine a C program pn gnome-icon-theme (no description available) ii ttf-dejavu2.25-3 Metapackage to pull in ttf-dejavu- ii vim-doc 1:7.1.314-3+lenny2 Vi IMproved - HTML documentation -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#409349: apcupsd: control queue full
On Thu, Mar 01, 2007 at 04:00:10PM +, Tom Wright wrote: > On Tuesday 20 February 2007 16:22, David H?rdeman wrote: > > Were you able to reproduce this with the latest Debian packaged Linux 2.6 > > kernel? I've seen this in the past, now it's with nearly the latest and greatest kernel. 2.6.29-rc4 generic-usb 0003:051D:0002.0001: control queue full It's been months since I had the issue last. Killing apcupsd and restarting it cleared the condition. I have the system e-mail /sbin/apcaccess every day, just to watch for the issue. I haven't seen this issue on my laptop which is also a ohci controller. apcupsd3.14.4-1 USB Controller: Silicon Integrated Systems [SiS] USB 1.1 Controller (rev 07) [OHCI]) -- David Fries http://fries.net/~david/ (PGP encryption key available) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#353559: update patch
I updated to Lenny and the bug is still present, but the patch doesn't apply cleanly, so here it is again. It is easy to verify the bug, strace rxvt tty completely cover the window with another echo test > /dev/pts/ --- rxvt-2.6.4/src/screen.c.orig2009-02-07 13:01:26.0 -0600 +++ rxvt-2.6.4/src/screen.c 2009-02-07 13:11:36.0 -0600 @@ -2204,10 +2204,17 @@ scr_refresh(int type) #endif int (*draw_string) (), (*draw_image_string) (); -if (type == NO_REFRESH) - return; -if (!TermWin.mapped) - return; + +if (type == NO_REFRESH || !TermWin.mapped) +{ +/* If the window is completely obscured (NO_REFRESH) we really don't + * want to refresh, that is why we are returning instead, but first we + * must say mark the flag don't refresh. A similar argument works for + * !mapped. + */ + want_refresh=0; + return; +} /* * A: set up vars -- David Fries http://fries.net/~david/ (PGP encryption key available) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#476880: fgfs-base: add joystick string "Logitech Inc. WingMan Strike Force 3D"
Package: fgfs-base Version: 0.9.10-1 Severity: normal *** Please type your report below this line *** The rudder and throttle are flipped here, it just needs the joystick named added to the right filename to use the fix. --- FlightGear/Input/Joysticks/Logitech/wingman-force-3d.xml.orig 2008-04-19 15:21:33.0 -0500 +++ FlightGear/Input/Joysticks/Logitech/wingman-force-3d.xml2008-04-19 15:21:36.0 -0500 @@ -27,6 +27,7 @@ Logitech Inc. WingMan Force 3D Logitech Inc. WingMan Force 3d Logitech WingMan Force 3D USB + Logitech Inc. WingMan Strike Force 3D Aileron -- System Information: Debian Release: 4.0 APT prefers stable APT policy: (500, 'stable') Architecture: amd64 (x86_64) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.22-rc6 Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#460314: milter-greylist: config file location comment update
Package: milter-greylist Version: 3.0-3 Severity: minor In /etc/default/milter-greylist it lists # Don't forget to edit the configurationfile /etc/mail/greylist.conf replace with, # Don't forget to edit the configuration file # /etc/milter-greylist/greylist.conf There needs to be a space between configuration and file and it specified the wrong location for the file. -- System Information: Debian Release: 4.0 APT prefers stable APT policy: (500, 'stable') Architecture: i386 (i586) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.22.2 Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Versions of packages milter-greylist depends on: ii adduser3.45 Add and remove users and groups ii libc6 2.3.6.ds1-13etch4 GNU C Library: Shared libraries ii libmilter0 8.13.8-3 Sendmail Mail Filter API (Milter) ii libspf2-2 1.2.5-4 Sender Policy Framework library, w Versions of packages milter-greylist recommends: ii sendmail 8.13.8-3 powerful, efficient, and scalable -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#442375: closed by Tollef Fog Heen <[EMAIL PROTECTED]> (Re: Bug#442375: libapr1-dev: update --includedir path, apr-1 not apr-1.0)
> You're supposed to use #include . > > Closing this non-bug. That doesn't address the original bug, just the commentary. apr-1.2.7/debian/rules lists --includedir=\$${prefix}/usr/include/apr-1.0 --with-installbuilddir=\$${prefix}/usr/share/apr-1.0/build The debian configuration overrides the default install location with a version directory apr-1.0 that doesn't match the program version 1.2.7. -- David Fries <[EMAIL PROTECTED]> http://fries.net/~david/ (PGP encryption key available) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#375787: I've analysed this some more...
On Thu, Oct 18, 2007 at 07:01:59PM -0700, Richard A Nelson wrote: > On Thu, 18 Oct 2007, Kees-Jan Dijkzeul wrote: > > This whole mess was concocted by me and a few others back when we were > forced to live on PPP/dialup or DHCP dsl/cable - and it worked just fine > for us, but at least I always ran a split horizon/view local DNS on my > mail servers - and thus never really ran into this issue. I manged to run for years without it being a problem. I don't know what happened that I got bitten that time either. The system serves DHCP and has static ip addresses. My fault for having dialup.m4 referenced. > I found a bug in the dynamic script that can cause early termination > of the host lookup loop - and explains the ';;' edit. > > I'm going to fix that, but before I put it out, I would like your > opinions on how to handle the remaining issue: > > /usr/share/sendmail/dynamic::update_host is called when the IP has > changed, and it calls find_host to do the resolution... just blindly > assuming it has done the job properly. > > If find_host works, there is no problem... > > If it fails, however, there are some options - none good ;( > > *) Keep sendmail from starting in an unknown state, and invent > a means to do the needed DNS lookup during > /etc/init.d/sendmail start > This likely means actually keeping some state around - and > we could even retry later on DHCP RENEWs ! > > *) Keep sendmail the 'logically offline state' - mail will > not be sent, I will have to investigate if it will accept > mail in this state or not (doing so could be *bad*) > > *) Use the prior hostname, which *may* be invalid - potentially > causing bounces, DNSBL additions, etc. I would say a modified last one. Keep the current configuration and sendmail running, but invalidate the peer and IP address cache, so that they don't have to change for the script to try again. Though I was looking through the /usr/share/sendmail/dynamic script I have and I don't see how it compares the current ip address to the last one. > -- > Rick Nelson > I stopped a long time ago to try to find anything in the bug list of dpkg. > We should run for an entry in the Guinness Book of Records. > -- Stephane Bortzmeyer -- David Fries <[EMAIL PROTECTED]> http://fries.net/~david/ (PGP encryption key available) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#442375: libapr1-dev: update --includedir path, apr-1 not apr-1.0
On Sun, Sep 16, 2007 at 09:31:30PM +0200, Stefan Fritsch wrote: > severity 442375 wishlist > thanks > > On Saturday 15 September 2007, [EMAIL PROTECTED] wrote: > > The header files are installed into /usr/include/apr-1.0 but as the > > version is 1.2.7 I think it is just a hangover from a previous > > version of apr as the package conflicts with libapr1.0-dev which is > > an old version and I assume where that version came from. ?If the > > package source is installed outside of the debian package it is > > installed into /usr/include/apr1 which makes me further think it > > should be installed to /usr/include/apr1 and it better matches the > > current package name libapr1-dev. > > Shouldn't other projects use > > apr-1-config --includedir > > anyway, instead of hardcoding a path? It is code not under my control in the Second Life viewer (Linden Labs). I suggested to the second life mailing list to use, pkg-config --cflags apr-1 Flame Bait, but if you aren't supposed to use, #include or #include then why put the header files in /usr/include for the compiler to find? Not really looking for an answer, thanks for looking into the original bug. -- David Fries <[EMAIL PROTECTED]> http://fries.net/~david/ (PGP encryption key available) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#404918: jpilot: missing showpage on postscript output
Package: jpilot Severity: important Tags: patch When printing a single page memo ghostscript outputs nothing at all. The end END command does not issue showpage so ghostscript discards the output that it does have and does not generate any output. The x11 ghostscript device displays the page. cat memo.ps | /usr/bin/gs -q -dSAFER -dNOPAUSE -r360 -sDEVICE=bjc800 -dMonochromePrint=true -sOutputFile=- - -c quit |wc --- print.c.origSat Nov 11 07:18:14 2006 +++ print.c Fri Dec 29 00:11:13 2006 @@ -946,7 +946,7 @@ int print_address_header() "pop pop pop ( )\n" "} if\n" "("Q_FLAG_CHAR"END) search { %%if\n" - " stop\n" + " showpage stop\n" "} if\n" "("Q_FLAG_CHAR"C12) search {\n" "/Courier-ISOLatin1 12 selectfont\n" -- System Information: Debian Release: 3.0 APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i586) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.4.33.5 Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#375787: sendmail-base: bad dialup.m4
Package: sendmail-base Version: 8.13.7-2 Severity: important For the last four or five days I've been tracking down a problem. I could receive e-mail, but only some of the e-mail would go out. I also saw the following, telnet ::1 25 Trying ::1... Connected to ::1. Escape character is '^]'. 220 ;; ESMTP connection timed out; no servers could be reached Sendmail 8.13.7/8.13.7/Debian-2; Mon, 26 Jun 2006 21:55:05 -0500; (No UCE/UBE) logging access from: ip6-localhost(OK)-ip6-localhost [IPv6:::1] I finally tracked it down to /etc/mail/m4/dialup.m4 (ip replaced with only one change) LOCAL_CONFIG # # # Dynamic host/ip updates from /etc/network/if-up.d/sendmail # # NOTE: the following line *MUST* be in /etc/mail/sendmail.mc dnl include(`/etc/mail/dialup.m4')dnl # This should *NOT* be included in submit.mc ! # # Make sure we accept mail as this ip (for bounces, etc) Cw # # Define our true hostname (from our ISP) - becomes $j define(`confDOMAIN_NAME', `;; connection timed out; no servers could be reached')dnl # # Make sure we accept mail as this name (for bounces, etc) Cw;; connection timed out; no servers could be reached # # Add our hostname to class G for genericstable support CG;; connection timed out; no servers could be reached # I was watching for failed DNS queries, using strace and ltrace to see what sendmail was doing starting up, and finally figured out the above. This is a server and doesn't have dynamically changing ip addresses, so I've disabled dialup.m4 in the sendmail configuration file. Maybe in the future the script could avoid setting confDOMAIN_NAME if the DNS query failed. -- System Information: Debian Release: 3.0 APT prefers unstable APT policy: (500, 'unstable'), (500, 'testing') Architecture: i386 (i586) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.4.26 Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Versions of packages sendmail-base depends on: ii adduser 3.45 Add and remove users and groups ii m41.4-15 a macro processing language ii make 3.79.1-15 The GNU version of the "make" util ii netbase 4.18 Basic TCP/IP networking system ii perl 5.8.7-4Larry Wall's Practical Extraction Versions of packages sendmail-base recommends: pn resolvconf (no description available) Versions of packages sensible-mda depends on: ii libc6 2.3.6-15 GNU C Library: Shared libraries ii procmail 3.22-2 Versatile e-mail processor. ii sendmail-bin [mail-transport- 8.13.7-2 powerful, efficient, and scalable Versions of packages rmail depends on: ii libc6 2.3.6-15 GNU C Library: Shared libraries ii libldap22.1.30-13+b1 OpenLDAP libraries ii sendmail-bin [mail-transpor 8.13.7-2 powerful, efficient, and scalable -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#335203: iftop: correct -N and -F documentation
Package: iftop Version: 0.16-1 Severity: minor Tags: patch -N use for disabling port service name lookups, -F for network filtering. The man page and the usage has -N instead of -F. I corrected the iftop.8 not the iftop.cat as I assumed the iftop.cat was generated from the iftop.8 file. diff -upr iftop-0.16.original/ChangeLog iftop-0.16/ChangeLog --- iftop-0.16.original/ChangeLog 2004-02-28 12:53:37.0 -0600 +++ iftop-0.16/ChangeLog2005-10-22 10:23:27.0 -0500 @@ -4,6 +4,11 @@ $Id: ChangeLog,v 1.23 2004/02/28 18:53:3 Attributions apply to all preceding items up to the next blank line. Unattributed items are by Paul Warren and Chris Lightfoot. +0.?? 02/10/2005 +* Corrected the man page and usage string for the -N and -F options. + -F network filter, -N don't resolve port numbers +David Fries <[EMAIL PROTECTED]> + 0.16 28/02/04 * Added support for DLT_NULL * Fix for pthread.c behaviour on Solaris. diff -upr iftop-0.16.original/iftop.8 iftop-0.16/iftop.8 --- iftop-0.16.original/iftop.8 2003-10-22 14:28:31.0 -0500 +++ iftop-0.16/iftop.8 2005-10-22 10:18:26.0 -0500 @@ -29,9 +29,9 @@ by using the \fB-n\fP option or by press By default, \fBiftop\fP counts all IP packets that pass through the filter, and the direction of the packet is determined according to the direction the packet -is moving across the interface. Using the \fB-N\fP option it is possible to +is moving across the interface. Using the \fB-F\fP option it is possible to get \fBiftop\fP to show packets entering and leaving a given network. For -example, \fBiftop -N 10.0.0.0/255.0.0.0\fP will analyse packets flowing in and +example, \fBiftop -F 10.0.0.0/255.0.0.0\fP will analyse packets flowing in and out of the 10.* network. Some other filter ideas: @@ -257,12 +257,12 @@ frozen. In this case some captured info you, but is included in the totals. A more subtle explanation comes about when running in promiscuous mode -without specifying a \fB-N\fP option. In this case there is no easy way +without specifying a \fB-F\fP option. In this case there is no easy way to assign the direction of traffic between two third parties. For the purposes of the main display this is done in an arbitrary fashion (by ordering of IP addresses), but for the sake of totals all traffic between other hosts is accounted as incoming, because that's what it is from the point of view of your -interface. The \fB-N\fP option allows you to specify an arbitrary network +interface. The \fB-F\fP option allows you to specify an arbitrary network boundary, and to show traffic flowing across it. \fBPeak totals don't add up\fP diff -upr iftop-0.16.original/options.c iftop-0.16/options.c --- iftop-0.16.original/options.c 2004-01-20 05:33:55.0 -0600 +++ iftop-0.16/options.c2005-10-22 10:08:48.0 -0500 @@ -236,7 +236,7 @@ static void usage(FILE *fp) { fprintf(fp, "iftop: display bandwidth usage on an interface by host\n" "\n" -"Synopsis: iftop -h | [-npbBP] [-i interface] [-f filter code] [-N net/mask]\n" +"Synopsis: iftop -h | [-nNpbBP] [-i interface] [-f filter code] [-F net/mask]\n" "\n" " -h display this message\n" " -n don't do hostname lookups\n" -- System Information: Debian Release: 3.0 APT prefers unstable APT policy: (500, 'unstable') Architecture: i386 (i586) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.4.26 Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Versions of packages iftop depends on: ii libc6 2.3.5-6GNU C Library: Shared libraries an ii libncurses5 5.4-9 Shared libraries for terminal hand ii libpcap0.70.7.1-1System interface for user-level pa iftop recommends no packages. -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#313098: dar: can't get --gzip to work
Package: dar Version: 2.2.1-1 Severity: minor I'm having a problem getting the --gzip[level] option to work. Both with or without a space it fails when trying to select the compression level. --gzip 1 The [list of path] is deprecated, thanks to use the -g option instead Writing archive contents... --gzip11 Ignoring unknown option - These options succeed. --gzip -z1 -z --bzip21 -y1 -y -- System Information: Debian Release: 3.0 Architecture: i386 Kernel: Linux SpacedOut 2.4.26 #12 Sat Jun 19 13:45:55 CDT 2004 i586 Locale: LANG=C, LC_CTYPE=C Versions of packages dar depends on: ii libbz2-1.0 1.0.1-14 A high-quality block-sorting file ii libc6 2.3.2.ds1-21 GNU C Library: Shared libraries an ii libdar2 2.2.1-1 Disk ARchive: Shared library ii libgcc1 1:3.4.1-5GCC support library ii libssl0.9.7 0.9.7-4 SSL shared libraries ii libstdc++5 1:3.3.4-9The GNU Standard C++ Library v3 ii zlib1g 1:1.2.1-3compression library - runtime -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#313094: libdar2 >=2.2.1-1 needed
Package: dar Version: 2.2.1-1 Severity: important Package: dar Version: 2.2.1-1 Depends: libbz2-1.0, libc6 (>= 2.3.2.ds1-4), libdar2, libgcc1 (>= 1:3.4.1-3), libssl0.9.7, libstdc++5 (>= 1:3.3.4-1), zlib1g (>= 1:1.2.1) But in reality it depends on libdar2 >=2.2.1-1 (or at least something newer than the 2.1.5-1 that I currently have installed). SpacedOut:~/backup/dar# ldd /usr/bin/dar libdar.so.3 => not found -- System Information: Debian Release: 3.0 Architecture: i386 Kernel: Linux SpacedOut 2.4.26 #12 Sat Jun 19 13:45:55 CDT 2004 i586 Locale: LANG=C, LC_CTYPE=C Versions of packages dar depends on: ii libbz2-1.0 1.0.1-14 A high-quality block-sorting file ii libc6 2.3.2.ds1-21 GNU C Library: Shared libraries an ii libdar2 2.1.5-1 Disk ARchive: Shared library ii libgcc1 1:3.4.1-5GCC support library ii libssl0.9.7 0.9.7-4 SSL shared libraries ii libstdc++5 1:3.3.4-9The GNU Standard C++ Library v3 ii zlib1g 1:1.2.1-3compression library - runtime -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#313062: xawtv: request per channel mixer option
Package: xawtv Version: 3.94-1 Severity: wishlist A per channel mixer selection/volume would be nice. The tuner's audio comes in the video port on my sound card. When I use the composite input selection the audio comes in line in on the sound card. Having one global mixer volume means half the time I can't control the volume from xawtv. It would also seem reasonable if a mixer per channel feature is added to record the volume on a mixer before setting it and leave it at the recorded volume when going to another mixer. That or I would be happy with mute on leaving a channel. I don't have a way to route the composite audio sound through the tuner card so it must get into the computer by a different way than the audio decoded on the tuner. I would expect other people to be in the same boat when selecting between various input types. -- System Information: Debian Release: 3.0 Architecture: i386 Kernel: Linux SpacedOut 2.4.26 #12 Sat Jun 19 13:45:55 CDT 2004 i586 Locale: LANG=C, LC_CTYPE=C Versions of packages xawtv depends on: ii debconf 1.2.35 Debian configuration management sy hi libasound2 1.0.8-3 ALSA library ii libc62.3.2.ds1-21GNU C Library: Shared libraries an ii libfontconfig1 2.3.2-1 generic font configuration library ii libfreetype6 2.1.5-2 FreeType 2 font engine, shared lib ii libjpeg626b-8The Independent JPEG Group's JPEG ii liblircclient0 0.6.4-10LIRC client library ii libncurses5 5.4-3 Shared libraries for terminal hand ii libpng12-0 1.2.8rel-1 PNG library - runtime ii libxaw7 4.1.0-13X Athena widget set library ii libxft2 2.1.1-2 advanced font drawing library for ii libxrandr2 4.3.0.dfsg.1-4 X Window System Resize, Rotate and ii libxrender1 0.8.2-1 X Rendering Extension client libra ii libxv1 4.3.0.dfsg.1-12.0.1 X Window System video extension li ii libzvbi0 0.2.3-1 Video Blank Interval decoder - run pi pia 3.80movie player ii scantv 3.94-1 scan TV channels for stations ii v4l-conf 3.94-1 tool to configure video4linux driv ii xawtv-plugins3.94-1 plugins for xawtv and motv ii xlibmesa3-gl [libgl1 4.2.1-9 Mesa 3D graphics library [XFree86] ii xlibs [libxpm4] 4.2.1-9 X Window System client libraries ii xutils 4.1.0-13X Window System utility programs ii zlib1g 1:1.2.1-3 compression library - runtime -- debconf information: * xawtv/channel-scan: true xawtv/makedev: true * xawtv/freqtab: us-bcast * xawtv/tvnorm: NTSC * xawtv/build-config: true -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#310908: gaim-otr: OTR button vanishes
Package: gaim-otr Version: 2.0.1-1 Severity: normal The OTR button on the conversation button bar vanishes if 'Show button as:' Preferences Conversations option is changed. It will only be visible when the button bar is shown when the conversation is started and changing the show buttons as option will cause it to go away. A related request is to show the OTR button in the Conversation menu. I run with show buttons as: None and this is the first plugin I've found that did not have the functionality available elsewhere. The buttons are rarely used by me and I don't see any reason to take up desktop space with them. -- System Information: Debian Release: 3.0 Architecture: i386 Kernel: Linux SpacedOut 2.4.26 #12 Sat Jun 19 13:45:55 CDT 2004 i586 Locale: LANG=C, LC_CTYPE=C Versions of packages gaim-otr depends on: ii gaim1:1.3.0-1multi-protocol instant messaging c ii libc6 2.3.2.ds1-21 GNU C Library: Shared libraries an ii libgcrypt11 1.2.0-11 LGPL Crypto library - runtime libr ii libotr1 2.0.2-1 Off-the-Record Messaging library -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]