Bug#1076151: firmware-amd-graphics: Please update extremely outdated amdgpu firmware to a new snapshot
Package: firmware-amd-graphics Version: 20230625-2 Severity: important X-Debbugs-Cc: tho...@glanzmann.de Dear Maintainer, *** Reporter, please consider answering these questions, where appropriate *** * What led up to the situation? I updated to Debian trixie. * What exactly did you do (or not do) that was effective (or ineffective)? I upgraded the linux-firmware: git clone https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git cd linux-firmware.git sudo make install sudo reboot *** End of the template - remove these template lines *** -- System Information: Debian Release: trixie/sid APT prefers testing APT policy: (500, 'testing') Architecture: amd64 (x86_64) Kernel: Linux 6.10.0-rc7 (SMP w/32 CPU threads; PREEMPT) Locale: LANG=C, LC_CTYPE=C.utf8 (charmap=UTF-8), LANGUAGE not set Shell: /bin/sh linked to /usr/bin/dash Init: systemd (via /run/systemd/system) firmware-amd-graphics depends on no packages. firmware-amd-graphics recommends no packages. Versions of packages firmware-amd-graphics suggests: ii initramfs-tools 0.142 -- no debconf information Mario Limonciello asked me to file a bug report to update the amdgpu firmware: * Mario Limonciello [2024-07-11 16:57]: > You're not the first person to complain about problems root caused with > extremely outdated firmware in Debian. > Can I ask you to please file a request with Debian to update all of amdgpu > firmware to a new snapshot? > Thanks,
Bug#1071562: Which NFS Version are you using?
Hello, since I have also three productions systems with proxmox and bookworm nfs servers which did not show the issue, I wonder which NFS version you're using? I use nfs version 3. Cheers, Thomas
Bug#1025419: libunwind 1.6.2-2 upgrade makes xorg crash on startup
Hello Axel, > And that fix seems to have been backported in 1.6.2-3 (now in stable, > testing and unstable), see https://bugs.debian.org/1026217 > So this bug report likely can be closed as fixed in libunwind/1.6.2-3 > or (force-) merged with #1026217. (But I'd like to have a second pair > of eyes agreeing that these two issues are indeed the same before > merging or closing.) yes, it was fixed within one week after I opened the bugreport. Cheers, Thomas
Bug#1010725: Since some weeks, gimp icons are displays as identical cross
Hello, I had the same issue. For me librsvg2-common was not installed. After installing it, the issue was gone. sudo apt install -y librsvg2-common Cheers, Thomas
Bug#1026217: libunwind 1.6.2-2.1 assumes 4k page sizes and crashes on systems with bigger page sizes
Package: libunwind8 Version: 1.6.2-2.1 Severity: important Tags: patch X-Debbugs-Cc: as...@lists.linux.dev Hello, the recent libunwind8 update 1.6.2-2.1 breaks for architectures which have a pagesize bigger than 4096 on arm64. The issue is fixed upstream. But no release has been made with the patch included. libunwind8 1.6.2-2.1 crashes Xorg on startup for systems with a pagesize bigger than 4096 on arm64. Linux on apple silicon (m1/m2) uses a 16 KB page size. https://github.com/libunwind/libunwind/commit/2d004eafc77f3c6a4bd9a44b1c35735273fd4e97 Please apply the above patch. Cheers, Thomas -- System Information: Debian Release: bookworm/sid APT prefers testing APT policy: (500, 'testing') Architecture: arm64 (aarch64) Kernel: Linux 6.1.0-asahi (SMP w/8 CPU threads) Kernel taint flags: TAINT_CPU_OUT_OF_SPEC Locale: LANG=C, LC_CTYPE=C.UTF-8 (charmap=UTF-8), LANGUAGE not set Shell: /bin/sh linked to /usr/bin/dash Init: systemd (via /run/systemd/system) Versions of packages libunwind8 depends on: ii libc6 2.36-6 ii liblzma5 5.2.9-0.0 libunwind8 recommends no packages. libunwind8 suggests no packages. -- no debconf information
Bug#1025419: libunwind 1.6.2-2 upgrade makes xorg crash on startup
Hello, the issue is already fixed upstream in libunwind. Cheers, Thomas
Bug#1025419: libunwind 1.6.2-2 upgrade makes xorg crash on startup
Hello, at least the problem on apple silicon systems is due to libunwind assuming 4k page size. Find the patch by Daniel Moody here: https://tg.st/u/0001-libunwind-1.6.2-dynamic-page-size.patch Janne Grunau pointed me to it. Cheers, Thomas
Bug#1025419: libunwind 1.6.2-2 upgrade makes xorg crash on startup
Hello, running Debian testing on apple silicon had the same issue. The culprit seems to be that libunwind was recently updated to 1.6.2-2 which triggers the issue for us. Downgrading to the old version works for us: wget http://ftp.de.debian.org/debian/pool/main/libu/libunwind/libunwind8_1.3.2-2_arm64.deb sudo dpkg -i libunwind8_1.3.2-2_arm64.deb I also tried to build Xorg against the new libunwind, which did __not__ resolve the issue. Output when libunwind is upgraded to 1.6.2-2: (EE) (EE) Backtrace: (EE) 0: /usr/lib/xorg/Xorg (OsLookupColor+0x188) [0xe58f96c8] (EE) unw_get_proc_info failed: no unwind info found [-10] (EE) (EE) Segmentation fault at address 0x0 (EE) Fatal server error: (EE) Caught signal 11 (Segmentation fault). Server aborting (EE) (EE) Please consult the The X.Org Foundation support at http://wiki.x.org for help. (EE) Cheers, Thomas
Bug#1005729: Automatic partitioning fails on arm64 d-i with error message partman: mkfs.vfat: Not enough or too many clusters for filesystem
Source: partman-efi X-Debbugs-Cc: tho...@glanzmann.de Version: 94 Severity: normal Tags: d-i patch Hello, when installing Debian on an arm64 m1 machine, I get the following error during auto partitioning or when creating a EFI System Partition with more than 256 MB: partman: mkfs.vfat: Not enough or too many clusters for filesystem - try less or more sectors per cluster When I try to manually create mkfs.vfat on the same system it works: (air) [~] lsblk NAMEMAJ:MIN RM SIZE RO TYPE MOUNTPOINTS nvme0n1 259:00 465.9G 0 disk ├─nvme0n1p1 259:10 500M 0 part ├─nvme0n1p2 259:20 371.6G 0 part ├─nvme0n1p3 259:30 2.3G 0 part ├─nvme0n1p4 259:40 1G 0 part /boot/efi ├─nvme0n1p5 259:50 85.5G 0 part / └─nvme0n1p6 259:60 5G 0 part nvme0n2 259:70 3M 0 disk nvme0n3 259:80 128M 0 disk (air) [~] umount /boot/efi (air) [~] mkfs.vfat /dev/nvme0n1p4 mkfs.fat 4.2 (2021-01-31) (air) [~] mkfs.vfat -F 16 /dev/nvme0n1p4 mkfs.fat 4.2 (2021-01-31) The sector size is report as 4096: (air) [~] blockdev --getss /dev/nvme0n1 4096 The issue seems to be related to: https://sources.debian.org/src/partman-efi/94/commit.d/format_efi/?hl=15#L66 Since mkfs.vfat don't have issue with creating file systems on block devices with a sector size other than 4096, it is probably best to remove the whole code block that sets the sectors per cluster and sector size. Find attached a patch. -- System Information: Debian Release: bookworm/sid APT prefers testing APT policy: (500, 'testing') Architecture: arm64 (aarch64) Kernel: Linux 5.16.0-asahi-next-20220118-g5fa3c85b5f14 (SMP w/8 CPU threads; PREEMPT) Kernel taint flags: TAINT_CPU_OUT_OF_SPEC Locale: LANG=C, LC_CTYPE=C.UTF-8 (charmap=UTF-8), LANGUAGE not set Shell: /bin/sh linked to /usr/bin/dash Init: systemd (via /run/systemd/system) Report will be sent to Debian Bug Tracking System --- ./wpa_supplicant/doc/docbook/wpa_supplicant.conf.sgml.orig 2022-02-13 09:33:30.099608502 +0100 +++ ./wpa_supplicant/doc/docbook/wpa_supplicant.conf.sgml 2022-02-13 09:33:51.187037059 +0100 @@ -45,8 +45,8 @@ EAP-TLS as work network. -# allow frontend (e.g., wpa_cli) to be used by all users in 'wheel' group -ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel +# allow frontend (e.g., wpa_cli) to be used by all users in 'root' group +ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=root # # home network; allow all valid ciphers network={ @@ -79,7 +79,7 @@ Aegis, Interlink RAD-Series) -ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel +ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=root network={ ssid="example" scan_ssid=1 @@ -101,7 +101,7 @@ -ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel +ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=root network={ ssid="example" scan_ssid=1 @@ -123,7 +123,7 @@ authentication -ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel +ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=root network={ ssid="1x-test" scan_ssid=1 @@ -148,7 +148,7 @@ use. -ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel +ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=root network={ ssid="example" scan_ssid=1 @@ -178,7 +178,7 @@ (-Dwired or -Droboswitch on command line). -ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel +ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=root ap_scan=0 network={ key_mgmt=IEEE8021X --- format_efi.orig 2022-02-14 04:42:29.023205868 +0100 +++ format_efi 2022-02-14 04:46:31.407838463 +0100 @@ -62,22 +62,8 @@ db_progress START 0 3 partman/text/formatting db_progress INFO $template db_progress SET 1 - log_sector_size="$(blockdev --getss "$(cat device)")" - if [ "$log_sector_size" = 512 ]; then - mkdosfs_opts= - else - # mkdosfs has trouble handling cluster - # calculations for non-512-byte logical - # sectors. Forcing one sector per cluster - # avoids this as long as the filesystem - # isn't too large, but that shouldn't be a - # problem for EFI System Partitions. - mkdosfs_opts='-s 1' - fi if log-output -t partman --pass-stdout \ $MKFS_FAT -F "${new_efi_fs#fat}" \ --S "$log_sector_size" \ -$mkdosfs_opts \ "$device" >/dev/null; then sync status=OK
Bug#1005639: wpa_supplicant.conf(5) has an example with group wheel which make it fail on Debian
Package: wpasupplicant X-Debbugs-Cc: tho...@glanzmann.de Version: 2:2.9.0-23 Severity: wishlist Tags: patch Hello, wpa_supplicant.conf(5) manpage includes multiple examples with group wheel. Group wheel does not exist on Debian as a result the example fails. It should either be changed to root or completly removed. The line appears multiple times in the man page: ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel Find attached a patch Cheers, Thomas --- ./wpa_supplicant/doc/docbook/wpa_supplicant.conf.sgml.orig 2022-02-13 09:33:30.099608502 +0100 +++ ./wpa_supplicant/doc/docbook/wpa_supplicant.conf.sgml 2022-02-13 09:33:51.187037059 +0100 @@ -45,8 +45,8 @@ EAP-TLS as work network. -# allow frontend (e.g., wpa_cli) to be used by all users in 'wheel' group -ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel +# allow frontend (e.g., wpa_cli) to be used by all users in 'root' group +ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=root # # home network; allow all valid ciphers network={ @@ -79,7 +79,7 @@ Aegis, Interlink RAD-Series) -ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel +ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=root network={ ssid="example" scan_ssid=1 @@ -101,7 +101,7 @@ -ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel +ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=root network={ ssid="example" scan_ssid=1 @@ -123,7 +123,7 @@ authentication -ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel +ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=root network={ ssid="1x-test" scan_ssid=1 @@ -148,7 +148,7 @@ use. -ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel +ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=root network={ ssid="example" scan_ssid=1 @@ -178,7 +178,7 @@ (-Dwired or -Droboswitch on command line). -ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel +ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=root ap_scan=0 network={ key_mgmt=IEEE8021X
Bug#997981: subject
Hello, I have the same issue. I'm on Debian 10 amd64 with 0.12.2-3. I also tried 1.1.2-2~bpo10+1. This issue is related with something Letsencrypt changed. The last Letsencrypt Certificate was from 8th October. Tonight I renewed my Letsencrypt Certificate autoamtically. After that before the login promot, ocserv was crashing. From the client it looked like that: (nuc) [~] openconnect vpn.company.com POST https://vpn.company.com/ Connected to 1.2.3.4:443 SSL negotiation with vpn.company.com Connected to HTTPS on vpn.company.com with ciphersuite (TLS1.3)-(ECDHE-SECP256R1)-(RSA-PSS-RSAE-SHA256)-(AES-256-GCM) Error reading HTTP response: Invalid argument GET https://vpn.company.com/ Connected to 1.2.3.4:443 SSL negotiation with vpn.company.com Connected to HTTPS on vpn.company.com with ciphersuite (TLS1.3)-(ECDHE-SECP256R1)-(RSA-PSS-RSAE-SHA256)-(AES-256-GCM) Error reading HTTP response: Invalid argument Failed to obtain WebVPN cookie >From the server site it looked like that: Dec 7 04:00:58 debian ocserv[6166]: main: main.c:983: Child 6178 died with sigsegv ... 180 ... similiar entries skipped. I was able to restore operation by compiling ocserv from source: sudo apt-get build-dep -y ocserv wget https://www.infradead.org/ocserv/download/ocserv-1.1.5.tar.xz tar xfJ ocserv-1.1.5.tar.xz cd ocserv-1.1.5 sudo mkdir -p /local/ocserv sudo chown /local/ocserv ./configure --prefix=/local/ocserv make make instsall sudo /etc/init.d/ocserv stop sudo /local/ocserv/sbin/ocserv -c /etc/ocserv/ocserv.conf However I'll upgrade to Debian 11 tonight. Debian 11 doesn't have this problem, because I have several other ocserv on Debian 11, which don't have the issue. Cheers, Thomas
Bug#927242: linphone unable to register account
Hello, I have the same problem. After trying for two hours, I found a way to register an account: - Run linphonec - Run proxy add Than the account also shows up in linphone. Example: linphonec> proxy add Adding new proxy setup. Hit ^D to abort. Enter proxy sip address: [2003:...::1]:5060 Your identity for this proxy: sip:614720-4@[2003:...::1] Do you want to register on this proxy (yes/no): yes Specify register expiration time in seconds (default is 600): Expiration: 600 seconds Specify route if needed: No route specified. sip address: sip:[2003:...::1]:5060 route: identity: sip:614720-4@[2003:...::1] register: yes expires: 600 registered: no Accept the above proxy configuration (yes/no) ?: yes Proxy added. linphonec> Registration on sip:[2003:...::1]:5060 failed: Unauthorized linphonec> Password for 614720-4 on my.host.nane: password Refreshing on sip:614720-4@[2003:...::1]... linphonec> Registration on sip:[2003:...::1]:5060 successful. Cheers, Thomas
Bug#271384: MAXSTR issue
Hello Jan, > I did a quick estimate by resizing my xterm to maximum (on 1600x1200 > with a rather tiny font, yielding 264 columns) and multiplying by 2 > for a possible xinerama/dualhead setup. The outcome was more than > 512, so I increased it to 768. I see. 512 is enough for my 1920x1200 btw. However. MAXSTR does have an effect on every buffer allocated in screen which basically means that screen blows up in size: My main screen for example has 217Mbyte _RSS_ for 44 windows and a scrollback buffer of 5000 lines per window. (mephisto) [~] ps axxuww | head -1; ps axuwww | grep sithglan | grep SCREEN USER PID %CPU %MEMVSZ RSS TTY STAT START TIME COMMAND sithglan 18194 0.0 21.0 221424 217812 ? Ss Jul22 3:58 bin/SCREEN -S main -c .screenrc-main Never the less increasing MAXSTR is a big step into the right direction. > If it's that much of an issue, I'll change these parts of the code to > allocate the buffers dynamically - but I don't think it's worth the > effort. There is more broken than that. For exmaple the following: Use a caption like the following: caption always '%H%{-b}%{= wb} %-Lw%{= wr}%50>%{+b}%n%f %t%{= wb}%{-b}%+Lw%{= wr}%-017= %c %Y-%m-%d' and open 40 windows (I have that in the configile, too) Start screen in a xterm (80x25) now dettach and reattach on a xterm running fullscreen on 1600x1200 => BOOM => Segfault (you have to open a few windows to see that effect: like 40 or so) But if you start screen in a Fullscreen xterm in the first place you never get the SEGAULT. So screen seems to have a look at the wide of the terminal you start it initial and uses that information to allocate internal buffers. I never tracked it down (but it should be really easy to track it down) because it always happens to me when I reattach my (at work started screen) at home (I have bigger screen at home). Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#271384: closed by Jan Christoph Nordholz <[EMAIL PROTECTED]> (Bug#271384: fixed in screen 4.0.3-2)
Hello, to what value did you increase MAXSTR? Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#76803: screen swiveling vs ssh agent
Hello, this is very easy to solve: ssh-agent | head -2 > .ssh/env . .ssh/env ssh-add screen -S main -c .screenrc-main => Problem solved. At least in practice. am I the only screen poweruser? Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#348657: screen doesn't spawn login shells
Hello, please close 348657 It is fixed since ages. Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#408941: artsd fails to start if you call it from command line using fvwm2 as window manager
Hello, I found the problem. You have to start kded beforehand manually. After that artsd starts. In woody this happened implcitly. However the new artsdsp doesn't work as good as the old release did, so it is useless for my current application. I think we can kick this bugreport. I have to look for another solution to my problem. Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#408941: artsd fails to start if you call it from command line using fvwm2 as window manager
Package: arts Version: 1.5.5-1 Severity: serious Hello, I used arts a lot on debian woody in the following way: window 1: artsd window 2: artsdsp -m /path/to/first/application/with/oss_interface window 3: artsdsp -m /path/to/second/application/with/oss_interface Since the update to etch artsd doesn't work anylonger in this way. When I try to start artsd in my environment which doesn't uses kde it fails in the following way: (faui00w) [~] artsd Creating link /home/cip/adm/sithglan/.kde/socket-faui00w. can't create mcop directory (faui00w) [~] ls -al /home/cip/adm/sithglan/.kde/socket-faui00w ls: /home/cip/adm/sithglan/.kde/socket-faui00w: No such file or directory (faui00w) [~] ls -al ~/.kde/ total 30 drwx-- 2 sithglan icipguru 512 Oct 3 18:20 ./ drwxr-xr-x 197 sithglan immdstud 29184 Jan 29 14:38 ../ What do I have to do to get it back working? Thomas -- System Information: Debian Release: 4.0 APT prefers testing APT policy: (1050, 'testing'), (500, 'testing') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.19.2 Locale: LANG=C, [EMAIL PROTECTED] (charmap=ISO-8859-15) Versions of packages arts depends on: ii libarts1c2a 1.5.5-1aRts sound system core components ii libartsc0 1.5.5-1aRts sound system C support librar arts recommends no packages. -- debconf-show failed -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#298121: mutt: header cache patch should use realpath for maildir
Hello, > Also consider symlinks in the path, etc. Really, your best approach as > far as I understand is to use realpath() on the filename/path of the > mailbox before md5-ing it. yes, it is. Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#364288: Randomly loses input since X11R7 upgrade
Hello Tore, I see the same problem on my mac mini using a self compiled X using usb keyboard and usb mouse. My solution so far for it is pulling the usb keyboard out and putting it after a second or so back in. After that my usb keyboard starts working again. I know that my pointer works to, but I didn't check the buttons. If you find anything further usefull please let me know. Next time it hangs, I try if magic sysrq works for me, too. (Nice trick with unrar and giving the kernel back the vt switching). Btw. when I do this on a working Xserver I simply can switch back to vt 5 (this is where my Xserver runs) and it seems to switch keyboard back to raw modus. Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#365733: watchdog: Option ping-count doesn't work and isn't documented
Package: watchdog Version: 5.2.4-4 Severity: important Hello everyone, the option ping-count isn't documented and doesn't work. Per default watchdog tried three times to ping a host. With the option ping-count it is supposed to change this number via the configfile. This is important because my server just rebooted because of this check failed because I guess 3 times wasn't enough. However the source code has this option in, but watchdog isn't able to parse the line because it is tested after the "ping" option and since "ping" is a prefix to "ping-count" it tries to parse "ping-count" as "ping" and fails. The fix is obvious: First check for "ping-count" than for "ping", a patch is attached. Thanks, Thomas --- a/src/watchdog.c +++ b/src/watchdog.c @@ -295,6 +295,11 @@ fprintf(stderr, "Ignoring invalid line in config file:\n%s\n", line); else add_list(&pidfile, strdup(line + i)); + } else if (strncmp(line + i, PINGCOUNT, strlen(PINGCOUNT)) == 0) { + if (spool(line, &i, strlen(PINGCOUNT))) + fprintf(stderr, "Ignoring invalid line in config file:\n%s\n", line); + else + pingcount = atol(line + i); } else if (strncmp(line + i, PING, strlen(PING)) == 0) { if (spool(line, &i, strlen(PING))) fprintf(stderr, "Ignoring invalid line in config file:\n%s\n", line); @@ -398,11 +403,6 @@ else { minpages = atol(line + i); } - } else if (strncmp(line + i, PINGCOUNT, strlen(PINGCOUNT)) == 0) { - if (spool(line, &i, strlen(PINGCOUNT))) - fprintf(stderr, "Ignoring invalid line in config file:\n%s\n", line); - else - pingcount = atol(line + i); } else { fprintf(stderr, "Ignoring invalid line in config file:\n%s\n", line); } -- System Information: Debian Release: 3.1 Architecture: i386 (i686) Kernel: Linux 2.6.17-rc3 Locale: LANG=C, [EMAIL PROTECTED] (charmap=ISO-8859-15) Versions of packages watchdog depends on: ii libc6 2.3.2.ds1-22sarge3 GNU C Library: Shared libraries an ii makedev 2.3.1-77 creates device files in /dev -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#361835: watchdog: Problem with the option interface
Hello Frederic, this part of the watchdog binary opens the file /proc/net/dev and looks if the byte counter of incomming and extern traffic changes, if it doesn't it resets the machine. This happened once for me with a sky networkcard (but I use watchdog only since last night). I guess that my networkcard driver doesn't increment the byte counter if it reached a 2Gbyte mark or something, but I will track it down. You should monitor /proc/net/dev as well manually and look out if the counters increment. Maybe add some debug code. Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#348657: Before the last update of screen you could get login shells per default by sepcifing 'shell -/bin/bash' this does no longer work which is very annoying
Hello, > What does ps say about the bash process that is spawned by screen? > It should be listed as -/bin/bash. yes, it is. But it isn't a login shell. That is for sure. I even tried it with a clean environment. What is the md5sum of your screen, if you can't reproduce it? (travelmate) [~] apt-cache policy screen screen: Installed: 4.0.2-4.1 Candidate: 4.0.2-4.1 Version table: *** 4.0.2-4.1 0 500 http://ftp2.de.debian.org testing/main Packages 100 /var/lib/dpkg/status (travelmate) [~] md5sum /usr/bin/screen 192288cde7fb2d99a1145d20a7de0082 /usr/bin/screen Greetings, Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#348657: Before the last update of screen you could get login shells per default by sepcifing 'shell -/bin/bash' this does no longer work which is very annoying
Hello, > I can't reproduce this. Let's try again on testing: (travelmate) [~] cat .screenrc-debian shell -/bin/bash (travelmate) [~] screen -c .screenrc-debian => in screen: [EMAIL PROTECTED]:~$ suspend # => Note the different prompt (which # indicates that my .bash_profile doesn't get sourced) # Shell suspends and gets restarted (login bashs can only be # suspended using suspend -f) Screen gives me: Child has been stopped, restarting. and now again on sarge: (faui01) [~] cat .screenrc-debian shell -/bin/bash (faui01) [~] screen -c .screenrc-debian => in screen: (faui01) [~] suspend -/bin/bash: suspend: cannot suspend a login shell Are you able to reproduce that? Greetings, Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#348661: vim doesn't handle its runtime files anymore: syntax highligt for help and in general is broken, it doesen't decoce gziped file per default and so on
Hello, > please just verify that your /etc/vimrc is up to date : verify that > the one in the package has been merged, and is not stayed under > /etc/vim/vimrc.dpkg-new. if yes, please close that bug, and fix that > yourself (and/or open a bug on dpkg which is here the one responsible) that was it: (travelmate) [~] ls -al /etc/vim/vimrc* -rw-r--r-- 1 root root 2843 Sep 1 18:45 /etc/vim/vimrc -rw-r--r-- 1 root root 2802 Nov 27 08:40 /etc/vim/vimrc.dpkg-dist (travelmate) [~] mv /etc/vim/vimrc.dpkg-dist /etc/vim/vimrc mv: overwrite `/etc/vim/vimrc', overriding mode 0644? (travelmate) [~] / !! / mv /etc/vim/vimrc.dpkg-dist /etc/vim/vimrc thank you for your immediate response. Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#348661: vim doesn't handle its runtime files anymore: syntax highligt for help and in general is broken, it doesen't decoce gziped file per default and so on
Hello, > Could you please give us the output of "ls -l /usr/bin/vim*"? I fear > there went something wrong when upgrading the package from a version > which used diversions to a newer package which now uses alternatives. lrwxrwxrwx 1 root root 21 Jan 15 13:33 /usr/bin/vim -> /etc/alternatives/vim* -rwxr-xr-x 1 root root 1090476 Dec 23 01:54 /usr/bin/vim.basic* -rwxr-xr-x 1 root root 1505828 Dec 23 01:54 /usr/bin/vim.full* -rwxr-xr-x 1 root root 1336524 Dec 23 01:54 /usr/bin/vim.gtk* lrwxrwxrwx 1 root root 3 Jan 15 13:32 /usr/bin/vimdiff -> vim* -rwxr-xr-x 1 root root1600 Dec 23 01:53 /usr/bin/vimtutor* and /etc/alternatives/vim points to /usr/bin/vim.full Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#348661: vim doesn't handle its runtime files anymore: syntax highligt for help and in general is broken, it doesen't decoce gziped file per default and so on
Package: vim Version: 1:6.4-006+1 Severity: serious Tags: experimental Hello, please test your packages before publishing them. :syn on gives me for example the following: Error detected while processing /usr/share/vim/vim64/syntax/syntax.vim: line 42: E216: No such group or event: filetypedetect BufRead travelmate:~# dpkg --get-selections \* | grep vim vim install vim-common install vim-fullinstall vim-gtk install vim-gui-common install vim-runtime install It doesn't work for vim.full nor vim.basic. Thomas -- System Information: Debian Release: testing/unstable APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.12-1-386 Locale: LANG=C, [EMAIL PROTECTED] (charmap=ISO-8859-15) Versions of packages vim depends on: ii libc62.3.5-8 GNU C Library: Shared libraries an ii libgpmg1 1.19.6-21 General Purpose Mouse - shared lib ii libncurses5 5.5-1 Shared libraries for terminal hand ii vim-common 1:6.4-006+1 Vi IMproved - Common files ii vim-runtime 1:6.4-006+1 Vi IMproved - Runtime files vim recommends no packages. -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#348657: Before the last update of screen you could get login shells per default by sepcifing 'shell -/bin/bash' this does no longer work which is very annoying
Package: screen Version: 4.0.2-4.1 Severity: important Tags: experimental Hello, before the last update specifing 'shell -/bin/bash' in my .screenrc gave me login shells (this happens also if I compile screen by myself). With current available screen package for testing, it doesn't work anylonger and I have to specify manual that screen should spawn login shells: screen -t l55 /bin/bash --login which is very annoying when I want to create a new window with a login shell, which I want all the time. -- System Information: Debian Release: testing/unstable APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.12-1-386 Locale: LANG=C, [EMAIL PROTECTED] (charmap=ISO-8859-15) Versions of packages screen depends on: ii base-passwd 3.5.11 Debian base system master password ii debconf 1.4.67 Debian configuration management sy ii libc6 2.3.5-8GNU C Library: Shared libraries an ii libncursesw5 5.5-1 Shared libraries for terminal hand ii libpam0g 0.79-3 Pluggable Authentication Modules l ii passwd1:4.0.13-7 change and administer password and screen recommends no packages. -- debconf information: screen/old_upgrade_prompt: false -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#295249: qdbm debian package
Hello UKAI, I wanted to ask, if there is any progress in the qdbm debian packages? Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#307216: xterm colorization problems
Package: xterm Version: 4.3.0.dfsg.1-12.0.1 Severity: important Tags: patch Xterm has some colorization problems which are pop-up for example if you read your eMail in mutt in a screen. The problem is: From: Thomas Dickey <[EMAIL PROTECTED]> To: Thomas Glanzmann <[EMAIL PROTECTED]> Subject: Re: Color problems with xterm-201 Date: Sun, 1 May 2005 15:31:09 -0400 I found the problem (took about an hour). It was when xterm was writing out pending data, updated the display's colors and didn't restore. That happened to be in one of those rarely-needed checks - it had some data leftover telling it to scroll just as it started writing new data. This fragment in charproc.c: if (!AddToRefresh(screen)) { /* make sure that the correct GC is current */ currentGC = updatedXtermGC(screen, flags, fg_bg, False); if (screen->scroll_amt) FlushScroll(screen); doesn't account for that one of the functions called by FlushScroll() could modify the display's colors as is done in updatedXtermGC(). What's odd is that the code is very old - and no one reported it before. I see the updatedXtermGC() call from 1996. (On the other hand, I've encountered bugs that old more than once). Anyway, the fix for that is to move the assignment past the if statement: if (!AddToRefresh(screen)) { if (screen->scroll_amt) FlushScroll(screen); /* make sure that the correct GC is current */ currentGC = updatedXtermGC(screen, flags, fg_bg, False); I'll check for other occurrences, of course. The attached patch fixes the described problem. Please apply and recompile. --- xterm-201/charproc.c~redraw +++ xterm-201/charproc.c @@ -3435,12 +3435,13 @@ InsertChar(screen, cells); } if (!AddToRefresh(screen)) { - /* make sure that the correct GC is current */ - currentGC = updatedXtermGC(screen, flags, fg_bg, False); if (screen->scroll_amt) FlushScroll(screen); + /* make sure that the correct GC is current */ + currentGC = updatedXtermGC(screen, flags, fg_bg, False); + if (flags & INVISIBLE) { if (cells > len) { str = temp_str = malloc(cells); -- System Information: Debian Release: 3.1 APT prefers testing APT policy: (1050, 'testing'), (500, 'testing') Architecture: i386 (i686) Kernel: Linux 2.4.30 Locale: LANG=C, [EMAIL PROTECTED] (charmap=ISO-8859-15) Versions of packages xterm depends on: ii libc62.3.2.ds1-21GNU C Library: Shared libraries an ii libexpat11.95.8-1XML parsing C library - runtime li ii libfontconfig1 2.3.1-2 generic font configuration library ii libfreetype6 2.1.7-2.4 FreeType 2 font engine, shared lib ii libice6 4.3.0.dfsg.1-12.0.1 Inter-Client Exchange library ii libncurses5 5.4-4 Shared libraries for terminal hand ii libsm6 4.3.0.dfsg.1-12.0.1 X Window System Session Management ii libxaw7 4.3.0.dfsg.1-12.0.1 X Athena widget set library ii libxext6 4.3.0.dfsg.1-12.0.1 X Window System miscellaneous exte ii libxft2 2.1.7-1 FreeType-based font drawing librar ii libxmu6 4.3.0.dfsg.1-12.0.1 X Window System miscellaneous util ii libxpm4 4.3.0.dfsg.1-12.0.1 X pixmap library ii libxrender1 0.8.3-7 X Rendering Extension client libra ii libxt6 4.3.0.dfsg.1-12.0.1 X Toolkit Intrinsics ii xlibs4.3.0.dfsg.1-12 X Keyboard Extension (XKB) configu ii xlibs-data 4.3.0.dfsg.1-12 X Window System client data -- debconf-show failed -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#298121: mutt: header cache patch should use realpath for maildir
Hello Dato, * Adeodato Simó <[EMAIL PROTECTED]> [050310 07:10]: > Just in case you didn't spot it already. I described from the lists. > > If I open a maildir using mutt's mailboxes list, and then open it using > > "mutt -f Mail/drow", the header cache creates a second cache. They're > > the same mailbox, but one is "Mail/drow" and the other is > > "/home/drow/Mail/drow". I could hack something up next monday to prepend cwd if mailbox path doesn't start with a '/'. > > This could conceivably lead to false matches, also, since it uses relative > > paths. In praxis not. Greetings, Thomas
Bug#277864: Using valgrind to track it down
Hello, > Hi Thomas. I don't know why, but seemed to have missed this message > when you sent it first. I just found it today when examining this bug > report via the web. I am running valgrind now. For how long should I > run mutt? Will have any way of knowing if the crash would have happened > normally? I only just started using valgrind a few days ago for a > project of my own, and I don't know a whole lot about it yet. no there is no indicator. AFAIK. Maybe in the log in /tmp Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#262272: I think this is the symtom
Hello, ==6299== Invalid read of size 1 ==6299==at 0x1B904A33: strcmp (mac_replace_strmem.c:250) ==6299==by 0x807581D: hash_find_hash (hash.c:104) ==6299==by 0x80A8E90: mutt_sort_threads (thread.c:782) ==6299==by 0x80A6C39: mutt_sort_headers (sort.c:287) ==6299==by 0x8061F24: update_index (curs_main.c:326) ==6299==by 0x80685A3: mutt_index_menu (curs_main.c:501) ==6299==by 0x807C01F: main (main.c:936) ==6299== Address 0x1C072300 is not stack'd, malloc'd or (recently) free'd ==6299== ==6299== Invalid read of size 1 ==6299==at 0x1B904A33: strcmp (mac_replace_strmem.c:250) ==6299==by 0x807581D: hash_find_hash (hash.c:104) ==6299==by 0x80A8C21: mutt_sort_threads (thread.c:917) ==6299==by 0x80A6C39: mutt_sort_headers (sort.c:287) ==6299==by 0x8061F24: update_index (curs_main.c:326) ==6299==by 0x80685A3: mutt_index_menu (curs_main.c:501) ==6299==by 0x807C01F: main (main.c:936) ==6299== Address 0x1C072300 is not stack'd, malloc'd or (recently) free'd this might be the problem. Could you produce a coredump and make it available to us. The other valgrind log would be helpful, too. Greetings, Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#262272: mutt: segfault while sorting mailbox
Hallo, > Just FYI, You undo Ctrl-S with Ctrl-Q. Screen has it's own flow control in addition to the terminals inside screen: Ctrl-A Ctrl-Q or Ctrl-A q Greetings, Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#262272: mutt: segfault while sorting mailbox
Hello, thanks for all the data. But we have a problem. The mutt binary has the debugging symbols stripped. We have to ask dato to build a mutt package with the debugging symbols in and let you re-valgrind it. I can say so far that we have at least a few invalid free(s). This could be even header cache related. And access to unallocated memory. I ask dato today to build a new package and send it over to you. Greetings, Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#262272: mutt: segfault while sorting mailbox
Hello, > Thomas: should I mail you as well or is mailing [EMAIL PROTECTED] > enough? Currently I'm group replying. [EMAIL PROTECTED] is enough. Greetings, Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#262272: mutt: segfault while sorting mailbox
Hello, the backtrace isn't very helpful expect that it indicates that already freed or never allocated memory was free'ed. This could be a problem of the header cache for example. > # Note from Maurits: "Invoer-/uitvoerfout" is Dutch for: Input/output > # error. you can do a export LD_LIBRARY_PATH=/usr/lib/debug export LANG=C gdb /path/to/mutt than we get this errors in english, too. Greetings, Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#262272: mutt: segfault while sorting mailbox
Hello, > I am using 1.5.6 from SID and I am connecting via SIMAP > to and the Server is a courier-imap-ssl could you please create a backtrace (see previous eMails) and/or run valgrind on it? Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#262272: mutt: segfault while sorting mailbox
Hello, > I also get segfaults from mutt several times a day. It always seems to > happen just after mutt displays a message "Sorting mailbox". This also > happens when I don't do anything. Mutt is running, I go to sleep and > next morning mutt has segfaulted. And no signs of fairies or grues. ;) > It only happens when I am in an IMAP folder. Which IMAP Server do you use? > I saved 23 core dumps of about 1.5 MB each. I don't have experience > analyzing them, but they are available in case anyone would want to > have a look. With some pointers I may be able to analyze it myself. For starters: gdb /usr/bin/mutt /path/to/core bt quit But if you produce this coredump on Sarge it is very likely that it is useless because the framepointers are stripped from the libs. If the backtrace of the above has many '???' in it produce a new coredump with libs which have framepointers in: apt-get install libc6-dbg # Run mutt with LD_LIBRARY_PATH=/usr/lib/debug set LD_LIBRARY_PATH=/usr/lib/debug mutt # GDB, too LD_LIBRARY_PATH=/usr/lib/debug gdb /usr/bin/mutt /path/to/core bt quit You can also run it in valgrind (if it runs in valgrind it might be possible that it doesn't crash anylonger) However invalid memory accesses and 'strange behaviour' will be logged in a logfile: Oh and mutt will be a bit slower using valgrind because valgrind is kind a virtual machine actually. apt-get install valgrind valgrind --tool=memcheck --run-libc-freeres=no --num-callers=10 \ --show-reachable=yes --leak-check=yes --error-limit=no \ --log-file=/tmp/valgrind /usr/bin/mutt After quiting mutt you will find a /tmp/valgrind.pid logfile. You can send this privately to me or make it accessible via a webserver (it might be very big). > Yesterday I tried "set strict_threads=yes" like was proposed earlier > in the thread as a workaround. Since then I haven't seen a > segfault. That post mentions that there may be unbearable side > effects, but I haven't noticed. So for now I am fine, thanks for the > tip. So turn that off again - than it might be possible to find that bug. I also have a ppc32 user who reports the same bug at a hand, I got an account on his machine but was unable to reproduce the bug. He said that it happens at least once a day but with external modification of the IMAP mailbox and after that moving arrow keys (but it might take some time). Maybe you can trigger this segfault faster if you do so. Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#282451: mutt: same problem, but everytime on IMAP+TLS connect attempt
Hello, could you please send me an useful backtrace? apt-get install libc6-dbg LD_LIBRARY_PATH=/usr/lib/debug mutt gdb /usr/bin/mutt attach pid reproduce the problem and type backtrace. :-) Hopefully we get the name of the calling function. Does this happen also with an empty mailbox? Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#290103: readline provided with bash: self-insert doesn't work correctly
Package: bash3 Version: 3.0-10 Severity: normal Tags: patch Hello, having the following in .inputrc: $if Bash ",2": " 2>&1" ",n": " 2>/dev/null" ",t": "/var/tmp/sithglan/" ",b": "/opt/build/thomas/" ",": self-insert $endif should expand ,2 to " 2>&1" but , still to ','. This isn't the case and I asked bash maintainer Chet Ramey to fix it and he did it. :-) The following patch will ship with the next version of bash. > > That is a bug in the released version of readline-5.0. Shadow keymaps > > are implemented by using an extra character. When a `shadowed' key > > (the `,' in your case) is bound to self-insert, the 5.0 code was > > trying to insert the extra character (256) rather than the `,'. *** ../../../bash-20050106.orig/lib/readline/readline.c Fri Oct 15 14:36:55 2004 --- readline.c Sat Jan 8 23:51:44 2005 *** *** 657,665 the input queue with _rl_unget_char. */ { - #if 0 - r = _rl_dispatch (ANYOTHERKEY, FUNCTION_TO_KEYMAP (map, key)); - #else - /* XXX - experimental code -- might never be executed. Save -for later. */ Keymap m = FUNCTION_TO_KEYMAP (map, key); int type = m[ANYOTHERKEY].type; --- 657,660 *** *** 667,673 if (type == ISFUNC && func == rl_do_lowercase_version) r = _rl_dispatch (_rl_to_lower (key), map); else r = _rl_dispatch (ANYOTHERKEY, m); - #endif } else if (r && map[ANYOTHERKEY].function) --- 662,681 if (type == ISFUNC && func == rl_do_lowercase_version) r = _rl_dispatch (_rl_to_lower (key), map); + else if (type == ISFUNC && func == rl_insert) + { + /* If the function that was shadowed was self-insert, we +somehow need a keymap with map[key].func == self-insert. +Let's use this one. */ + int nt = m[key].type; + rl_command_func_t *nf = m[key].function; + + m[key].type = type; + m[key].function = func; + r = _rl_dispatch (key, m); + m[key].type = nt; + m[key].function = nf; + } else r = _rl_dispatch (ANYOTHERKEY, m); } else if (r && map[ANYOTHERKEY].function) -- System Information: Debian Release: 3.1 APT prefers testing APT policy: (1050, 'testing') Architecture: i386 (i686) Kernel: Linux 2.4.29-rc1 Locale: LANG=C, [EMAIL PROTECTED] (charmap=ISO-8859-15) Versions of packages bash3 depends on: ii base-files 3.1.2Debian base system miscellaneous f ii libc6 2.3.2.ds1-20 GNU C Library: Shared libraries an ii libncurses5 5.4-4Shared libraries for terminal hand ii passwd 1:4.0.3-30.7 Change and administer password and -- debconf-show failed -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]