Re: Problems with cygwin in Windows 10
Hi, (This is the same reply I sent when Robert emailed cygwin-xfree rather than cygwin.) On Tue, Jul 21, 2015 at 11:13 PM, Robert McBroom mcfo...@bellsouth.net wrote: Having problems with the X interface in Windows 10 (currently build 10240). /bin/startxwin fails with the symptoms in the attached file xwin.txt. The output from cygcheck -cv is also attached. What am I missing? I have been using Cygwin/X on Windows 10 64-bit without issue, across numerous builds (currently 10240 also.) The 2nd time you ran startxwin, It looks like Cygwin thinks you lack 3D acceleration. Try this command: $ startxwin -- -extension GLX -multiwindow Also, the UnknownGroup might be an issue that prevented you from writing the lock file to /tmp/. Someone else would know better than me though. -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Does /etc/profile need to set MANPATH?
On Mon, Jul 20, 2015 at 8:09 AM, Mikhail Usenko cyg...@inbox.ru wrote: On Sat, 17 May 2014 10:37:30 +0200 Achim Gratz ... wrote: Corinna Vinschen writes: Whatever, it looks like MANPATH can really go away. Achim, do your worst. Done. Besides defining MANPATH (and the other variables such as INFOPATH and PATH which can be modified by the user's .bash_profile) /etc/profile also did export the MANPATH variable. Now it does not and I suppose that user's skeleton file /etc/skel/.bash_profile should provide user-defined MANPATH pro-forma as an exported environment variable or else it will have no effect. -- Mike [...] Why does Cygwin need to define MANPATH by default? Cygwin uses /etc/man_db.conf, just like the Red Hat family of Linux distros. (The Debian family uses /etc/manpath.config). -Mike D -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Problems with cygwin in Windows 10
Hi, On Tue, Jul 21, 2015 at 12:52 AM, Robert McBroom mcfo...@bellsouth.net wrote: Having problems with the X interface in Windows 10 (currently build 10240). /bin/startxwin fails with the symptoms in the attached file xwin.txt. The output from cygcheck -cv is also attached. What am I missing? I have been using Cygwin/X on Windows 10 64-bit without issue, across numerous builds (currently 10240 also.) The 2nd time you ran startxwin, It looks like Cygwin thinks you lack 3D acceleration. Try this command: $ startxwin -- -extension GLX -multiwindow Also, the UnknownGroup might be an issue that prevented you from writing the lock file to /tmp/. Someone else would know better than me though. -Mike -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/
Re: Updated: xorg-server-1.17.2-1 (TEST)
On Thu, Jul 9, 2015 at 12:30 PM, Jon TURNEY jon.tur...@dronecode.org.uk wrote: The following packages have been updated in the Cygwin distribution: *** xorg-server-*1.17.2-1 These packages contain XWin and the other X.Org X11 servers. In addition to upstream fixes [1], the following cygwin-specific changes have been made since 1.17.1-5: * Only apply MINIMIZE or MAXIMIZE styles from XWinrc when mapping a window * Ensure WM_TRANSIENT_FOR windows are always parented * Downgrade a warning when the experimental crashreporter isn't present to debug * When a SEGV occurs checking WGL, disable WGL and continue, rather than exiting * Report invalid options in combination with -nodecoration * Add support for a SKIPTASKBAR style in XWinrc These packages are currently marked as a test release because they contain changes to code critical to the correct functioning of multiwindow mode, so there is a higher than normal risk of regressions. They will be made stable in approximately two weeks if no major regressions are reported. [...] I ran 3 tests and did not notice any regressions: 1. XDMCP into a RHEL 7.1 host with GDM and launch a gnome classic session (this tests GLX) 2. SSH -Y into a Fedora 21 host (with XWin in MultiWindow mode) and launch gnome-terminal 3. SSH -Y into a Fedora 21 host (with XWin in MultiWindow mode) and launch Firefox 38 -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Isuse with xwin-xdg-menu View logfile
On Wed, Jul 8, 2015 at 10:43 AM, Jon TURNEY jon.tur...@dronecode.org.uk wrote: On 07/07/2015 04:13, Michael DePaulo wrote: First of all, let me say how much I like xwin-xdg-menu. Thumbs up to Jon Turney, and to all the other developers like Yaakov Selkowitz who work on related components/package for Cygwin/X. Anyway, when I do the following: $ xinit -- -multiwindow From the xterm that automatically launches: $ xwin-xdg-menu XDG Menu View logfile I then see this: executing 'xterm -title '/dev/pty1' -e less +F /dev/pty1', pid 1292 (pid 1292 stderr) Warning: Missing charsets in String to FontSet conversion (pid 1292 stderr) Warning: Unable to load any usable fontset pid 1292 exited with status 0 (The xterm window flashes briefly, and then disappears.) Is /dev/pty1 the log file we should be trying to open? I am not even sure what log file should be opened. Personally, I would have assumed it is the X Server's log file. That's hard to do with full generality, as we would need to know or discover what -logfile option the server was given. I see that this is the relevant code: https://github.com/jon-turney/xwin-xdg-menu/blob/51ae0ec7f225b8257ba64ce5318b857bf9ce8450/execute.c#L236 I made a small update to today, which includes a change to remove the View logfile menu item if stdout is a tty, which should avoid this problem scenario. -- Jon TURNEY Volunteer Cygwin/X X Server maintainer Thanks. I tested it successfully. And Yaakov, thanks for explaining that. (In a split email thread, Yaakov explained that I should use the XWin Server shortcut, which uses startxwin). -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Return codes over 1 byte
mark06 mentioned this on IRC today and then left the channel about 1 hour later:
mark06 has anyone ever discussed exit codes above one byte? they are
valid on modern windows, but cygwin's bash will mess them
I was curious, so I googled it (I could not find an answer) and then
tried it out. I can confirm the what he said.
Cygwin Bash:
mike@executor ~
$ uname -a
CYGWIN_NT-10.0 executor 2.0.4(0.287/5/3) 2015-06-09 12:22 x86_64 Cygwin
mike@executor ~
$ cat return.c
int main (){ return 512; }
mike@executor ~
$ x86_64-w64-mingw32-gcc return.c -o return.exe
mike@executor ~
$ ./return.exe
mike@executor ~
$ echo $?
0
cmd.exe:
C:\cygwin64\home\mikereturn.exe
C:\cygwin64\home\mikeecho %errorlevel%
512
-Mike
--
Problem reports: http://cygwin.com/problems.html
FAQ: http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Return codes over 1 byte
On Thu, Jul 9, 2015 at 7:42 PM, Yaakov Selkowitz yselkow...@cygwin.com wrote: On Thu, 2015-07-09 at 19:30 -0400, Michael DePaulo wrote: mark06 mentioned this on IRC today and then left the channel about 1 hour later: mark06 has anyone ever discussed exit codes above one byte? they are valid on modern windows, but cygwin's bash will mess them I was curious, so I googled it (I could not find an answer) and then tried it out. I can confirm the what he said. http://pubs.opengroup.org/onlinepubs/9699919799/functions/exit.html https://www.gnu.org/software/libc/manual/html_node/Exit-Status.html -- Yaakov Right, only the least significant 8 bits are outputted. So 257 becomes 1. I thought I read somewhere that the Cygwin mintty terminal + bash shell is supposed to be suitable for running native windows apps. Maybe I was thinking of the 1st paragraph on this page after the list of features: https://code.google.com/p/mintty/ Or maybe I was thinking about this reply: https://cygwin.com/ml/cygwin/2007-03/msg00758.html Either way, I feel like this should be documented somewhere. Perhaps I should submit a patch to add a section like return codes to to this page? https://cygwin.com/cygwin-ug-net/using-effectively.html https://cygwin.com/git/gitweb.cgi?p=newlib-cygwin.git;a=blob;f=winsup/doc/effectively.xml -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: [PATCH mintty] Add an XWin XDG menu entry
On Tue, Jul 7, 2015 at 4:00 PM, Thomas Wolff t...@towo.net wrote: Am 07.07.2015 um 13:47 schrieb Yaakov Selkowitz: On Tue, 2015-07-07 at 11:47 +0200, Thomas Wolff wrote: On 07.07.2015 01:12, Yaakov Selkowitz wrote: --- Thomas, could you please ship a new release with this patch ASAP? Are you sure this is desired? Mintty isn't related to X Windows; Well, not for a full X desktop, and that's the reason for the OnlyShowIn. But within the context of a multiwindow session it makes sense to provide this option. OK, anyone else to approve this? Sorry I'm asking but some people might think it's alien in an X menu. (Would that addition actually pass without a mintty dependency on xdg? Otherwise, since mintty is in cygwin-base, xdg itself could also include the entry.) Thomas I think this makes sense because: 1. Many people prefer mintty over X11 terminals. 2. Many non-technical users do not understand what X11 is. They probably just know that they they can launch Cygwin Apps from this taskbar menu, and the Cygwin terminal is seemingly another Cygwin App. Believe me, I have seen misconceptions about X11 at work, even from users who use terminals. [1] 3. From an end-user's perspective, mintty isn't significantly different from an X11 app in MultiWindow mode. For example, mintty has native window decorations like most X11 apps do, you can share clipboard between mintty and X11 apps, etc. [1]. A user was using a proprietary Windows X server with XDMCP, and she thought that the remote machine's login screen (display manager's greeter), desktop environment, terminal emulator, etc was part of the X server. She was amazed that they did not change when I migrated her to Cygwin XWin with XDMCP. (Don't worry, I later migrated all my users to X2Go or X11 forwarded over SSH.)
Moving desktop environment packages out of the X11 category
Hi everyone, We recently added new package categories for desktop environments https://cygwin.com/ml/cygwin-apps/2015-05/msg00013.html However, some desktop-environment-specific packages remain under the X11 category. For example: gnome-backgrounds gnome-menus metacity Note that I think that toolkits should remain under X11. (Although one could argue that GTK3 should be under GNOME.) I also think that window managers that are not associated with a desktop environment should remain X11. Ideally, I would like to be able to install the entire X11 category without numerous desktop environment packages (like libgtop2.0_10) being pulled in due to dependencies. -Mike
cygport lua error (while packaging NEdit 5.6)
Hi. I am working on updating NEdit to 5.6. yselkowitz last updated it to 5.5-6 in 2013. I am not finished yet though. Anyway, is line 112 a bug in cygport? http://pastebin.com/p4aNCSvB The line is: lua: (command line):1: invalid escape sequence near '\;' I ran into it when using this version of the cygport file: https://github.com/mikedep333/nedit-cygport/blob/a9ffaac3f33a9d87144220bd8c5558ba9a094998/nedit.cygport -Mike
Re: [ITP] unison2.48
On Fri, Jun 5, 2015 at 9:59 PM, Andrew Schulman schulman.and...@epa.gov wrote: On Fri, 2015-06-05 at 16:16 -0400, Andrew Schulman wrote: On Fri, 2015-06-05 at 12:06 -0400, Andrew Schulman wrote: BTW do you have any plans to build the GTK GUI versions? Not immediately. The last time I tried was several years ago, and there was some problem with lablgtk that kept it from working. That was many versions of unison and lablgtk ago, so it would be worth trying again, but only if people want it. I have had a working lablgtk in Ports for some time now. Have you tried that? No, I haven't looked at it in at least 5 years. For myself, I use command line unison every day, but have no need for the GUI. And there's a Windows native build GUI, which probably meets many users' needs, although not necessarily Cygwin users. So I don't know how many people would want the GTK GUI. Well, I wouldn't outright replace the CUI version with the GUI (even though IIRC the latter can work from the command-line as well with DISPLAY unset) due to the additional dependencies, but as a separate command it would be a plus. I always thought the GUI versions would be separate packages - unison2.48gtk, etc. I'll poll to find out what the interest is. Andrew I am interested. I will try to package unison2.48gtk when I get a chance; hopefully within the next few weeks. -Mike
Re: [ITP] unison2.48
On Thu, Jun 11, 2015 at 8:51 AM, Andrew Schulman schulman.and...@epa.gov wrote: I will try to package unison2.48gtk when I get a chance; hopefully within the next few weeks. Thanks for offering to help. It seems a little odd to have different maintainers for the text and GUI versions of Unison, but I think it would be okay. Since I don't use the GUI myself, it would probably be better to have a different maintainer for it, someone who uses it. When you're ready to start, let's discuss on this list. First step would be to package LablGTK. I took Yaakov's work from ports and reworked it some, and the result works with a few minor problems still to solve. Andrew Right now, I'd like to at least take a look at what you have done so far. Can you the cygport files on public git repos, such as on GitHub? Also, I am mikedep333 on IRC. -Mike
Re: Following update to XWin or xorg-server: changed syntax to get a xterm terminal window
On Wed, Jun 10, 2015 at 7:28 AM, Fergus Daly ferg...@frontier-science.co.uk wrote: For ages I used XWin -nolock -nolisten local -multiwindow xterm -display localhost:0.0 to get a xterm terminal. Following recent updates I get a fatal error: Cannot establish any listening sockets. In the past, updates to XWin have sometimes led to similar difficulties, but I have always managed to iterate to a new successful joint syntax XWin arguments xterm arguments but this time I am totally stymied. Can anybody offer me a working syntax please? Thank you. Fergus By default, XWin 1.17 no longer listens on TCP sockets, only on unix domain sockets. https://www.cygwin.com/ml/cygwin-announce/2015-06/msg3.html However, -nolisten local disables unix domain sockets., So to use unix domain sockets: XWin -nolock -multiwindow xterm -display :0.0 Or to keep using TCP sockets: XWin -nolock -nolisten local -listen tcp -multiwindow xterm -display localhost:0.0 -Mike -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/
Re: From Microsoft: Windows 10 Console and Cygwin
On Mon, May 4, 2015 at 3:51 PM, Rich Eizenhoefer ri...@microsoft.com wrote: snip I've created a backlog item for this request so we can track the ask. It's possible, but would probably need to pick your brain in-depth more about the ask in the future. In the meantime, is it okay if I attach a copy of this email thread to the internal tracking item in our database? If not, I'll just keep the summary that I've already added. --Rich I am not an official spokesperson for Cygwin, but I am sure that you are allowed to copy this email thread. The cygwin mailing lists are public and they are publicly archived: https://cygwin.com/ml/cygwin/2015-05/ https://cygwin.com/ml/cygwin/2015-05/msg00063.html -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: From Microsoft: Windows 10 Console and Cygwin
Hi, see bottom post. On Wed, Apr 29, 2015 at 4:06 PM, Corinna Vinschen corinna-cyg...@cygwin.com wrote: Hi Rich, nice to read from you. On Apr 29 19:01, Rich Eizenhoefer wrote: Hi, I'm the Program Manager at Microsoft for the updated Windows 10 console. I searched the Cygwin FAQ and mailing list archives for issues related to Windows 10 and found an item about multiple windows which should already be fixed. We have received a couple reports about crashes when running on Windows 10 with the new console enabled. For example: cygwin is dying when i run a bunch of the git tools. For example: grep -rin log .\ 0 [main] us 0 init_cheap: VirtualAlloc pointer is null, Win32 error 487 AllocationBase 0x0, BaseAddress 0x6857, RegionSize 0x3A, State 0x1 C:\Program Files (x86)\Git\bin\grep.exe: *** Couldn't reserve space for cygwin's heap, Win32 error 0 This error message is a bit weird. For one thing, it would be nice if you could give us a pointer to the message. I'm not aware of such a bug report. Also, this is about a non-Cygwin version of git as the path to the git tools suggests. On the other hand, if it's not a Cygwin version, the error message itself only makes marginal sense. So maybe it's a Cygwin DLL as part of some third party package which is often outdated in relation to the official Cygwin distro version. As for the error message itself, it refers to the fact that, after a Cygwin fork() call, the forked child can't reserve the space required for the so-called cygwin heap, a region used for internal usage especially important for fork/exec related stuff. This usually points to Cygwin DLLs which are not rebased, or a non-Cygwin DLL loaded into a spot required for a Cygwin DLL after a fork. This is especially a problem on 32 bit due to the restricted virtual memory. However, if I'm not completely off-track, the bug report does not really have anything to do with a bad interaction between the W10 console and Cygwin. While my own W10 virtual machine under Qemu/KVM fails to update to versions after build 9926, I didn't encounter any unusual behaviour of the W10 console in comparison with, say, the W8.1 console yet. I'm usually running Cygwin tools in mintty, but I did test the W10 console, too, of course. [...] Allow me to clarify. C:\Program Files (x86)\Git\ is Git for Windows, AKA msysgit: https://msysgit.github.io/ As explained on README.md, it uses a MSys: https://github.com/msysgit/msysgit And as explained on that README.md, MSys is a fork of an old version of Cygwin. MSys's home page is here, it is part of the MinGW project: http://mingw.org/wiki/msys -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: MATE category ?
On Fri, Apr 17, 2015 at 5:57 PM, Marco Atzeri marco.atz...@gmail.com wrote: Yaakov, there are several device with MATE as category. Are we going over the list defined in: https://cygwin.com/setup.html Regards Marco There are also dozens of packages spread out over these categories that don't exist on that list either: KDE LXDE Xfce IMHO, all 4 categories should be made official. Lots of Linux distros use them as categories. The alternative is to put things like window managers and mate-settings-daemon under X11. -Mike
[ANNOUNCEMENT] New: gtkperf-0.40-1
Version 0.40-1 of gtkperf has been uploaded. Note that this utility tests GTK2 performance, not GTK3 performance. DESCRIPTION === GtkPerf is an application designed to test GTK+ performance. The point is to create common testing platform to run predefined GTK+ widgets and this way define the speed of device/platform. HOMEPAGE http://gtkperf.sourceforge.net/ -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
New: gtkperf-0.40-1
Version 0.40-1 of gtkperf has been uploaded. Note that this utility tests GTK2 performance, not GTK3 performance. DESCRIPTION === GtkPerf is an application designed to test GTK+ performance. The point is to create common testing platform to run predefined GTK+ widgets and this way define the speed of device/platform. HOMEPAGE http://gtkperf.sourceforge.net/
Re: [ITP] gtkperf 0.40
On Fri, Feb 27, 2015 at 6:00 PM, Michael DePaulo mikedep...@gmail.com wrote:
On Fri, Feb 27, 2015 at 12:37 PM, Yaakov Selkowitz
yselkow...@cygwin.com wrote:
On Fri, 2015-02-27 at 08:49 -0500, Michael DePaulo wrote:
The benefit to Cygwin is that we will have a commonly used[1]
benchmark for testing Cygwin/X performance.
Thank you Jon Turney for helping me package this.
Now at revision c81e08d
https://github.com/mikedep333/gtkperf-cygport
https://github.com/mikedep333/gtkperf-cygport
Currently at commit e707eaa
Why the NO_AUTOPOINT=1?
Otherwise, you run into this problem with the very old autotools:
make all-recursive
make[1]: Entering directory
'/usr/src/gtkperf-cygport/gtkperf-0.40-1.i686/build'
Making all in intl
make[2]: Entering directory
'/usr/src/gtkperf-cygport/gtkperf-0.40-1.i686/build/intl'
make[2]: Nothing to be done for 'all'.
make[2]: Leaving directory
'/usr/src/gtkperf-cygport/gtkperf-0.40-1.i686/build/intl'
Making all in po
make[2]: Entering directory
'/usr/src/gtkperf-cygport/gtkperf-0.40-1.i686/build/po'
make[2]: *** No rule to make target 'Makevars', needed by 'Makefile'. Stop.
make[2]: Leaving directory
'/usr/src/gtkperf-cygport/gtkperf-0.40-1.i686/build/po'
Makefile:363: recipe for target 'all-recursive' failed
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory
'/usr/src/gtkperf-cygport/gtkperf-0.40-1.i686/build'
Makefile:273: recipe for target 'all' failed
make: *** [all] Error 2
*** ERROR: make failed
I added a 3 line comment about this.
You should also change the following:
-SRC_URI=http://downloads.sourceforge.net/gtkperf/gtkperf_0.40.tar.gz
+HOMEPAGE=http://gtkperf.sourceforge.net/;
+SRC_URI=mirror://sourceforge/gtkperf/gtkperf_${VERSION}.tar.gz
${VERSION} should be used in SRC_URI to prevent mistakes when (if?) this
package gets a new version upstream.
Done.
-Mike
Questions/Comments/Complaints?
Re: [ANNOUNCEMENT] Updated: cygport-0.18.0.1-1
On Wed, Mar 4, 2015 at 11:37 PM, Yaakov Selkowitz yselkow...@cygwin.com wrote: I have uploaded a new version of cygport to the Cygwin distribution and the Fedora/EPEL Cygwin copr. Highlights in this version: * A new 'upload' command for package maintainers; ... Thank you! Don't forget to update the documentation for us newbie packagers: https://cygwin.com/setup.html -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Display issue
(I know that cygwin-xfree is deprecated. But I believe I should continue this existing conversation on it.) On Sun, Mar 1, 2015 at 1:46 PM, Maarten Hoes hoes.maar...@gmail.com wrote: Hi, On 1-3-2015 19:15, Michael DePaulo wrote: I just tried to reproduce this, but I was unable to. The icons looked correct even before I selected help - about: http://i.imgur.com/Tp50WK8.png I can reproduce it 100% of the time. I am using XWin 1.17.1-2 32-bit in multiwindow mode on Windows 10 64-bit. I launched XWin via startxwin and then ran ssh -Y fedora-host from the xterm. I then ran libreoffice. I am running a single window XDCMP session with a KDM display manager (xwin -query x.x.x.x or Xlaunch). There is no ssh usage. I am running Windows 8.1 64 bit. I open a terminal window and run ./soffice. I recreated your setup on my Fedora 21 64-bit test VM, but I was unable to reproduce it: 1. Install package group: KDE 2. Install package kdm (not sure if it was part of the package group or not) 3. Enable XDMCP in KDM (this involved rebooting) 4. X :1 -query fedora-host 5. Login to KDM 6. Open up Konsole 7. Install package libreoffice 8. soffice (did not reproduce the bug) 9. Install the libreoffice-kde package 10. soffice (did not reproduce the bug) Server is Fedora 21 64-bit. Same here. There was no command-line output from LibreOffice. My guess is that my building with --enable-dbg-util etc. has something to do with that. That's possible. Did you build from upstream source or from the Fedora SRPM? The Fedora SRPM has a TON of patches: http://pkgs.fedoraproject.org/cgit/libreoffice.git/tree/?h=f21 -Mike -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/
Re: Display issue
On Tue, Feb 24, 2015 at 6:06 PM, Maarten Hoes hoes.maar...@gmail.com wrote: Hi, Im experiencing an issue where the graphics arent displayed correctly. When I start LibreOffice on my Fedora 21 Linux system from a cygwin/x system, the left hand side (open file, template, writer document, etc.) isnt displayed properly initially. After clicking around some, the display updates and gets displayed right. Here's a screenshot of the badly displayed graphics: http://imgur.com/bTD4yQ3 And this is what it looks like (correctly) after some clicking around: http://imgur.com/adj5mKD I have no idea on how to start troubleshooting the issue; all thoughts and ideas are more than welcome. I just tried to reproduce this, but I was unable to. The icons looked correct even before I selected help - about: http://i.imgur.com/Tp50WK8.png I am using XWin 1.17.1-2 32-bit in multiwindow mode on Windows 10 64-bit. I launched XWin via startxwin and then ran ssh -Y fedora-host from the xterm. I then ran libreoffice. Server is Fedora 21 64-bit. There was no command-line output from LibreOffice. -Mike -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/
[ITP] gtkperf 0.40
The benefit to Cygwin is that we will have a commonly used[1] benchmark for testing Cygwin/X performance. Thank you Jon Turney for helping me package this. https://github.com/mikedep333/gtkperf-cygport Currently at commit e707eaa https://apps.fedoraproject.org/packages/gtkperf https://packages.debian.org/jessie/gtkperf [1] It is included in the Phoronix Test Suite: http://openbenchmarking.org/test/pts/gtkperf Note that 1.2.1 is their version number for their packaging and benchmarking scripts, not a new upstream gtkperf version. category: X11 requires: libatk1.0_0 libgdk_pixbuf2.0_0 libglib2.0_0 libgtk2.0_0 libintl8 libpango1.0_0 font-adobe-dpi75 sdesc: GTK+ performance test ldesc: GtkPerf is an application designed to test GTK+ performance. The point is to create common testing platform to run predefined GTK+ widgets and this way define the speed of device/platform.
Re: [ITP] gtkperf 0.40
On Fri, Feb 27, 2015 at 12:37 PM, Yaakov Selkowitz
yselkow...@cygwin.com wrote:
On Fri, 2015-02-27 at 08:49 -0500, Michael DePaulo wrote:
The benefit to Cygwin is that we will have a commonly used[1]
benchmark for testing Cygwin/X performance.
Thank you Jon Turney for helping me package this.
Now at revision c81e08d
https://github.com/mikedep333/gtkperf-cygport
https://github.com/mikedep333/gtkperf-cygport
Currently at commit e707eaa
Why the NO_AUTOPOINT=1?
Otherwise, you run into this problem with the very old autotools:
make all-recursive
make[1]: Entering directory '/usr/src/gtkperf-cygport/gtkperf-0.40-1.i686/build'
Making all in intl
make[2]: Entering directory
'/usr/src/gtkperf-cygport/gtkperf-0.40-1.i686/build/intl'
make[2]: Nothing to be done for 'all'.
make[2]: Leaving directory
'/usr/src/gtkperf-cygport/gtkperf-0.40-1.i686/build/intl'
Making all in po
make[2]: Entering directory
'/usr/src/gtkperf-cygport/gtkperf-0.40-1.i686/build/po'
make[2]: *** No rule to make target 'Makevars', needed by 'Makefile'. Stop.
make[2]: Leaving directory
'/usr/src/gtkperf-cygport/gtkperf-0.40-1.i686/build/po'
Makefile:363: recipe for target 'all-recursive' failed
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory '/usr/src/gtkperf-cygport/gtkperf-0.40-1.i686/build'
Makefile:273: recipe for target 'all' failed
make: *** [all] Error 2
*** ERROR: make failed
I added a 3 line comment about this.
You should also change the following:
-SRC_URI=http://downloads.sourceforge.net/gtkperf/gtkperf_0.40.tar.gz
+HOMEPAGE=http://gtkperf.sourceforge.net/;
+SRC_URI=mirror://sourceforge/gtkperf/gtkperf_${VERSION}.tar.gz
${VERSION} should be used in SRC_URI to prevent mistakes when (if?) this
package gets a new version upstream.
Done.
-Mike
Re: Cygwin/X XDMCP Issue
On Sun, Feb 22, 2015 at 4:38 PM, Maarten Hoes hoes.maar...@gmail.com wrote: On 22-2-2015 22:09, Maarten Hoes wrote: 28 1.690209 192.168.0.21 - 192.168.178.20 TCP 74 60221â+'6000 [SYN] Seq=0 Win=29200 Len=0 MSS=1460 SACK_PERM=1 TSval=6862378 TSecr=0 WS=128 Wait, what ? send to 192.168.178.20 instead of 192.168.0.20 ? i changed the ip address range of my home network. thats whats wrong, now i guess i need to figure out where i forgot to update that ip range / address. Oh, well. I can think of 2 possible issues/solutions: With XDMCP under VcXsrv or Cygwin XWin, you often need to use this argument to, usually when you have multiple (virtual) network adapters: -from 192.168.0.20 XDMCP involves a forward-confirmed reverse DNS lookup. So if you have incorrect DNS records like the following: PTR: 192.168.0.20 - windows-pc A: windows-pc - 192.168.178.20 Then this problem will happen. -Mike -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/
Re: Cygwin DLLs being modified somehow?
On Sun, Feb 22, 2015 at 10:31 PM, René Berber r.ber...@computer.org wrote: On 2/22/2015 7:53 PM, Michael DePaulo wrote: I am seeing very weird behavior and I am hoping that someone can explain it. This happens on 2 different machines. My personal Windows 10 64-bit machine, and a Windows 7 64-bit VM hosted by the X2Go project on another continent. It seems to be happening to multiple DLLs, but I will list one example below. I install 32-bit Cygwin with the 2.870 installer libopenssl100 is one of the packages that gets installed. It is at version. 1.0.1k-1 /bin/cygcrypto-1.0.0.dll always appears to be 1,820,199 bytes, and it always appears to be last modified on 1/8/2015 20:30 UTC. When I extract it from the tarball: Its md5sum is: c79b900428d91e5882c24bbb6bc36969 And its sha256sum is: 53376a3d69be996cb71abda41b7cc6bea7a599b027fe819323fafc3c433b7767 Yet on both systems, once installed, its m5sum and sha1sum differ! [snip] The installation, you may have noticed, always runs rebase at the end. There's very little documentation left about rebase, the only one I found is in: /usr/share/doc/Cygwin/_autorebase.README -- René Berber Thank you René and Andrey. I found this too: http://cygwin.wikia.com/wiki/Rebaseall However, this leads to a another question. On the X2Go project, we compile nxproxy, a 3rd-party piece of software that we are now maintaining, against Cygwin. We also recompile Cygwin OpenSSH with an X2Go-compatibility patch applied and with Kerberos support disabled. (We make sure to comply with the GPL.) So far we have just been distributing these .exe files and .dll files[1] in the X2Go Client directory (C:\Program Files (x86)\x2goclient\), which also contains regular win32 code. We do not run the Cygwin installer or call rebase. We use NSIS for our installer, and it simply copies the files. So my question is: Is there a better way to distribute/install these 2 programs? (We want to make X2Go Client for Windows easy to install, even for people who have never used Cygwin. We also want to update/upgrade these files in between X2Go Client maintenance releases (like 4.0.3.x) only when there is a need to, such as a security update.) Also, in the absence of a doing things differently, should we just distribute the original .exe and .dll files from the cygwin binary packages? -Mike [1] original cygwin files: cygcrypto-1.0.0.dll cygedit-0.dll cyggcc_s-1.dll cygjpeg-8.dll cygncursesw-10.dll cygpng15-15.dll cygssp-0.dll cygstdc++-6.dll cygwin1.dll cygz.dll sh.exe (renamed from dash.exe) recompiled OpenSSH: sftp.exe sftp-server.exe sshd.exe ssh-keygen.exe nxproxy: cygXcomp.dll nxproxy.exe -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Cygwin DLLs being modified somehow?
On Mon, Feb 23, 2015 at 1:21 PM, Andrey Repin anrdae...@yandex.ru wrote: Greetings, Michael DePaulo! I found this too: http://cygwin.wikia.com/wiki/Rebaseall However, this leads to a another question. On the X2Go project, we compile nxproxy, a 3rd-party piece of software that we are now maintaining, against Cygwin. We also recompile Cygwin OpenSSH with an X2Go-compatibility patch applied and with Kerberos support disabled. Is this really necessary? What parts of the OpenSSH need patching, and why disabling Kerberos support? 1. It looks like if I followed your recommended approach below, we would not need to apply our 1 patch. http://code.x2go.org/gitweb?p=x2goclient-contrib.git;a=blob;f=openssh-server/6.6p1-3-x2go1_src/openssh-6.6p1-1.x2go.patch;h=4ba4148fc7351c9a24986e340d59d88cdef2475a;hb=HEAD Note that with my user account, C:\Users\mike.DEPAULO\.x2go\etc\sshd_config is written to by x2goclient.exe to contain these 3 lines (the 1st 2 lines are referenced in the patch), plus 2 other lines: Subsystem shell C:/PROGRA~2/X2GOCL~4.2-2/sh Subsystem sftp C:/PROGRA~2/X2GOCL~4.2-2/sftp-server AuthorizedKeysFile /cygdrive/C/Users/%u.DEPAULO/.x2go/.ssh/authorized_keys 2. I compile it without Kerberos support in order to slim down dependencies, and reduce the number of necessary security updates. We don't use MIT Kerberos, and it receives security updates fairly often. (That said, if we do follow your recommended approach below, then linking to MIT Kerberos is a trade-off I would accept.) (We make sure to comply with the GPL.) Noted. Though, it took several attempts to find links to a source code on your website. And Download binary links dumping visitor into a directory full of different versions isn't speaking to your advantage. Take example from virtualbox.org and rework your downloads page. I will improve our wiki after I update this email, thanks. (I always update the link on the homepage though. It links to the latest version.) So far we have just been distributing these .exe files and .dll files[1] in the X2Go Client directory (C:\Program Files (x86)\x2goclient\), which also contains regular win32 code. We do not run the Cygwin installer or call rebase. We use NSIS for our installer, and it simply copies the files. So my question is: Is there a better way to distribute/install these 2 programs? (We want to make X2Go Client for Windows easy to install, even for people who have never used Cygwin. We also want to update/upgrade these files in between X2Go Client maintenance releases (like 4.0.3.x) only when there is a need to, such as a security update.) Also, in the absence of a doing things differently, should we just distribute the original .exe and .dll files from the cygwin binary packages? I have a counter-question. Any objection you have to distribute your application as part of Cygwin infrastructure? You can still keep your NSIS installer, assuming you change it to download appropriate setup.exe, and you could offer an option to make portable installation from live system, knowing well that it's already rebased and ready to work. Good question :) tl;dr: We would have to port a lot of other code from native win32 to Cygwin. 1st of all, note that X2Go Client itself (x2goclient.exe) uses libssh to authenticate and tunnel the nx-libs traffic. (nxagent on the X2Go Server, nxproxy on the X2Go Client) traffic. OpenSSH is used for our folder and printer sharing feature specifically. 2nd, nxproxy and openssh are only 2 of our Windows components. We also have the following native win32 executables/components (plus all their native win32 libraries): 1. x2goclient.exe 2. VcXsrv 3. PulseAudio 4. PuTTY 1. x2goclient.exe I assume we are keeping as a native win32 executable because we want to have it use the user's %USERPROFILE% for the user's config files, ssh known_hosts, etc. If we could make a cygwin port of x2goclient.exe use %USERPROFILE%, that would be great. I would also need to package libssh for Cygwin. Cygwin only has libssh2. Fortunately, I am in contact with upstream libssh and they are very helpful. (Note: libssh is being actively developed, despite the competing libssh2 existing.) The x2goclient source code is also full of ifdefs for Linux vs Windows vs Mac that I would need to fix for cygwin. 2. VcXsrv could be replaced with Cygwin XWin. However, I've tried it and there are some integration bugs between nxproxy and Cygwin XWin that I/we need to resolve 1st. I really want to move from VcXsrv to Cygwin XWin anyway, even if x2goclient remains native win32, because the upstream VcXsrv developer has only responded to 1 bug report / pull request I ever sent him. (He even ignores bug reports and pull requests like CVE-2013-6462 in libXfont 1.4.6). He also does git updates to the latest master branch versions of X11 libraries, rather than stable releases. In contrast, the Cygwin X11 devs (Jon Turney for example) are very responsive. Also, I need
Help porting the XLaunch feature to autoselect the display number
Hi, i was talking with Jon Turney about this on IRC. I am trying to port this feature of VcXsrv (and XMing also I think) to Cygwin XLaunch: https://sourceforge.net/p/vcxsrv/code/ci/460182676a960385dff96c1563f781213060f6fc/ Attached is my WIP patch. (I know it needs the comments updated for main.cc). There's a bug in main.cc that is causing this to happen when -1 is specified: http://imgur.com/Jv4tpip I am not very familiar with C or C++. I am hoping that someone could give me some advice. (I am much better at Java and bash.) Thanks, -Mike 0001-Now-the-display-number-can-be-set-to-automatic-by-sp.patch Description: Binary data -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/
Cygwin DLLs being modified somehow?
I am seeing very weird behavior and I am hoping that someone can explain it. This happens on 2 different machines. My personal Windows 10 64-bit machine, and a Windows 7 64-bit VM hosted by the X2Go project on another continent. It seems to be happening to multiple DLLs, but I will list one example below. I install 32-bit Cygwin with the 2.870 installer libopenssl100 is one of the packages that gets installed. It is at version. 1.0.1k-1 /bin/cygcrypto-1.0.0.dll always appears to be 1,820,199 bytes, and it always appears to be last modified on 1/8/2015 20:30 UTC. When I extract it from the tarball: Its md5sum is: c79b900428d91e5882c24bbb6bc36969 And its sha256sum is: 53376a3d69be996cb71abda41b7cc6bea7a599b027fe819323fafc3c433b7767 Yet on both systems, once installed, its m5sum and sha1sum differ! On my personal machine 8b6881e6aab4b8438a6db784003a39fb 0a4db441c5faff305b28bcddbaeae91d824ed7b39146a55f423d970d362fe6df On the X2Go project machine: 3ebfce45e66f227bc8f753ad1ef314a7 9305561f1417f2121ae2f05fd64ca9405814f757ab1eee8a4746963520e1724f I do not know assembly. Another X2Go Developer, Mihai Moldovan (CC'd) and I took a brief look at the disassembled output from objdump. It looks like the addresses differ. -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Updated: xorg-server-1.17.1-1 (TEST)
On Mon, Feb 16, 2015 at 10:21 AM, Jon TURNEY jon.tur...@dronecode.org.uk wrote: On 15/02/2015 14:49, Michael DePaulo wrote: Unfortunately, I have experienced 2 regressions vs 1.16.3-1. (In the results below, I am assuming 1 and 2 are the same regression/issue, while 3 is separate.) Thanks for testing and reporting the problem. With the excellent detail in your email I was able to build a Centos 7 VM which reproduces the problem, which was due to the RANDR changes in 1.17.1-1 not being quite correct. I've had a go at fixing this and uploaded a snapshot [1]. Perhaps you could try that and see if it improves things for you? [1] ftp://cygwin.com/pub/cygwinx/x86/XWin.20150216-git-2323ff7b08e56b79.exe.bz2 That fixes both issues. Thanks! -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/
Re: TEST RELEASE: Cygwin 1.7.34-002
On Sat, Dec 6, 2014 at 2:49 PM, Corinna Vinschen corinna-cyg...@cygwin.com wrote: I finally released another TEST version of the next upcoming Cygwin release. The version number is 1.7.34-002. I *think* I am experiencing a very bad regression. These are the Windows permissions on my ~/.ssh/id_rsa file: C:\cygwin\home\mike\.sshicacls id_rsa id_rsa NT AUTHORITY\SYSTEM:(F) DEPAULO\mike:(R,W,D,WDAC,WO) Under cygwin 1.7.33-2, I am able to use the file fine: mike@executor ~ $ uname -a CYGWIN_NT-6.3-WOW64 executor 1.7.33-2(0.280/5/3) 2014-11-13 15:45 i686 Cygwin mike@executor ~ $ ssh galactica Enter passphrase for key '/home/mike/.ssh/id_rsa': Last login: Fri Dec 12 08:36:39 2014 from executor.depaulo.org mike@galactica:~ :) [1] $ exit logout Connection to galactica closed. mike@executor ~ $ cd .ssh mike@executor ~/.ssh $ ls -latr id_rsa -rw--- 1 mike mkpasswd 1743 Dec 7 2013 id_rsa But under 1.7.34-002, I get a permissions error: mike@executor ~ $ uname -a CYGWIN_NT-6.3-WOW64 executor 1.7.34(0.282/5/3) 2014-12-06 18:03 i686 Cygwin mike@executor ~ $ ssh galactica @@@ @ WARNING: UNPROTECTED PRIVATE KEY FILE! @ @@@ Permissions 0670 for '/home/mike/.ssh/id_rsa' are too open. It is recommended that your private key files are NOT accessible by others. This private key will be ignored. key_load_private_type: bad permissions mike@galactica's password: mike@executor ~ $ cd .ssh mike@executor ~/.ssh $ ls -latr id_rsa -rw-rwx---+ 1 mike Domain Users 1743 Dec 7 2013 id_rsa -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: TEST RELEASE: Cygwin 1.7.34-002
On Sat, Dec 6, 2014 at 2:49 PM, Corinna Vinschen corinna-cyg...@cygwin.com wrote: For your convenience I wrote new documentation. Since this is a TEST prerelease, the new documentation is not part of the official docs yet. Rather have a look at https://cygwin.com/preliminary-ntsec.html If you read it (which I seriously hope for) and it's all just incomprehensible gobbledygook to you, please say so on the mailing list Might I suggest a clarification for the section Samba account mapping? Remember that there are multiple components to Samba nowadays, such as the AD DC and the NT4-style DC. But this part of the documentation concerns itself with only the Samba file server, smbd. At the beginning of it, change: A fully set up Samba with domain integration is running winbindd to map Window SIDs to artificially created UNIX uids and gids, and this mapping is transparent within the domain, so Cygwin doesn't have to do anything special. to something like: A fully set up Samba file server with domain integration is running winbindd to map Window SIDs to artificially created UNIX uids and gids, and this mapping is transparent within the domain, so Cygwin doesn't have to do anything special to access Samba shares. -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: setup-x86.exe stops downloading after one or two packages
On Sun, Dec 7, 2014 at 5:57 AM, jropph jro...@gmail.com wrote: For the few last weeks setup-x86.exe 2.852 stops downloading packages after 1 or 2 files. When I run it again, it downloads 1 or 2 of the remaining files, and then it stops again. I verified this also in another mirror. I reported this problem also. No response so far: https://cygwin.com/ml/cygwin/2014-08/msg00157.html -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: setup-x86.exe stops downloading after one or two packages
On Mon, Dec 8, 2014 at 9:44 AM, Corinna Vinschen corinna-cyg...@cygwin.com wrote: On Dec 8 09:00, Michael DePaulo wrote: On Sun, Dec 7, 2014 at 5:57 AM, jropph jro...@gmail.com wrote: For the few last weeks setup-x86.exe 2.852 stops downloading packages after 1 or 2 files. When I run it again, it downloads 1 or 2 of the remaining files, and then it stops again. I verified this also in another mirror. I reported this problem also. No response so far: https://cygwin.com/ml/cygwin/2014-08/msg00157.html Nothing has changed in setup for some time which might explain this. I just tried it myself with about 15 packages or so, and there was no hang. Is there a chance that some virus checker is interfering? Corinna [...] Back when I originally posted, I was using ESET NOD32 Antivirus 7 (not to be confused with ESET Smart Security 7) with cygwin setup-x86 2.850. I used the NOD32 GUI to temporarily disable protection, but it did not help. Now I am using ESET NOD32 Antivirus 8.0.304.0 with cygwin setup-x86 2.852. I just downloaded 61 packages from http://mirrors.kernel.org/sourceware/cygwin/ without issue, so the problem is probably resolved for me. Thanks, -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: [ANNOUNCEMENT] Updated: xorg-server-1.16.0-1 (TEST)
On Wed, Aug 20, 2014 at 7:18 AM, Jon TURNEY jon.tur...@dronecode.org.uk wrote: On 18/08/2014 14:43, Michael DePaulo wrote: On Sat, Aug 16, 2014 at 10:41 AM, Jon TURNEY wrote: The following packages have been updated in the Cygwin distribution: *** xorg-server-*1.16.0-1 I think I found a bug: There is no stdout or stderr. So for example, xwin --version does not output anything. Nothing is logged to /var/log/xwin/Xorg.0.log when I type that command. This is the same as with xorg-server 1.15. If I specify an invalid argument (e.g. xwin --foobar). then I see the information dialog box, and the xwin usage info is outputted to that log file. I did all my testing with 32-bit Cygwin. Thanks for reporting this. Are you using the same terminal in both cases? (For obscure reasons, XWin --version doesn't work in a cmd.exe terminal, see [1]) You also seem to be using a separate cygwin installation to test xorg-server 1.16, which might also be related to this in some way. [1] https://sourceware.org/bugzilla/show_bug.cgi?id=9763 -- Jon TURNEY Volunteer Cygwin/X X Server maintainer Hi Jon, Your suspician was correct. I was launching my main cygwin installation with the start menu shortcut (`C:\cygwin\bin\mintty.exe -i /Cygwin-Terminal.ico -`), but I was launching the X.org 1.16 cygwin installation by double-clicking C:\cygwin-xorg16\cygwin.bat. Thus I was running my X.org 1.16 cygwin installation with cmd.exe. When I launched my X.org 1.16 Cygwin installation by running `C:\cygwin-xorg16\bin\mintty.exe -i /Cygwin-Terminal.ico -`, this issue was resolved. Thanks, -Mike -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/
Re: Cygwin installer hangs while downloading files
On Thu, Aug 7, 2014 at 8:32 AM, Michael DePaulo mikedep...@gmail.com wrote: HI, I'm having a weird issue with the Cygwin installer. It occurs whether I am using it to update my 2 installations of Cygwin (both 32-bit and 64-bit), or to perform a new install. The issue is that the installer hangs while downloading files. In other words, the progress bar makes 0% more progress at some point during a download. I've let the installer sit for over several hours without it making any progress. Sometimes it's setup.bz2. Usually it's one of the packages. On average, it happens on the 3rd file or so. Sometimes the 1st, sometimes the 6th, etc. It usually occurs in the middle of a file download, but sometimes at the beginning of the file download (when progress is 0%.) The UI is still responsive when this happens; so I can quit the installer. My Setup: 1. Windows 8.1 64-bit with the latest updates. 2. Joined to a domain. I Tried the following, but the issue still occurs: 1. Disabling my antivirus (NOD32 Antivirus 7) 2. Using both 32-bit and 64-bit Cygwin installer 3. Using the download path C:\Users\mike.DEPAULO\Downloads\cygtemp instead of the path on the U: network drive with spaces in it. 4. Selecting different mirrors. I also tried using another computer (Same OS, joined to the same domain, but different AV) and the issue did NOT occur. My workaround right now is just repeatedly run setup.exe -q and quitting the install whenever it does hang, until the entire download completes successfully. Once the entire download does complete, the packages install without any issue. -Mike The issue seems to be happening less frequently now. Today I downloaded about 60 packages and it only happened once. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: Running a cygwin application on Windows
On Sat, Aug 16, 2014 at 8:57 AM, Fernando Gont fg...@si6networks.com wrote: On 08/16/2014 06:49 AM, Corinna Vinschen wrote: On Aug 16 12:29, Achim Gratz wrote: Fernando Gont writes: My idea was to use cygwin, since I'm not much of a Windows programmer. Is there any way to produce and ship an exe with the relevant libraries? If I understand correctly what you're trying to do, you need to obtain a commercial license from RedHat for doing so. Otherwise, if you ship binaries linked against cygwin1.dll you need to also include all sources (for Cygwin and your own application) with each such shipment. You only need the buyout license if you want to provide binaries only. But this is a GPL'ed project, so I assume there's no desire to create binary-only packages. To be as honest and straightforward as possible :-): My goal is to provide as may options as possible. I will be producing a Cygwin package since that seems to be a fair way to contribute to this project. But I also want to address the user that just knows how to run commands from the windows command line and wants to quickly download install my toolkit to try it. -- hence my goal of also producing some sort of binary distribution (even if the binary package also includes the source). I'm in the process of reading the online information to produce the Cygwin package. Are there any pointers to get the other stuff done? Thanks! -- Fernando Gont SI6 Networks e-mail: fg...@si6networks.com PGP Fingerprint: 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492 [...] Hi Fernando, I'm a volunteer contributor to the open source project X2Go. In addition to producing a Cygwin package, you might want to do what we currently do for the Windows builds of X2Go Client. We provide 5 binaries linked against cygwin .DLL files. They are nxproxy.exe (part nx-libs, maintained by our project), and 4 patched OpenSSH binaries: sftp.exe sftp-server.exe ssh-keygen.exe sshd.exe (We of course provide the source code.) Our build process needs improvement. But for now, after building those binaries, our solution is to manually copy all of the following into one single directory: 1. Those 5 binaries 2. All their cygwin .DLL dependencies 3. sh.exe We then call the cygwin-built binaries from our native win32 app, x2goclient.exe, like any other command-line binary. We can also call them from cmd.exe. We include the files (in one directory) as part of our NSIS installer That's all there is to it if you only care about a small number of our binaries working, and their dependencies don't require any directory structure. More info: 1. Our entire Windows build process is documented here. : http://wiki.x2go.org/doku.php/wiki:development:build-howto-mswin:x2goclient 2. How we handle Cygwin updates is a larger discussion. Basically, whenever we update cygwin, we rebuild our .binaries. And I monitor cygwin-announce for any security updates. Usually when there is a security update (e.g., to openssl), we update the vulnerable.DLL files individually and refer to it as a new build of x2goclient at version X. 3. We do not package nx-libs for Cygwin because it is useless without X2Go Client for Windows, which is only available for native win32, and which in turn has native win32 dependencies. 4. We do not try to upstream our SSH patch because it is specific to X2Go's needs. 5. You can determine what source package a file is from by running cygcheck. For example: $ cygcheck -f /usr/bin/cygcrypto-1.0.0.dll libopenssl100-1.0.1i-1 I do not know of an easy way to determine what binary package it is from. However, the cygwin mirrors keep all the binary packages (and the 1 source package) under a directory named after the source package. For example: ftp://mirrors.kernel.org/sourceware/cygwin/x86/release/openssl/ So after looking at the contents of the tarballs, I found it in this binary package ftp://mirrors.kernel.org/sourceware/cygwin/x86/release/openssl/libopenssl100/libopenssl100-1.0.1i-1.tar.xz -Mike DePaulo -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: [ANNOUNCEMENT] Updated: man-db-2.6.7-2
On Fri, Aug 15, 2014 at 1:52 AM, Yaakov Selkowitz yselkow...@cygwin.com wrote: The following packages have been updated in the Cygwin distribution: * man-db-2.6.7-2 man-db is an implementation of the standard Unix documentation system accessed using the man command. It uses a Berkeley DB database in place of the traditional flat-text whatis databases. This release removes the cache database generation from the postinstall step due to its often excessive length. Users will need to manage the database themselves with mandb(1) in order to use whatis/apropos(1). -- Yaakov (on behalf of Chris Breisch) IMHO, this sounds like a serious decrease in Cygwin's usability. How do Linux distributions handle this? Linux distros install many more packages by default, so doesn't their cache database generation typically take much longer? Also, should documentation (or perhaps the info a user sees when they start Cygwin for the 1st time) be updated? Respectfully, -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: [ANNOUNCEMENT] Updated: man-db-2.6.7-2
On Fri, Aug 15, 2014 at 10:20 AM, Yaakov Selkowitz yselkow...@cygwin.com wrote: On 2014-08-15 07:22, Michael DePaulo wrote: This release removes the cache database generation from the postinstall step due to its often excessive length. Users will need to manage the database themselves with mandb(1) in order to use whatis/apropos(1). IMHO, this sounds like a serious decrease in Cygwin's usability. How do Linux distributions handle this? Linux distros install many more packages by default, so doesn't their cache database generation typically take much longer? Exactly why they don't seem to do it during postinstall either. For example, in Fedora this is handled by a cron job. A future release may add that functionality, but it is clear that postinstall is the wrong place for it. I disagree, but I need to read more about the subject in order to have a well-informed opinion. Also, should documentation (or perhaps the info a user sees when they start Cygwin for the 1st time) be updated? How so? Yaakov Consider the example of where I work. After I install the corporate IT department's SCCM package/script for Cygwin (1.7.16, last updated August 2012), I am greeted by this message every time I start the Cygwin Terminal: Your group is currently mkpasswd. This indicates that your gid is not in /etc/group and your uid is not in /etc/passwd. The /etc/passwd (and possibly /etc/group) files should be rebuilt. See the man pages for mkpasswd and mkgroup then, for example, run mkpasswd -l [-d] /etc/passwd mkgroup -l [-d] /etc/group Note that the -d switch is necessary for domain users. I can get rid of the message if do I run those commands. However, here's what happens if I do run them: mkpasswd -l -d takes 9m29.159s outputs 37062 lines mkgroup -l -d takes 4m26.470s outputs 36866 lines (I have to run those commands with admin rights, but fortunately mandb does not require admin rights because of user DBs.) I am not suggesting that users see a mandb message every time they launch cygwin. But I am suggesting that that users see it when they 1st launch cygwin. They already see this message: Copying skeleton files. These files are for the users to personalise their cygwin experience. They will never be overwritten nor automatically updated. './.bashrc' - '/home/mike//.bashrc' './.bash_profile' - '/home/mike//.bash_profile' './.inputrc' - '/home/mike//.inputrc' './.profile' - '/home/mike//.profile' Also, users who are particularly reliant on the apropos command probably don't know about running the mandb command. It's analogous to users launching the Windows Help and Support Center, but its search bar returning 0 results they haven't run another utility 1st. -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: [ANNOUNCEMENT] Updated: man-db-2.6.7-2
On Fri, Aug 15, 2014 at 8:05 PM, Andrey Repin anrdae...@yandex.ru wrote: Greetings, Michael DePaulo! This release removes the cache database generation from the postinstall step due to its often excessive length. Users will need to manage the database themselves with mandb(1) in order to use whatis/apropos(1). IMHO, this sounds like a serious decrease in Cygwin's usability. How do Linux distributions handle this? Linux distros install many more packages by default, so doesn't their cache database generation typically take much longer? Exactly why they don't seem to do it during postinstall either. For example, in Fedora this is handled by a cron job. A future release may add that functionality, but it is clear that postinstall is the wrong place for it. I disagree, but I need to read more about the subject in order to have a well-informed opinion. The TL;DR version of the issue is that compilation of database is slow and prone to fail, plus the database should be re-indexed each time a page is added, removed or changed. If Cygwin do this only at man-db postinstall, the database WILL go out of reality pretty soon. Understood. I was thinking that the db was generated every time you updated or installed packages. That's how Debian and its derivatives do keep it up-to-date. (Actually, they only run mandb if 1 or more of the updated/installed packages have manpage. It's handled by their dpkg trigger feature.) Reference: http://askubuntu.com/questions/178773/is-there-a-way-to-see-what-exactly-the-processing-triggers-does-per-package-ba Also, should documentation (or perhaps the info a user sees when they start Cygwin for the 1st time) be updated? How so? Yaakov Consider the example of where I work. After I install the corporate IT department's SCCM package/script for Cygwin (1.7.16, last updated August 2012), I am greeted by this message every time I start the Cygwin Terminal: Your group is currently mkpasswd. This indicates that your gid is not in /etc/group and your uid is not in /etc/passwd. This is about to change in a short while. I am not suggesting that users see a mandb message every time they launch cygwin. This could be resolved in a more graceful way. I.e. motd/fortune/etc. But I am suggesting that that users see it when they 1st launch cygwin. They already see this message: Copying skeleton files. These files are for the users to personalise their cygwin experience. They will never be overwritten nor automatically updated. Users will only see this message, if the $HOME directory is created anew. I, for one, do not see a reason to keep Cygwin users' $HOMEs separate from a system %USERPROFILE%. Also, users who are particularly reliant on the apropos command probably don't know about running the mandb command. It's analogous to users launching the Windows Help and Support Center, but its search bar returning 0 results they haven't run another utility 1st. That's an issue worth resolving. We just need to find a way to do it in a non-abusive fashion. -- WBR, Andrey Repin (anrdae...@yandex.ru) 16.08.2014, 03:57 Sorry for my terrible english... Thanks, I have a much better understanding of the situation now. I agree that there is no simple solution. -Mike -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: How do fonts work when you have no font packages installed?
On Sun, Jul 6, 2014 at 11:08 AM, Jon TURNEY jon.tur...@dronecode.org.uk wrote: On 04/07/2014 08:16, Michael DePaulo wrote: I just installed cygwin64 with only the base packages, xorg-server, and its dependencies installed. The FAQ (2014-04-29) states: 3.5. My favourite font has gone! The font Emacs uses is just boxes Only minimal fonts will be installed after the upgrade. However, on my system, /usr/share/fonts/ is empty. The X server has a version of 6x13 fixed font built-in, to allow it and (some) apps which use core fonts, to operate in this situation. If some local X clients were installed, this would (hopefully) cause any X core fonts they require to be installed. Yet I am able to XDMCP into a GNOME2 CentOS 6.5 machine and all the fonts show up fine with the handful of apps I tested. Can someone explain how text is being rendered? Is Cygwin Xwin using fonts from the Windows OS? While this is technically possible by adding the Windows font directory to the X server font path, that is not done. (Although those fonts are made available to local clients by telling fontconfig to look in the Windows font directory, although this is not technically perfect as there isn't any mechanism to tell fontconfig to update it's cache when Windows fonts are added or removed) Are the fonts being rendered client-side via XRender? Yes, if the apps you tested are modern, client-side fonts are almost certainly being used. [1] explains this fairly well: The first X11 clients used the core X11 protocol to draw text, as that was the only choice. [...] Because GTK+ and Qt, the toolkits behind several applications including all GNOME and KDE applications, switched to Xft, many programs on most desktops [...] now use Xft. Thank you very much, this is a very good answer. I was not aware of Xft. Core fonts are a legacy feature. If you want to use an older remote application which requires a particular core font, you will have to install it. Yes, I've observed this too. At work we run an old (but actively maintained) commercial Linux application written in Motif, and it requires core fonts to be installed on the Windows machines running Cygwin X11 or VcXsrv. [1] http://en.wikibooks.org/wiki/Guide_to_X11/Fonts -- Jon TURNEY Volunteer Cygwin/X X Server maintainer Thanks again, -Mike DePaulo -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/
How do fonts work when you have no font packages installed?
I tried to send this email previously, but I think it did not go through because I was not subscribed to the list yet. Hi, I just installed cygwin64 with only the base packages, xorg-server, and its dependencies installed. The FAQ (2014-04-29) states: 3.5. My favourite font has gone! The font Emacs uses is just boxes Only minimal fonts will be installed after the upgrade. However, on my system, /usr/share/fonts/ is empty. Yet I am able to XDMCP into a GNOME2 CentOS 6.5 machine and all the fonts show up fine with the handful of apps I tested. Can someone explain how text is being rendered? Is Cygwin Xwin using fonts from the Windows OS? Are the fonts being rendered client-side via XRender? Thanks in advance, -Mike mike@executor /usr $ uname -a CYGWIN_NT-6.3 executor 1.7.30(0.272/5/3) 2014-05-23 10:36 x86_64 Cygwin mike@executor /usr $ find . | grep fonts ./share/doc/groff-1.22.2/examples/hdtbl/fonts_n.ps ./share/doc/groff-1.22.2/examples/hdtbl/fonts_n.roff ./share/doc/groff-1.22.2/examples/hdtbl/fonts_x.ps ./share/doc/groff-1.22.2/examples/hdtbl/fonts_x.roff mike@executor /usr $ cd share mike@executor /usr/share $ ls awk file icons locale man p11-kitpki terminfo texmf zoneinfo doc groff info magic misc pkgconfig tabset texinfo X11 mike@executor /usr/share $ X -query galactica -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/
