Re: bug report: 64-bit cygwin setup crashes under Wine
On Tue, Sep 03, 2013 at 10:08:44PM -0700, Austin English wrote: >I recently noticed the 64-bit cygwin installer crashes under wine. >After further debugging, it appears that the issue is that cygwin is >misaligning the stack, causing a crash. I've copy/pasted the analysis >below: > >Hello folks, > >confirming. > >Reminded me of bug 27680 (violation of the Windows 64-bit ABI, stack pointer >(RSP) must be 16-byte aligned when making a call to Win64 API)... > >Similarly the gcc generated prologue code for MSABI (SSE2 callee save) causes a >crash due to misaligned stack (needs to be 16-byte aligned). If you'd like to submit a patch to get setup working in Wine it would be thoughtfully considered. Otherwise, I can say with a fair amount of certainty that no one is going to be interested in installing Wine to debug issues running (in this case non-cygwin!) programs there. -- 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
bug report: 64-bit cygwin setup crashes under Wine
Howdy,
I recently noticed the 64-bit cygwin installer crashes under wine.
After further debugging, it appears that the issue is that cygwin is
misaligning the stack, causing a crash. I've copy/pasted the analysis
below:
Hello folks,
confirming.
Reminded me of bug 27680 (violation of the Windows 64-bit ABI, stack pointer
(RSP) must be 16-byte aligned when making a call to Win64 API)...
Similarly the gcc generated prologue code for MSABI (SSE2 callee save) causes a
crash due to misaligned stack (needs to be 16-byte aligned).
--- snip ---
...
0027:Starting process L"Z:\\home\\focht\\Downloads\\setup-x86_64.exe"
(entryproc=0x691616)
0027:Call KERNEL32.GetModuleHandleA(0057ce81 "KERNEL32") ret=0057ce90
0027:Ret KERNEL32.GetModuleHandleA() retval=7b82 ret=0057ce90
0027:Call KERNEL32.GetProcAddress(7b82,0057ce9a "VirtualProtect")
ret=0057ceb1
0027:Ret KERNEL32.GetProcAddress() retval=7b83007c ret=0057ceb1
0027:trace:seh:raise_exception code=c005 flags=0 addr=0x7fb738137842
ip=7fb738137842 tid=0027
0027:trace:seh:raise_exception rax=7b83007c rbx=7b83007c
rcx=7bae33d0 rdx=000403dc
0027:trace:seh:raise_exception rsi=0069 rdi=0040016f
rbp=0033fc08 rsp=0033fb18
0027:trace:seh:raise_exception r8=0033fc80 r9=0033fc88
r10= r11=003629581400
0027:trace:seh:raise_exception r12=7bf00c50 r13=7fffb49477a0
r14= r15=
...
--- snip ---
Winedbg isn't useful here because it doesn't support SSE2 instructions.
Running winedbg --gdb yields more useful results.
Debug session:
--- snip ---
...
Dump of assembler code for function NtProtectVirtualMemory:
=> 0x7f8e5e51703a <+0>:push %rbp
0x7f8e5e51703b <+1>:mov%rsp,%rbp
0x7f8e5e51703e <+4>:push %rdi
0x7f8e5e51703f <+5>:push %rsi
0x7f8e5e517040 <+6>:sub$0x1d0,%rsp
0x7f8e5e517047 <+13>:movaps %xmm6,-0xb0(%rbp)
0x7f8e5e51704e <+20>:movaps %xmm7,-0xa0(%rbp)
0x7f8e5e517055 <+27>:movaps %xmm8,-0x90(%rbp)
0x7f8e5e51705d <+35>:movaps %xmm9,-0x80(%rbp)
0x7f8e5e517062 <+40>:movaps %xmm10,-0x70(%rbp)
0x7f8e5e517067 <+45>:movaps %xmm11,-0x60(%rbp)
0x7f8e5e51706c <+50>:movaps %xmm12,-0x50(%rbp)
0x7f8e5e517071 <+55>:movaps %xmm13,-0x40(%rbp)
0x7f8e5e517076 <+60>:movaps %xmm14,-0x30(%rbp)
0x7f8e5e51707b <+65>:movaps %xmm15,-0x20(%rbp)
0x7f8e5e517080 <+70>:mov%rcx,0x10(%rbp)
0x7f8e5e517084 <+74>:mov%rdx,0x18(%rbp)
0x7f8e5e517088 <+78>:mov%r8,0x20(%rbp)
0x7f8e5e51708c <+82>:mov%r9d,0x28(%rbp)
0x7f8e5e517090 <+86>:movl $0x0,-0xb4(%rbp)
0x7f8e5e51709a <+96>:mov0x20(%rbp),%rax
0x7f8e5e51709e <+100>:mov(%rax),%rax
...
--- snip ---
Reg dump before faulting instruction (xmm6 save):
--- snip ---
Wine-gdb> info reg
rax0x33fc503406928
rbx0x7b89f3602072638304
rcx0x-1
rdx0x33fc503406928
rsi0x696881280
rdi0x40016f4194671
rbp0x33fbe80x33fbe8
rsp0x33fa080x33fa08
r8 0x33fc583406936
r9 0x44
r100x3629532710232621549328
r110x00
r120x7bf00c502079329360
r130x7fff237e6030140733788872752
r140x00
r150x00
rip0x7f8e5e5170470x7f8e5e517047
eflags 0x202[ IF ]
cs 0x3351
ss 0x2b43
ds 0x00
es 0x00
fs 0x00
gs 0x00
Program received signal SIGSEGV, Segmentation fault.
...
0x7f8e5e517047 in NtProtectVirtualMemory (process=0x0, addr_ptr=0x0,
size_ptr=0x0, new_prot=6881280, old_prot=0x33fc88)
at /home/focht/projects/wine/wine-git/dlls/ntdll/virtual.c:2120
2120{
--- snip ---
Windows x86_64 ABI specifies that registers xmm6-xmm15 be callee saved.
Gcc generates the MSABI prologue accordingly (see previous disassembly).
At the entry point of VirtualProtect() the stack is 16-byte aligned ...
--- snip ---
Stopped on breakpoint 1 at 0x7b89f360 VirtualProtect
[/home/focht/projects/wine/wine-git/dlls/kernel32/virtual.c:214] in kernel32
VirtualProtect () at
/home/focht/projects/wine/wine-git/dlls/kernel32/virtual.c:214
214{
Wine-dbg>info reg
Register dump:
rip:7b89f360 rsp:0033fc80 rbp:0033fd20 eflags:0206
( - -- I - -P- )
rax:7b89f360 rbx:7b89f360 rcx:0040016f
rdx:0078
rsi:0069 rdi:0040016f r8:0004
r9:0033fc88 r10:003629532710
r11: r12:7bf00c50 r13:7f7bdb10
r14: r15:
Wine-dbg>
--- snip ---
*dingding*
Re: lpr command line syntax
Greetings, Rockefeller, Harry! > The Cygwin lpr command requires both server and printer names, i.e., > $ lpr foo > Sends a print job to a network printer > Usage: lpr -S server -P printer [-C class] [-J job] [-o option] [-x] [-d] > filename I'm afraid, you're using Windows lpr command, rather than Cygwin. (Yes, Windows DO have lpr command-line utility - it is a part of TCP/IP installation, thus comes preinstalled on systems starting Windows '98.) Though, I don't have Cygwin's lpr installed to compare records. Never had the need for it. > And then the command fails. > Is this intended? > Linux will use environment variable defaults, will it not? -- WBR, Andrey Repin (anrdae...@yandex.ru) 04.09.2013, <03:07> Sorry for my terrible english... -- 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: Mirror sites/archives - "download incomplete" problem
On 9/3/2013 5:23 PM, Mark Sheffield wrote: Is there some likelihood that some of the mirrors have corrupted packages? Sure, it's possible. Just like it's possible for files to get corrupted during your download. Is there a method to determine whether or not any specific download mirror has received the current build correct and intact? There's an MD5SUM for the package in setup.ini. You can check that value against the value you get for the package. That's essentially what "setup*.exe" does though. It sounds to me like it was reporting that case to you (i.e. download incomplete). -- Larry _ A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting annoying in email? -- 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
Mirror sites/archives - "download incomplete" problem
Hello - I have been struggling with getting gcc running under cygwin and after struggling for the better part of a week decided that the problem may be in what I downloaded from a mirror site. I had downloaded my copy originally (a few months ago) from ftp://cygwin.mirrorcatalogs.com and had gotten a "download incomplete" error. The installation went (mostly) OK, with some reading from null errors, but I thought "Oh, that linux" and let it go. The installation mostly worked, completely worked as far as I knew. I was concerned, though, and tried maybe 5 times in the following several days to get the download to complete OK, but after needing to get gcc running this past week (nothing I was doing made any difference) I started investigating this package download nuance as perhaps the culprit. I tried downloading from ftp://ftp.gtlib.gatech.edu and ftp://lug.meu.edu and got the same error and was still thinking that it was some unimportant artifact that hadn't been fixed because it was so very unimportant. I then tried ftp://mirror.cs.vt.edu and got NO error messages. I installed this package and got NO errors. Since I did this, my gcc linker began working for the first time and I am now compiling the latest version of gcc which is a massive undertaking for the gcc compiler and cygwin. My series of questions out of all of this: Is there some likelihood that some of the mirrors have corrupted packages? Is there a method to determine whether or not any specific download mirror has received the current build correct and intact? I think that my problem is solved, but there are enough hits on the web to see that other users have the same problem. -- 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: Trying to build rxvt on Cygwin64.
On 9/3/2013 7:40 AM, Andrey Repin wrote: Greetings, Robert McBroom! I'm accustomed to using rxvt. I know how to set all the options that I want to use. mintty is to limited and xterm has too much stuff. mintty is based on the PuTTY source code, is very robust and well-tested. If you find a feature missing in mintty, you may ask the author to include it, or to explain it's absence. Last I checked, rxvt do not have support for UNICODE, aside from one not very well working hack. -- WBR, Andrey Repin (anrdae...@yandex.ru) 03.09.2013, <15:37> I don't need UNICODE. It has been nice to have one command window system that works the same on lots of unix flavors for the last thirty years. It is useful to find out the quirks that allow building a X program on Cygwin64 for lots of other applications. Provided that those that know are willing to share. Robert McBroom -- 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: lpr command line syntax
Problem resolved. I was using the Windows 'lpr' command. Even though cygutils is part of the 'base' Cygwin install this package didn't have lpr. So, I installed the other cygutils packages and /usr/bin/lpr was there. Well, it was either that, or running lprsetup.sh and restarting Cygwin that brought about the Cygwin lpr command. -- 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: ZSH "Address Space Needed By" problem
Sure enough, updating my installation to the most recent version seems to have fixed it. Thank you for the help. -- 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: ZSH "Address Space Needed By" problem
On 9/3/2013 1:43 PM, Chris wrote: Hey folks, I fired up Cygwin this morning to find that my ZSH is having a problem starting. I'm getting the following error message anytime I try to start it: Judging by the fact that your Cygwin package is 8 versions out-of-date, I'm going to hazard a guess that you have not updated your Cygwin installation between today and the last time you used it and that you have rebooted your computer well before this problem appeared. If my guess on all this is correct, it points to something else on your system changing in the interim and instigating this complaint. That "something else" would be BLODA (http://cygwin.com/acronyms/#BLODA). Take a look and see what you can find, keeping in mind that recent installs of other software may provide a clue. -- Larry _ A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting annoying in email? -- 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
ZSH "Address Space Needed By" problem
Hey folks, I fired up Cygwin this morning to find that my ZSH is having a problem starting. I'm getting the following error message anytime I try to start it: ZSH initiated. Syncing profile with zorro...done 0 [main] zsh 5960 child_info_fork::abort: address space needed by 'rlimits.dll' (0x4A) is already occupied /home/goose/.zsh/alias:15: fork failed: resource temporarily unavailable 0 [main] zsh 3608 child_info_fork::abort: address space needed by 'rlimits.dll' (0x4A) is already occupied /home/goose/.zsh/completion:46: fork failed: resource temporarily unavailable 0 [main] zsh 1080 child_info_fork::abort: unable to remap parameter.dll to same address as parent (003F) - try running rebaseall /home/goose/.zsh/prompt:7: fork failed: resource temporarily unavailable 0 [main] zsh 1312 child_info_fork::abort: address space needed by 'parameter.dll' (0x3F) is already occupied precmd:4: fork failed: resource temporarily unavailable (goose@crucio)[128] ~ % I'm unable to execute any commands, even "bash" or "echo"; however, if I manually edit my /etc/passwd file and change my shell from ZSH to bash, I'm able to use bash without any issues. Here's the locations of the affected DLL's: goose@crucio / $ find . -depth -type f -name "*parameter.dll*" ./lib/zsh/5.0.2/zsh/parameter.dll ./lib/zsh/5.0.2/zsh/zleparameter.dll ./usr/lib/zsh/5.0.2/zsh/parameter.dll ./usr/lib/zsh/5.0.2/zsh/zleparameter.dll goose@crucio / $ find . -depth -type f -name "*rlimits.dll*" ./lib/zsh/5.0.2/zsh/rlimits.dll ./usr/lib/zsh/5.0.2/zsh/rlimits.dll I've tried the following commands, none of which have fixed it: D:\cygwin\bin\dash.exe -c '/usr/bin/rebaseall -v' D:\cygwin\bin\dash.exe -c '/usr/bin/rebaseall -v -s *.dll' D:\cygwin\bin\dash.exe -c '/usr/bin/rebase -s /usr/lib/zsh/5.0.2/zsh/*.dll' D:\cygwin\bin\dash.exe -c '/usr/bin/rebase -O /usr/lib/zsh/5.0.2/zsh/*.dll' The first time I attempted to send this email with my "cygcheck.out" attached, I received a response of "ezmlm-reject: fatal: Sorry, I don't accept messages larger than 100 bytes (#5.2.3)" (Not sure why I got this, as my cygcheck.out measures 815908 bytes). The second time I attempted to send this email, I uploaded the file to my personal web server and provided a direct link to the cygcheck.txt file the, and I received the error reply: "552 spam score exceeded threshold" So forgive me for making you fix this URL by hand, but hopefully this third try it will actually send to the mailing list: https (colon slash slash) captain (dash) kickass (dot) net (slash) files (slash) cygcheck (dot) txt Thanks. -- 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
lpr command line syntax
The Cygwin lpr command requires both server and printer names, i.e., $ lpr foo Sends a print job to a network printer Usage: lpr -S server -P printer [-C class] [-J job] [-o option] [-x] [-d] filename And then the command fails. Is this intended? Linux will use environment variable defaults, will it not? -- 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
[ANNOUNCEMENT] Updated: {lcms2/liblcms2_2/liblcms2-devel}-2.5-1: Little color management engine
Hi New versions of 'lcms2/liblcms2_2/liblcms2-devel' have been uploaded to a server near you. o Update to latest upstream release o Build for cygwin 1.7.24 with gcc-4.7.3 lcms2 NEWS: === User defined parametric curves can now be saved in ICC profiles. RGB profiles using same tone curves for several channels are storing now only one copy of the curve (saves space) Fixed a multithead bug on optimization (StageDEF) Fixed some 64 bit warnings on size_t to uint32 conversions update black point detection algorithm to reflect ICC changes Added new cmsPlugInTHR() and fixed some race conditions (thanks to Artifex) Added error descriptions on cmsSmoothToneCurve Fixed a bug on big endian platforms not supporting uint64 or long long. Fixed some bugs on floating point curves. Several improvements in cgats parser. Several minor issues found by cppcheck Fixed devicelink generation for 8 bits Added a reference for Mac MLU tag Added a way to read the profile creator from header Added identity curves support for write V2 LUT More changes from Artifex Added TIFF Lab16 handling on tifficc Added some fixes from XhmikosR Fixed a bug in parametric curves Remove spurious tabs added by git merge of pull request Added support for VS2012 Added a simple project for cppcheck Rendering intent used when creating the transform is now propagated to profile header in cmsTransform2Devicelink. This is because 7.2.15 in spec 4.3 Transform2Devicelink now keeps white point when guessing deviceclass is enabled Added some checks for non-happy path, mostly failing mallocs CYGWIN-ANNOUNCE UNSUBSCRIBE INFO If you want to unsubscribe from the cygwin-announce mailing list, please use the automated form at: http://cygwin.com/lists.html#subscribe-unsubscribe If this does not work, then look at the "List-Unsubscribe: " tag in the email header of this message. Send email to the address specified there. It will be in the format: cygwin-announce-unsubscribe-you=yourdomain.com cygwin.com If you need more information on unsubscribing, start reading here: http://sourceware.org/lists.html#unsubscribe-simple Please read *all* of the information on unsubscribing that is available starting at this URL. -- 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 to update an app for both x86 and x64?
On Tue, Sep 03, 2013 at 03:45:38PM +0400, Andrey Repin wrote: >Greetings, Christopher Faylor! > >>>I'm the maintainer of ???pv' which is only available for x86. What are >>>the proper steps to upload a new version of pv for both architectures? > >> It is the responsibility of the pv package maintainer to package a new >> version and make it available for upload. You have to wait for them to >> do this. > >Oh, but he said he's a maintainer >.> It's not yet a Thursday, is it? This isn't my first rodeo. He is *not* the Cygwin maintainer, or, if he is, then he's using a different email address and not identifying himself as such. -- 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
[ANNOUNCEMENT] Updated: {ming/libming1/libming-devel/python-ming/perl-ming/tcl-ming}-0.4.5-1: A SWF output library
Hi New versions of 'ming/libming1/libming-devel/python-ming/perl-ming/tcl-ming' have been uploaded to a server near you. o Updated to latest upstream release o Build for cygwin 1.7.24 with gcc-4.7.3 ming NEWS: == Fix a build problem with perl module and Perl 5.13.10 (Debian package patch 07) Fix an issue with a shared lib version of the perl module (Debian package patch 05) Fix a runtime warnign in the PHP module (Issue #6) Fix build probem with PHP 5.4 (Issue #9, Debian Bug #662620) Took in some build warning fixes found by openSUSE Add support for giflib 4.2.0+ (bugzilla #119) Fix build with SunStudio compiler (bugzilla #103, #118) Include test file missing in 0.4.4 (Oops) Ensure perl tests use uninstalled package Allow running testsuite with interpreters in custom paths Improve detection of TCL files Fix testsuite runs on systems w/out GNU vasprintf Fix run of ActionScriptTests with BSD make Fix short allocation parsing empty glyph record CYGWIN-ANNOUNCE UNSUBSCRIBE INFO If you want to unsubscribe from the cygwin-announce mailing list, please use the automated form at: http://cygwin.com/lists.html#subscribe-unsubscribe If this does not work, then look at the "List-Unsubscribe: " tag in the email header of this message. Send email to the address specified there. It will be in the format: cygwin-announce-unsubscribe-you=yourdomain.com cygwin.com If you need more information on unsubscribing, start reading here: http://sourceware.org/lists.html#unsubscribe-simple Please read *all* of the information on unsubscribing that is available starting at this URL. -- 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
[ANNOUNCEMENT] Updated: {libEMF/libEMF1/libEMF-devel}-1.0.7-2: A library for generating Enhanced Metafiles
Hi New versions of 'libEMF/libEMF1/libEMF-devel' have been uploaded to a server near you. o Build for cygwin 1.7.24 with gcc-4.7.3 o Fixed libEMF.la file pointing to old /usr/lib/gcc/i686-pc-cygwin/4.5.3/libstdc++.la file CYGWIN-ANNOUNCE UNSUBSCRIBE INFO If you want to unsubscribe from the cygwin-announce mailing list, please use the automated form at: http://cygwin.com/lists.html#subscribe-unsubscribe If this does not work, then look at the "List-Unsubscribe: " tag in the email header of this message. Send email to the address specified there. It will be in the format: cygwin-announce-unsubscribe-you=yourdomain.com cygwin.com If you need more information on unsubscribing, start reading here: http://sourceware.org/lists.html#unsubscribe-simple Please read *all* of the information on unsubscribing that is available starting at this URL. -- 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: Disabling focus reporting in mintty
Thomas, Thanks for your pointers. The issue did indeed lie in my tmux configuration, but it is confusing why it did not manifest itself before. When I created a completely new user account with no settings whatsoever, tmux did not exhibit the issue I have been describing. On a whim, I blew away my tmux configuration and tried again (important to note: if you have active tmux sessions, you must use a different socket (-L option) to test a new configuration, or close your open sessions). When I eliminated my tmux configuration, the problem was gone. The only thing that looked incriminating was this line: set -g escape-time 0 I set this to 0 because hitting escape would cause delays in e.g. vim when escape was pressed. See here: http://superuser.com/questions/252214/slight-delay-when-switching-modes-in-vim-using-tmux-or-screen. As a compromise, I set the value to 100 instead of 0. Problem solved! As I mentioned, I have never noticed this issue before. The setting has existed in my tmux configuration for several years, so I can only conclude something must have changed recently in tmux. Thanks for your help! On Thu, Aug 29, 2013 at 2:26 AM, Thomas Wolff wrote: > Am 27.08.2013 14:52, schrieb John Koelndorfer: > >> Thomas, thanks for the suggestion. I thought I had tried that and >> indeed, when I attempted to disable focus reporting within tmux there >> wasn't any change. I'd have to guess that tmux is intercepting the >> control sequence and probably discarding it. I don't know enough >> about terminal emulation to say if that's a sane thing to do. >> >> In any case, I'm told in #tmux on freenode that focus reporting can be >> disabled in tmux 1.9, so it's just a matter of waiting for that >> release (or building the latest sources). > > Hmm. I've just tried to reproduce the issue by compiling tmux 1.8 myself. > On my PPC/Debian system it supports focus reporting as expected and does not > switch it on by default. It both passes through window focus and generates > its own focus report if switching panes, but *only if* enabled in the > respective pane. > I don't see any configuration option which could have affected that, either. > The behaviour you described cannot be observed; it could be observed, > obviously, if you enable focus reporting explicitly, or run an application > that does so (like my editor mined) and kill it with kill -9 (so it cannot > reset it). > Otherwise I don't observe your problem. Please check again, is there any > other source of getting the reports enabled in your environment? Do you > enable it in your .profile or .bashrc? > -- > Thomas > > >> >> Thanks again! >> >> On Mon, Aug 26, 2013 at 4:09 PM, Thomas Wolff wrote: >>> >>> Am 26.08.2013 22:54, schrieb John Koelndorfer: >>> I seem to have run into some trouble with focus reporting in mintty 1.1.2 and tmux 1.8 running on a remote Arch Linux host. tmux does not seem to catch the focus reporting control characters properly and will allow the "^[[O" and "^[[I" to bleed through to applications. Here is how I have tested this: 1. Connect to remote host over ssh. 2. Run `cat'. 3. Click on the mintty taskbar entry repeatedly. Note that there are no control characters showing up. Now, I repeat the above steps, but before running cat, I launched a tmux session using `tmux'. Now, when I click the mintty taskbar, I see: ^[[O^[[I repeatedly. In other applications things will manifest differently, but my issue is specifically with irssi where "[I" shows up in the input box every time I focus mintty, which is quite annoying. In order to try and mitigate this issue, I attempted to echo ^[[?1004l per http://code.google.com/p/mintty/wiki/Changes (under version 0.4.1). Exact steps were: 1. Launch mintty. 2. Run `cat' locally. 3. Press "ESC [ ? 1 0 0 4 l", followed by ^D. I believe this should echo the correct control sequence, can someone confirm? Nothing shows up on stdout from cat, so it appears mintty is indeed interpreting this as a control sequence. 4. Connect to same remote host, run tmux and cat as described before. 5. Focus and unfocus mintty by clicking the taskbar entry. However, the above steps still show the focus reporting control characters bleeding through to cat. Am I going about disabling focus reporting correctly? Should it do what I think it should be doing? Is there some other solution that I might be missing? >>> >>> Reading your report above, it seems tmux enables focus reporting (and >>> doesn't handle it then...). >>> So as a workaround, you would need to disable it (sending that escape >>> sequence) *after* you've started tmux, inside it. >>> -- >>> Thomas >>> >>> -- >>> Problem reports: http://cygwin.com/problems.html >>> FAQ: http://cygwin.com/faq/ >>> Documentation:
Re: [ANNOUNCEMENT] Updated: Cygwin 1.7.25
On 3 September 2013 07:36, Achim Gratz wrote: > Chris Sutcliffe writes: >> I've checked 5 different mirrors again this morning, and given that >> the release was 3 days ago I don't think it's a sync issue. 1.7.25 >> seems to be missing on the mirrors for x86. > > The files have been there all the time, but setup.ini wasn't updated. That > is now fixed, depending on when the mirrors update you may or may not see it > yet. Confirmed, I just tried again and 1.7.25 has shown up at the mirrors I frequent. Thanks, Chris -- Chris Sutcliffe http://emergedesktop.org http://www.google.com/profiles/ir0nh34d -- 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 to update an app for both x86 and x64?
Greetings, Christopher Faylor! >>I'm the maintainer of ???pv' which is only available for x86. What are >>the proper steps to upload a new version of pv for both architectures? > It is the responsibility of the pv package maintainer to package a new > version and make it available for upload. You have to wait for them to > do this. Oh, but he said he's a maintainer >.> It's not yet a Thursday, is it? -- WBR, Andrey Repin (anrdae...@yandex.ru) 03.09.2013, <15:44> Sorry for my terrible english... -- 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: Trying to build rxvt on Cygwin64.
Greetings, Robert McBroom! >>> The source tar balls I can find of rxvt are quite old and the config.guess >>> in the package doesn't recognize a win 7 x64, Cygwin64 system. Any pointers >>> to getting a x64 rxvt to build? >> If you're not building it out of pure curiosity, I suggest you state your >> intentions more clearly. >> If all you want is a graphical terminal window, look into mintty shipped with >> modern cygwin installations. > I'm accustomed to using rxvt. I know how to set all the options that I > want to use. mintty is to limited and xterm has too much stuff. mintty is based on the PuTTY source code, is very robust and well-tested. If you find a feature missing in mintty, you may ask the author to include it, or to explain it's absence. Last I checked, rxvt do not have support for UNICODE, aside from one not very well working hack. -- WBR, Andrey Repin (anrdae...@yandex.ru) 03.09.2013, <15:37> Sorry for my terrible english... -- 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: Cygwin 1.7.25
Chris Sutcliffe gmail.com> writes: > I've checked 5 different mirrors again this morning, and given that > the release was 3 days ago I don't think it's a sync issue. 1.7.25 > seems to be missing on the mirrors for x86. The files have been there all the time, but setup.ini wasn't updated. That is now fixed, depending on when the mirrors update you may or may not see it yet. Regards, Achim. -- 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: Cygwin 1.7.25
Corinna Vinschen cygwin.com> writes: > - Fix bug in sbrk breaking extending the heap beyond its default size. > Fixes: http://cygwin.com/ml/cygwin/2013-08/msg00520.html I confirm that this problem is fixed and the memory allocation from xz grows correctly beyond the default heap and is using the already commited memory now. Thank you very much. Regards, Achim. -- 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: Cygwin 1.7.25
On 2 September 2013 21:36, Larry Hall (Cygwin) wrote: > On 9/2/2013 11:34 AM, Angelo Graziosi wrote: >> >> Corinna Vinschen wrote: >>> >>> I just released Cygwin 1.7.25. >> >> Perhaps I am misunderstanding something.. but is 1.7.25 for x86_64 only or >> also for x86 systems? The mirrors have the 1.7.25 64 bit but not the 32 >> bit. >> In this case, it is still 1.7.24 the "release".. > > Releases now are for both platforms. Check another mirror or wait a bit if > what you're seeing doesn't match my statement. I've checked 5 different mirrors again this morning, and given that the release was 3 days ago I don't think it's a sync issue. 1.7.25 seems to be missing on the mirrors for x86. Chris -- Chris Sutcliffe http://emergedesktop.org http://www.google.com/profiles/ir0nh34d -- 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
