Emacs 24.0.90 Pretest Windows Binaries

Emacs 24.0.90 Pretest Windows Binaries

[Christoph Scholtes]

The Emacs 24.0.90 Pretest Windows Binaries have been published in

http://alpha.gnu.org/gnu/emacs/pretest/windows/

Re: Emacs 24.0.90 Pretest Windows Binaries

[Eli Zaretskii]

> Date: Tue, 18 Oct 2011 18:43:09 -0600
> From: Christoph Scholtes <cschol2112@googlemail.com>
> Cc: help-emacs-windows@gnu.org
> 
> The Emacs 24.0.90 Pretest Windows Binaries have been published in
> 
> http://alpha.gnu.org/gnu/emacs/pretest/windows/

Thank you!

RE: Emacs 24.0.90 Pretest Windows Binaries

[Drew Adams]

> The Emacs 24.0.90 Pretest Windows Binaries have been published in
> http://alpha.gnu.org/gnu/emacs/pretest/windows/

Thanks.

Re: Emacs 24.0.90 Pretest Windows Binaries

[Kevin]

Christoph Scholtes <cschol2112 <at> googlemail.com> writes:

> The Emacs 24.0.90 Pretest Windows Binaries have been published in
> 
> http://alpha.gnu.org/gnu/emacs/pretest/windows/
Hi,
  Have you tried to turn on the -O2 option when compiling emacs with MinGW
gcc-4.6.1 on WindowsXP? Emacs crashes when you press Ctrl-g. Win7 doesn't have
this issue.

Re: Emacs 24.0.90 Pretest Windows Binaries

[Richard Stallman]

      Have you tried to turn on the -O2 option when compiling emacs with MinGW
    gcc-4.6.1 on WindowsXP? Emacs crashes when you press Ctrl-g. Win7 doesn't have
    this issue.

I can't help with the substance of the problem, but please don't call
Windows "win" -- that term praises it.

-- 
Dr Richard Stallman
President, Free Software Foundation
51 Franklin St
Boston MA 02110
USA
www.fsf.org  www.gnu.org
Skype: No way! That's nonfree (freedom-denying) software.
  Use free telephony http://directory.fsf.org/category/tel/

Re: Emacs 24.0.90 Pretest Windows Binaries

[Eli Zaretskii]

> From: Kevin <yujie052@gmail.com>
> Date: Wed, 19 Oct 2011 08:52:18 +0000 (UTC)
> 
>   Have you tried to turn on the -O2 option when compiling emacs with MinGW
> gcc-4.6.1 on WindowsXP? Emacs crashes when you press Ctrl-g.

Please show a backtrace from this crash.

Also, where did you download the MinGW port of GCC 4.6.1?  There are
several sites offering this port, so it's important to know your
source.  Please also tell which versions of MinGW runtime and the
w32api you used to build Emacs.

Re: Emacs 24.0.90 Pretest Windows Binaries

[Kevin Yu]

On Wed, Oct 19, 2011 at 7:53 PM, Eli Zaretskii <eliz@gnu.org> wrote:
>> From: Kevin <yujie052@gmail.com>
>> Date: Wed, 19 Oct 2011 08:52:18 +0000 (UTC)
>>
>>   Have you tried to turn on the -O2 option when compiling emacs with MinGW
>> gcc-4.6.1 on WindowsXP? Emacs crashes when you press Ctrl-g.
>
> Please show a backtrace from this crash.
gdb says corrupt stack. There's no windows-xp machine nearby. I'll
give you the backtrace tomorrow.
> Also, where did you download the MinGW port of GCC 4.6.1?  There are
> several sites offering this port, so it's important to know your
> source.  Please also tell which versions of MinGW runtime and the
> w32api you used to build Emacs.

I installed mingw-gcc 4.6.1 by mingw-get. Here's the package version:
MinGW runtime: 3.20
w32api: 3.17-2

Re: Emacs 24.0.90 Pretest Windows Binaries

[Eli Zaretskii]

> Date: Wed, 19 Oct 2011 21:42:29 +0800
> From: Kevin Yu <yujie052@gmail.com>
> Cc: emacs-devel@gnu.org
> 
> gdb says corrupt stack. There's no windows-xp machine nearby. I'll
> give you the backtrace tomorrow.

Thanks.  Please be sure to use the latest GDB available from the MinGW
site.

Re: Emacs 24.0.90 Pretest Windows Binaries

[Kevin Yu]

On Wed, Oct 19, 2011 at 11:13 PM, Eli Zaretskii <eliz@gnu.org> wrote:
> Thanks.  Please be sure to use the latest GDB available from the MinGW
> site.
>

gdb backtrace:

Environment variable "DISPLAY" not defined.
TERM = dumb
Breakpoint 1 at 0x10bb0ec: file w32fns.c, line 7180.
Temporary breakpoint 2 at 0x10c2f56: file sysdep.c, line 855.
(gdb) r -q
Starting program: d:\emacs\src\oo-spd\i386\emacs.exe -q
[New Thread 4352.0x105c]
[New Thread 4352.0xac8]

Program received signal SIGSEGV, Segmentation fault.
0x77c0554a in msvcrt!_abnormal_termination ()
   from C:\windows\system32\msvcrt.dll
(gdb) bt
#0  0x77c0554a in msvcrt!_abnormal_termination ()
   from C:\windows\system32\msvcrt.dll
#1  0x77c09bc6 in strerror () from C:\windows\system32\msvcrt.dll
Backtrace stopped: previous frame inner to this frame (corrupt stack?)
(gdb)


C:\Documents and Settings\user>gdb --version
GNU gdb (GDB) 7.3
Copyright (C) 2011 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "mingw32".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>.

C:\Documents and Settings\user>mingw-get upgrade gdb
mingw-get: *** ERROR *** upgrade gdb: package is not installed
upgrade: gdb-7.3-2-mingw32-bin.tar.lzma
mingw-get: *** INFO *** package gdb-7.3-2-mingw32-bin.tar.lzma is up to date

Re: Emacs 24.0.90 Pretest Windows Binaries

[Eli Zaretskii]

> Date: Thu, 20 Oct 2011 10:16:16 +0800
> From: Kevin Yu <yujie052@gmail.com>
> Cc: emacs-devel@gnu.org
> 
> gdb backtrace:
> 
> Environment variable "DISPLAY" not defined.
> TERM = dumb
> Breakpoint 1 at 0x10bb0ec: file w32fns.c, line 7180.
> Temporary breakpoint 2 at 0x10c2f56: file sysdep.c, line 855.
> (gdb) r -q
> Starting program: d:\emacs\src\oo-spd\i386\emacs.exe -q
> [New Thread 4352.0x105c]
> [New Thread 4352.0xac8]
> 
> Program received signal SIGSEGV, Segmentation fault.
> 0x77c0554a in msvcrt!_abnormal_termination ()
>    from C:\windows\system32\msvcrt.dll
> (gdb) bt
> #0  0x77c0554a in msvcrt!_abnormal_termination ()
>    from C:\windows\system32\msvcrt.dll
> #1  0x77c09bc6 in strerror () from C:\windows\system32\msvcrt.dll
> Backtrace stopped: previous frame inner to this frame (corrupt stack?)
> (gdb)

Thanks.  But this is only for one thread; as you see above, there are
others.  "info threads" will show all the threads; please produce a
backtrace from all of them.  To switch to thread number N, type
"thread N" at the GDB prompt.

TIA

Re: Emacs 24.0.90 Pretest Windows Binaries

[Kevin Yu]

> Thanks.  But this is only for one thread; as you see above, there are
> others.  "info threads" will show all the threads; please produce a
> backtrace from all of them.  To switch to thread number N, type
> "thread N" at the GDB prompt.

(gdb) info threads
  Id   Target Id         Frame
  2    Thread 5748.0x112c 0x7c92e514 in ntdll!LdrAccessResource ()
   from C:\windows\system32\ntdll.dll
* 1    Thread 5748.0x214 0x77c0554a in msvcrt!_abnormal_termination ()
   from C:\windows\system32\msvcrt.dll
(gdb) thread 2
[Switching to thread 2 (Thread 5748.0x112c)]
#0  0x7c92e514 in ntdll!LdrAccessResource ()
   from C:\windows\system32\ntdll.dll
(gdb) bt
#0  0x7c92e514 in ntdll!LdrAccessResource ()
   from C:\windows\system32\ntdll.dll
#1  0x77d191be in USER32!GetProcessWindowStation ()
   from C:\windows\system32\user32.dll
#2  0x77d2776b in USER32!GetMessageA () from C:\windows\system32\user32.dll
#3  0x010bbf37 in w32_msg_pump (msg_buf=<optimized out>) at w32fns.c:2246
#4  0x010bc5ac in w32_msg_worker@4 (arg=0x0) at w32fns.c:2465
#5  0x7c80b729 in KERNEL32!GetModuleFileNameA ()
   from C:\windows\system32\kernel32.dll
#6  0x00000000 in ?? ()
(gdb)

Thanks.

Re: Emacs 24.0.90 Pretest Windows Binaries

[Eli Zaretskii]

> Date: Thu, 20 Oct 2011 17:11:04 +0800
> From: Kevin Yu <yujie052@gmail.com>
> Cc: emacs-devel@gnu.org
> 
> > Thanks.  But this is only for one thread; as you see above, there are
> > others.  "info threads" will show all the threads; please produce a
> > backtrace from all of them.  To switch to thread number N, type
> > "thread N" at the GDB prompt.
> 
> (gdb) info threads
>   Id   Target Id         Frame
>   2    Thread 5748.0x112c 0x7c92e514 in ntdll!LdrAccessResource ()
>    from C:\windows\system32\ntdll.dll
> * 1    Thread 5748.0x214 0x77c0554a in msvcrt!_abnormal_termination ()
>    from C:\windows\system32\msvcrt.dll
> (gdb) thread 2
> [Switching to thread 2 (Thread 5748.0x112c)]
> #0  0x7c92e514 in ntdll!LdrAccessResource ()
>    from C:\windows\system32\ntdll.dll
> (gdb) bt
> #0  0x7c92e514 in ntdll!LdrAccessResource ()
>    from C:\windows\system32\ntdll.dll
> #1  0x77d191be in USER32!GetProcessWindowStation ()
>    from C:\windows\system32\user32.dll
> #2  0x77d2776b in USER32!GetMessageA () from C:\windows\system32\user32.dll
> #3  0x010bbf37 in w32_msg_pump (msg_buf=<optimized out>) at w32fns.c:2246
> #4  0x010bc5ac in w32_msg_worker@4 (arg=0x0) at w32fns.c:2465
> #5  0x7c80b729 in KERNEL32!GetModuleFileNameA ()
>    from C:\windows\system32\kernel32.dll
> #6  0x00000000 in ?? ()
> (gdb)

Hmm... looks like thread 1 is the one that crashes.  Your previous
message indicated that it crashes inside strerror:

  Program received signal SIGSEGV, Segmentation fault.
  0x77c0554a in msvcrt!_abnormal_termination ()
     from C:\windows\system32\msvcrt.dll
  (gdb) bt
  #0  0x77c0554a in msvcrt!_abnormal_termination ()
     from C:\windows\system32\msvcrt.dll
  #1  0x77c09bc6 in strerror () from C:\windows\system32\msvcrt.dll

Interestingly enough, I cannot put a breakpoint on that function in my
Emacs binary: GDB says it is not linked in.  Can you put a breakpoint
on strerror?

Also, do I understand correctly that to crash Emacs, all you need is
start it with "emacs -Q" and type "C-g"?  Or is something else
required?

Re: Emacs 24.0.90 Pretest Windows Binaries

[Kevin Yu]

On Thu, Oct 20, 2011 at 10:03 PM, Eli Zaretskii <eliz@gnu.org> wrote:

> Interestingly enough, I cannot put a breakpoint on that function in my
> Emacs binary: GDB says it is not linked in.  Can you put a breakpoint

I can't set a breakpoint either.

> on strerror?
>
> Also, do I understand correctly that to crash Emacs, all you need is
> start it with "emacs -Q" and type "C-g"?  Or is something else
> required?
>
Yes.

I'm using the same tool chain to build emacs on windows 7. everything goes well.
I wonder if it's a bug of mingw-gcc 4.6.1 :(. it came out after
upgrading mingw-gcc from
4.5.1 to 4.6.1

Any way, thank you for tracking this issue.

Best wishes.

Re: Emacs 24.0.90 Pretest Windows Binaries

[Eli Zaretskii]

> Date: Fri, 21 Oct 2011 09:44:23 +0800
> From: Kevin Yu <yujie052@gmail.com>
> Cc: emacs-devel@gnu.org
> 
> I wonder if it's a bug of mingw-gcc 4.6.1

This thread:

  http://lists.gnu.org/archive/html/qemu-devel/2011-10/msg01697.html

suggests that it probably is.  IIUC, there's an issue with setjmp in
the MS runtime not working when -fomit-frame-pointer is used, and the
MinGW GCC 4.6.1 turns on -fomit-frame-pointer by default when
optimizing.  Since Emacs uses setjmp/longjmp _a_lot_, and in
particular C-g throws to the top level via this mechanism, it seems
like we have a smoking gun.

Perhaps rebuild Emacs with -fno-omit-frame-pointer, and see if that
helps.

Re: Emacs 24.0.90 Pretest Windows Binaries

[Kevin Yu]

On Fri, Oct 21, 2011 at 4:04 PM, Eli Zaretskii <eliz@gnu.org> wrote:

> Perhaps rebuild Emacs with -fno-omit-frame-pointer, and see if that
> helps.

Oh, it works. Thanks for your help. :)

Re: Emacs 24.0.90 Pretest Windows Binaries

[Kevin Yu]

[-- Attachment #1: Type: text/plain, Size: 345 bytes --]

Another question. :)
It seems like the windows font back-end assumes fonts with bold and
normal weights have same font-width. When I set the default font to a
font has different width in normal and bold weights, for example
Monaco, DejaVu Sans Mono, some of the glyphs are overlapped. The
letters 'e', 'D' in the attached picture is overlapped.

[-- Attachment #2: att1.png --]
[-- Type: image/png, Size: 1427 bytes --]

Re: Emacs 24.0.90 Pretest Windows Binaries

[Eli Zaretskii]

> Date: Fri, 21 Oct 2011 18:10:59 +0800
> From: Kevin Yu <yujie052@gmail.com>
> Cc: emacs-devel@gnu.org
> 
> It seems like the windows font back-end assumes fonts with bold and
> normal weights have same font-width. When I set the default font to a
> font has different width in normal and bold weights, for example
> Monaco, DejaVu Sans Mono, some of the glyphs are overlapped. The
> letters 'e', 'D' in the attached picture is overlapped.

I cannot reproduce this with DejaVu Sans Mono (I don't have the other
font you mention).  Can you show a reproducible recipe starting from
"emacs -Q"?

In general, Emacs doesn't "assume" anything about fonts, it receives
all the info about a font from the font itself.  At least AFAIK.

Re: Emacs 24.0.90 Pretest Windows Binaries

[Kevin Yu]

On Fri, Oct 21, 2011 at 8:12 PM, Eli Zaretskii <eliz@gnu.org> wrote:
> I cannot reproduce this with DejaVu Sans Mono (I don't have the other
> font you mention).  Can you show a reproducible recipe starting from
> "emacs -Q"?

Sorry i couldn't reproduce it with DejaVu Sans Mono either.

I got Monaco from here http://www.gringod.com/2006/02/24/return-of-monacottf/
If you do not want install the Monaco font, just forget it. :)

Here's reproducing steps:

1. emacs -Q
2. M-x load-theme misterioso
3. set the buffer (*scratch* buffer) font to Monaco (shift-left-mouse-button)
4. type in following text in *scratch* buffer
     (defun D(
    the letter D is displayed in 'bold' face but the last '(' isn't.
5. put your cursor on last '(', you will find out that the last pixel
column of 'D' is overlapped
6. if you move your cursor two chars back, the overlapped pixels come out again.

The monaco font doesn't provide a bold version. Does emacs generated
it for displaying?

Re: Emacs 24.0.90 Pretest Windows Binaries

[Eli Zaretskii]

> Date: Fri, 21 Oct 2011 22:12:29 +0800
> From: Kevin Yu <yujie052@gmail.com>
> Cc: emacs-devel@gnu.org
> 
> On Fri, Oct 21, 2011 at 8:12 PM, Eli Zaretskii <eliz@gnu.org> wrote:
> > I cannot reproduce this with DejaVu Sans Mono (I don't have the other
> > font you mention).  Can you show a reproducible recipe starting from
> > "emacs -Q"?
> 
> Sorry i couldn't reproduce it with DejaVu Sans Mono either.
> 
> I got Monaco from here http://www.gringod.com/2006/02/24/return-of-monacottf/

Yes, I see it with Monaco.

> Here's reproducing steps:
> 
> 1. emacs -Q
> 2. M-x load-theme misterioso
> 3. set the buffer (*scratch* buffer) font to Monaco (shift-left-mouse-button)
> 4. type in following text in *scratch* buffer
>      (defun D(
>     the letter D is displayed in 'bold' face but the last '(' isn't.
> 5. put your cursor on last '(', you will find out that the last pixel
> column of 'D' is overlapped
> 6. if you move your cursor two chars back, the overlapped pixels come out again.

Please submit a bug report about this, including the URL where the
font can be downloaded.  Font drivers and drawing are really not my
cup of tea in Emacs, but maybe someone could reproduce this on a Posix
system, or maybe Jason will find time to dig into this.

> The monaco font doesn't provide a bold version. Does emacs generated
> it for displaying?

Sorry, I don't know.

Re: Emacs 24.0.90 Pretest Windows Binaries

[Kevin Yu]

reported. thanks for your help.

Re: Emacs 24.0.90 Pretest Windows Binaries

[Eli Zaretskii]

> Date: Fri, 21 Oct 2011 16:45:20 +0800
> From: Kevin Yu <yujie052@gmail.com>
> Cc: emacs-devel@gnu.org
> 
> On Fri, Oct 21, 2011 at 4:04 PM, Eli Zaretskii <eliz@gnu.org> wrote:
> 
> > Perhaps rebuild Emacs with -fno-omit-frame-pointer, and see if that
> > helps.
> 
> Oh, it works. Thanks for your help. :)

Thanks, I added this issue to etc/PROBLEMS.

Re: Emacs 24.0.90 Pretest Windows Binaries

[Christoph Scholtes]

On 10/19/2011 2:52 AM, Kevin wrote:

> Hi,
>    Have you tried to turn on the -O2 option when compiling emacs with MinGW
> gcc-4.6.1 on WindowsXP? Emacs crashes when you press Ctrl-g. Win7 doesn't have
> this issue.

The pretest was compiled unoptimized with gcc 4.6.1 (tdmMinGW32) on 
Windows 7. I will try to reproduce this, if you can confirm which gcc 
distribution you used to compile. Thanks.