Emacs 23.4 released

Emacs 23.4 released

[Chong Yidong]

GNU Emacs 23.4 has been released, and is now available at
ftp.gnu.org/gnu/emacs/ and the GNU FTP mirrors (see
http://www.gnu.org/order/ftp.html).

The SHA1 checksums are:

 0a78466fff52d1f43c0db0ea66651a0b13b7e473  emacs-23.4.tar.gz
 5effbd2af6bf7a4feb2d80810e49b0ce47b463cb  emacs-23.4.tar.bz2

Please send bug reports to bug-gnu-emacs@gnu.org.  You can use the
command M-x report-emacs-bug to do this.

Emacs 23.4 mainly fixes a security flaw in EDE (CVE-2012-0035),
described at

  http://lists.gnu.org/archive/html/emacs-devel/2012-01/msg00387.html

It also contains a number of other bugfixes.

Re: Emacs 23.4 released

[Stefan Monnier]

> GNU Emacs 23.4 has been released, and is now available at
> ftp.gnu.org/gnu/emacs/ and the GNU FTP mirrors (see
> http://www.gnu.org/order/ftp.html).

Yay!  Than1


        Stefan

Re: Emacs 23.4 released

[Bastien]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> GNU Emacs 23.4 has been released, and is now available at
>> ftp.gnu.org/gnu/emacs/ and the GNU FTP mirrors (see
>> http://www.gnu.org/order/ftp.html).
>
> Yay!  Than1

+1

Than2!

-- 
 Bastien

Re: Emacs 23.4 released

[Christoph Scholtes]

Chong,

On Windows, Emacs 23.4 crashes during dumping with return code 
-1073741819 (0xC0000005, Access Violation).

The 23.3.90 built fine, but unfortunately I didn't get to build the 23.4 
RC1, which crashes also. I just did a test build. I apologize for that.

Eli,
any pointers on how to troubleshoot this would be appreciated.

"../nt/oo-spd/i386/addsection" "oo-spd/i386/temacs.bin" 
"oo-spd/i386/temacs.exe"
  EMHEAP 21
Dumping from oo-spd/i386/temacs.bin
           to oo-spd/i386/temacs.exe
"./oo-spd/i386/temacs.exe" -batch -l loadup dump
mingw32-make[1]: *** [oo-spd/i386/emacs.exe] Error -1073741819
mingw32-make[1]: Leaving directory 
`D:/devel/emacs/pretest/23.4rc1/emacs-23.4/sr
c'
mingw32-make: *** [all-other-dirs-gmake] Error 2
Using C:\Windows\system32\cmd.exe as shell.

After cancelling the crash dialog it goes on to

"./oo-spd/i386/temacs.exe" -batch -l loadup dump
mingw32-make[1]: *** [oo-spd/i386/emacs.exe] Error -1073741819
mingw32-make[1]: Leaving directory 
`D:/devel/emacs/pretest/23.4rc1/emacs-23.4/sr
c'
mingw32-make: *** [all-other-dirs-gmake] Error 2

and crashes again.

Thanks,
Christoph


On 1/28/2012 10:36 AM, Chong Yidong wrote:

> GNU Emacs 23.4 has been released, and is now available at
> ftp.gnu.org/gnu/emacs/ and the GNU FTP mirrors (see
> http://www.gnu.org/order/ftp.html).
>
> The SHA1 checksums are:
>
>   0a78466fff52d1f43c0db0ea66651a0b13b7e473  emacs-23.4.tar.gz
>   5effbd2af6bf7a4feb2d80810e49b0ce47b463cb  emacs-23.4.tar.bz2
>
> Please send bug reports to bug-gnu-emacs@gnu.org.  You can use the
> command M-x report-emacs-bug to do this.
>
> Emacs 23.4 mainly fixes a security flaw in EDE (CVE-2012-0035),
> described at
>
>    http://lists.gnu.org/archive/html/emacs-devel/2012-01/msg00387.html
>
> It also contains a number of other bugfixes.

Re: Emacs 23.4 released

[Eli Zaretskii]

> Date: Sun, 29 Jan 2012 10:53:15 -0700
> From: Christoph Scholtes <cschol2112@googlemail.com>
> CC: emacs-devel@gnu.org, Eli Zaretskii <eliz@gnu.org>
> 
> On Windows, Emacs 23.4 crashes during dumping with return code 
> -1073741819 (0xC0000005, Access Violation).
> 
> The 23.3.90 built fine, but unfortunately I didn't get to build the 23.4 
> RC1, which crashes also. I just did a test build. I apologize for that.
> 
> Eli,
> any pointers on how to troubleshoot this would be appreciated.

Does the unoptimized build (and with debug info) crash as well?  If
so, try running this command:

> "./oo-spd/i386/temacs.exe" -batch -l loadup dump

under GDB, and see where it crashes.

Re: Emacs 23.4 released

[Christoph Scholtes]

On 1/29/2012 11:32 AM, Eli Zaretskii wrote:

> Does the unoptimized build (and with debug info) crash as well?

No, --no-opt does not crash.

Does it still makes sense to run it in the debugger?

Christoph

Re: Emacs 23.4 released

[Juanma Barranquero]

On Sun, Jan 29, 2012 at 19:41, Christoph Scholtes
<cschol2112@googlemail.com> wrote:

> No, --no-opt does not crash.

Could it be related to the omit-frame-pointer problem, if you're using
a 4.x gcc? Does it build with optimizations if you use
-fno-omit-frame-pointer?

    Juanma

Re: Emacs 23.4 released

[Christoph Scholtes]

On 1/29/2012 11:51 AM, Juanma Barranquero wrote:
> Could it be related to the omit-frame-pointer problem, if you're using
> a 4.x gcc? Does it build with optimizations if you use
> -fno-omit-frame-pointer?

I tried that. Still crashes.

Backtrace from optimized build with debug symbols:

Program received signal SIGSEGV, Segmentation fault.
0x010719e8 in lookup_volume_info (root_dir=0x88fcdc "d:\\") at w32.c:2051
2051        if (xstrcasecmp (info->root_dir, root_dir) == 0)
(gdb) bt
#0  0x010719e8 in lookup_volume_info (root_dir=0x88fcdc "d:\\") at 
w32.c:2051
#1  GetCachedVolumeInformation (root_dir=0x88fcdc "d:\\") at w32.c:2114
#2  0x01071bf6 in get_volume_info (
     name=0x2d543ee "/devel/emacs/pretest/23.4rc1/emacs-23.4/etc/charsets",
     pPath=0x88fe3c) at w32.c:2198
#3  0x01071cb4 in is_fat_volume (
     name=0x2d543ec 
"d:/devel/emacs/pretest/23.4rc1/emacs-23.4/etc/charsets",
     pPath=0x88fe3c) at w32.c:2213
#4  0x01071d01 in map_w32_filename (pPath=0x0,
     name=0x2d543ec 
"d:/devel/emacs/pretest/23.4rc1/emacs-23.4/etc/charsets")
     at w32.c:2235
#5  map_w32_filename (name=<optimized out>, pPath=0x0) at w32.c:2220
#6  0x010723b5 in sys_access (
     path=0x2d543ec 
"d:/devel/emacs/pretest/23.4rc1/emacs-23.4/etc/charsets",
     mode=0) at w32.c:2618
#7  0x010a7dad in init_charset () at charset.c:2384
#8  0x0117db15 in main (argc=<optimized out>, argv=0xd92e00) at emacs.c:1743
(gdb)

Re: Emacs 23.4 released

[Eli Zaretskii]

> Date: Sun, 29 Jan 2012 12:05:59 -0700
> From: Christoph Scholtes <cschol2112@googlemail.com>
> CC: Eli Zaretskii <eliz@gnu.org>, cyd@gnu.org, emacs-devel@gnu.org
> 
> On 1/29/2012 11:51 AM, Juanma Barranquero wrote:
> > Could it be related to the omit-frame-pointer problem, if you're using
> > a 4.x gcc? Does it build with optimizations if you use
> > -fno-omit-frame-pointer?
> 
> I tried that. Still crashes.

I just tried to build Emacs 23.4, and I'm sorry to say that neither
the optimized build nor the unoptimized one crash for me.  They both
build just fine and come up without crashing.

> Backtrace from optimized build with debug symbols:
> 
> Program received signal SIGSEGV, Segmentation fault.
> 0x010719e8 in lookup_volume_info (root_dir=0x88fcdc "d:\\") at w32.c:2051
> 2051        if (xstrcasecmp (info->root_dir, root_dir) == 0)

What's the problem that causes the crash?  Is `info' or
`info->root_dir' a NULL pointer or something?

If nothing else gives a clue, diff Emacs 23.4 sources against 23.3.90,
and then apply the diffs one by one until you get the crash.  This
should pinpoint the change which causes this.

Re: Emacs 23.4 released

[Juanma Barranquero]

On Sun, Jan 29, 2012 at 20:15, Eli Zaretskii <eliz@gnu.org> wrote:

> I just tried to build Emacs 23.4, and I'm sorry to say that neither
> the optimized build nor the unoptimized one crash for me.  They both
> build just fine and come up without crashing.

Same here.

    Juanma

Re: Emacs 23.4 released

[Christoph Scholtes]

On 1/29/2012 1:35 PM, Juanma Barranquero wrote:
> On Sun, Jan 29, 2012 at 20:15, Eli Zaretskii<eliz@gnu.org>  wrote:
>
>> I just tried to build Emacs 23.4, and I'm sorry to say that neither
>> the optimized build nor the unoptimized one crash for me.  They both
>> build just fine and come up without crashing.
>
> Same here.
>
>      Juanma

Confirmed. gcc 4.5.2 builds the 23.4 release fine. Builds with gcc 4.6.1 
crash with optimizations enabled.

I will build the release with gcc 4.5.2 then. I also will verify that 
the problem persists in the trunk build with gcc 4.6.1 and switch to 
4.5.2 for the weekly builds also. I will add a note to INSTALL in the 
trunk, if applicable.

As time permits I am going to try to figure out why it is crashing.

Thanks guys.

Christoph

Re: Emacs 23.4 released

[Juanma Barranquero]

On Sun, Jan 29, 2012 at 22:23, Christoph Scholtes
<cschol2112@googlemail.com> wrote:

> Confirmed. gcc 4.5.2 builds the 23.4 release fine. Builds with gcc 4.6.1
> crash with optimizations enabled.
>
> I will build the release with gcc 4.5.2 then. I also will verify that the
> problem persists in the trunk build with gcc 4.6.1 and switch to 4.5.2 for
> the weekly builds also. I will add a note to INSTALL in the trunk, if
> applicable.

In the past, I've had assorted random crashes building with 4.5.X and
4.6.1; sometimes optimized builds, sometimes not.

I've been using 4.6.2 for a while now and so far it works beautifully.

    Juanma

Re: Emacs 23.4 released

[Christoph Scholtes]

On 1/29/2012 2:37 PM, Juanma Barranquero wrote:

> In the past, I've had assorted random crashes building with 4.5.X and
> 4.6.1; sometimes optimized builds, sometimes not.

4.6.1 does not crash with the Emacs 24 pretest and the current trunk 
when building optimized. So it must be Emacs 23 code that makes it crash.

> I've been using 4.6.2 for a while now and so far it works beautifully.

Where did you get gcc 4.6.2? I installed the latest mingw and gcc 
--version says 4.6.1.

Christoph

Re: Emacs 23.4 released

[Juanma Barranquero]

On Mon, Jan 30, 2012 at 00:50, Christoph Scholtes
<cschol2112@googlemail.com> wrote:

> Where did you get gcc 4.6.2? I installed the latest mingw and gcc --version
> says 4.6.1.

http://sourceforge.net/projects/mingw/files/MinGW/Base/gcc/Version4/gcc-4.6.2-1/

    Juanma