Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 6890 - texmacs-1.0.0.14 segfaults with gcc 3.2
Summary: texmacs-1.0.0.14 segfaults with gcc 3.2
Status: RESOLVED FIXED
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: New packages (show other bugs)
Hardware: x86 Linux
: High normal (vote)
Assignee: Matthew Turk (RETIRED)
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2002-08-22 11:57 UTC by Jeremiah Mahler
Modified: 2003-02-04 19:42 UTC (History)
0 users

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Jeremiah Mahler 2002-08-22 11:57:49 UTC
Texmacs installs properly and even starts properly but then segfaults with a
small amount of usage.

I tried compiling without optimizations (in make.conf) but it did not make any
difference.
Comment 1 Seemant Kulleen (RETIRED) gentoo-dev 2002-08-22 22:51:40 UTC
any chance you could submit an strace?
Comment 2 Jeremiah Mahler 2002-08-22 23:40:19 UTC
Here is the tail of an strace, crashed when I tried to save the file:

read(3, "\2E5]\37\250]\325K\0\0\0\3\0\340\2\0\0\0\0{\2\343\1w\2"..., 32) = 32
brk(0x8920000)                          = 0x8920000
open("/home/jeri/practice.tm", O_WRONLY|O_CREAT|O_TRUNC, 0666) = 4
fstat64(4, {st_mode=S_IFREG|0644, st_size=0, ...}) = 0
old_mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =
0x40017000
write(4, "<TeXmacs|1.0.0.15>\n\n<style|artic"..., 2071) = 2071
close(4)                                = 0
munmap(0x40017000, 8192)                = 0
--- SIGSEGV (Segmentation fault) ---
+++ killed by SIGSEGV +++

There were also lots of message like these during compilation:

c++ -I./Basic/Data -I./Basic/Misc -I./Basic/System -I./Basic/Types -Wall
-Wno-return-type -O3 -fexpensive-optimizations -fno-exceptions -c
./Basic/Types/array.cc -o Objects/array.o
In file included from /usr/include/g++-v32/backward/iostream.h:31,
                 from Basic/Types/basic.hh:18,
                 from Basic/Types/array.hh:16,
                 from Basic/Types/array.cc:15:
/usr/include/g++-v32/backward/backward_warning.h:32:2: warning: #warning This
file includes at least one deprecated or antiquated header. Please consider
using one of the 32 headers found in section 17.4.1.2 of the C++ standard.
Examples include substituting the <X> header for the <X.h> header for C++
includes, or <sstream> instead of the deprecated header <strstream.h>. To
disable this warning use -Wno-deprecated.

Segfaults after compiling with a new compiler (gcc-3.2) that has issues with
backwards compatibility along with warnings about problems of using depreceated
headers, interesting.
Comment 3 Jeremiah Mahler 2002-08-23 11:10:55 UTC
Here the response to an email I sent regarding this problem:

From: Joris van der Hoeven <TeXmacs@math.u-psud.fr>
Subject: Re: TeXmacs and gcc-3.2, any problems?
To: Jeremiah Mahler <jmahler@pacbell.net>
Cc: bugs@texmacs.org


> Quick question. Have there been any problems with TeXmacs when built
> using gcc-3.2? I have built it with gcc-3.2 and it segfaults with
> minor use.

I have not tried yet. If it segfaults, then this
probably means that gcc-3.2 is bugged again :^(((
I still use gcc-2.95.3 myself, but it also seemed
to work fine with gcc-3.1.

Joris
Comment 4 Jeremiah Mahler 2002-08-25 00:08:08 UTC
Until this problem is fixed, the binary can be used:

ftp://ftp.texmacs.org/pub/TeXmacs/targz/TeXmacs-1.0.0.15-static-gnu-linux.tar.gz

It works perfectly.
Comment 5 Pierre-Henri Jondot 2002-09-03 14:22:06 UTC
I am afraid I must confirm texmacs built with gcc 3.2 segfaults...

Yet I have another texmacs related problem (if I am sure I am not the only one
to experience this, it will be good then to open a new bug about it I think)

When using texmacs as an ordinary user, some fonts can't be loaded :
ecti10.600pk and ecsl10.600pk, no problem when launching texmacs as root.
Majority of fonts load just fine yet...

Could someone else check this : 
  launch texmacs as an ordinary user, and click and the italic E on the context
dependant icon (text mode) and see the standard output.

I get :
TeXmacs] Loading cmr10 at 600 dpi
TeXmacs] Loading ecti10 at 600 dpi
Warning: I could not find the TeX font ecti10 at 600 dpi
         I will use the standard font cmr10 instead
when I am a standard user
(and of course, the output is certainly not italic text...)

and 
TeXmacs] Loading cmr10 at 600 dpi
TeXmacs] Loading ecti10 at 600 dpi

as root.
(normal output there of course)

Of course I checked the permissions... I have absolutely not the slightest idea
about why some fonts just won't load as a normal user.

I precise this is not gcc 3.2 specific. Same problem with a gentoo organized
aroung gcc 3.1 compiler...
AND : 
the statically built texmacs binary from ftp.texmacs.org behaves identically !

Could someone check this ?
Comment 6 Matthew Turk (RETIRED) gentoo-dev 2002-09-05 21:28:11 UTC
Ok, has anyone tried building TeXmacs 1.0.0.15 with GCC 3.2?  And for those of 
you that have had problems with .14 and GCC 3.2, did you use a libsigc++ built 
with GCC 3.2, or were you using the new GCC with the old std C++ lib?  ABI is 
not compatible between the two, and could result in segfaults.  This bug is 
critical for 1.4, so it is getting high priority from me. 
Comment 7 Pierre-Henri Jondot 2002-09-07 14:13:16 UTC
Yes, I can confirm that Texmacs 1.0.0.15 as well as 1.0.0.14 does segfault when
built with gcc 3.2. (I just copied the 1.0.0.14 ebuild to texmacs-1.0.0.15 in
order to test it)

Tested on a fresh 1.4_rc1 gentoo install. (so with everything built with gcc 3.2.).

This is not gentoo specific... It is a problem, I was told, for the inclusion of
texmacs in the next (gcc 3.2 based) version of mandrake too. 

As concerns my problem regarding fonts, it might be solved by the installation
of the fonts package to /usr/share/TeXmacs-1.0.0.??/fonts rather than
/usr/share/texmf/fonts.
Comment 8 Matthew Turk (RETIRED) gentoo-dev 2002-09-12 11:52:26 UTC
Ok, since this is upstream, I'm going to close it with WONTFIX.  But if the 
fonts issue is still applicable, please open a new bug for me to deal with.  
:)