[prev in list] [next in list] [prev in thread] [next in thread] 

List:       sylpheed
Subject:    [sylpheed:21827] Re: Frequent segfaults on corrupted messages when
From:       cognite () zianet ! com
Date:       2003-10-29 1:45:17
[Download RAW message or body]

if you want to avoid [other] hangs [a crash where the screen stays intact but
NOTHING changes or can be done except a cold boot, often requiring lots of
inode changes since the partition will not have been closed properly], you
might want to avoid SuSE8.2Pro.  Their tech staff has no explanation.  The
hang often happens during mail use because it can happen any time, as far
as I can ascertain, but most often occurs when you are doing the things
you do most often!



On Tue, 28 Oct 2003 21:13:12 -0500
Allen Ziegenfus <arzmail@mindspring.com> wrote:

> I have been using sylpheed for a couple months now, all the time
> experiencing frequent crashes when opening sylpheed or opening certain
> emails. The cause of this seems to be spam that has junk characters in
> it like the following:   òÒ»ËͶþ. It's gotten so bad that I would have
> to sort through my /var/spool/mail file looking for these type of
> messages and deleting them in order to get sylpheed to run. Finally I
> decided to do something about it!
> 
> I think I've traced the problem to the ISO-8859-1 char set library in
> glibc. Here is a sample stack trace: 
> 
> #0  0x200007075b0 in gconv () from /usr/lib/gconv/ISO8859-1.so
> #1  0x200007799d0 in gconv () from /usr/lib/gconv/EUC-CN.so
> #2  0x20000551e54 in __gconv (cd=0x1202de500, inbuf=0x11fff9cc8,
>     inbufend=0xfffff452202d3ca7 <Address 0xfffff452202d3ca7 out of
> bounds>,    outbuf=0x11fff9cd0, outbufend=0x0, irreversible=0x11fff9c70)
> at gconv.c:63
> #3  0x20000551188 in iconv (cd=0x1202ddc50, inbuf=0x11fff9cc8,
>     inbytesleft=0x11fff9cc0, outbuf=0x11fff9cd0,
> outbytesleft=0x11fff9cc4)    at iconv.c:53
> #4  0x1201031a0 in conv_iconv_strdup (inbuf=0x12 <Address 0x12 out of
> #bounds>,
>     src_code=0x1202e6520 "________", dest_code=0x120155d4f "ISO-8859-1")
>     at codeconv.c:845
> #5  0x120102c9c in conv_codeset_strdup (inbuf=0x1202d3ca0 " òÒ»ËͶþ",
>     src_code=0x11fff9d50 "GB2312", dest_code=0x0) at codeconv.c:708
> 
> In this case the message contained some encoded text that was using the
> GB2312 charset. After spending some time debugging the code I realized
> that if I modify the code so that conv_codeset_strdup is called with a
> value of "US_ASCII" for dest_code instead of NULL, that everything works
> fine. This amounts to changing the default charset for these operations.
> 
> I don't know much about these charsets.. Is there an easier way of
> changing the default charset? I tried fiddling with the options in the
> menus to no avail. 
> 
> Anybody seen this problem or know if it was fixed? I'm hoping that
> when I upgrade to a more recent distro this problem will go away. I'm
> running a rather old Red Hat 7.1 on a DEC Alpha PC164LX. glibc version
> is 2.2.3-11 and sylpheed version is the latest, 0.9.7. Can't wait to
> move to gentoo when I have time.
> 
> Don't know enough about signal handling but is there someway to insulate
> sylpheed from these types of crashes? It certainly didn't give me a good
> impression of sylpheed till I realized the problem was somewhere else. 
> 
> Allen
> 
> 


-- 
 ;););) blessed are the geeks -- for they shall inherit the source code :):):) 

[prev in list] [next in list] [prev in thread] [next in thread]