Re: [Nano-devel] Crash report - segfault reproduceable

[Nick Warne]

OK, update on this - I can reproduce the crash on two machines at work also.

First, some info of them all.

Machine at home - Slackware 10  - Linux 2.6.15 gcc version 3.3.4
Machine1 at work - Redhat 7.1 - Linux 2.4.31 - gcc version 3.4.4
machine2 at work2 - Mandrake 6.5 Linux 2.2.26 - gcc version pgcc-2.91.66 
19990314 (egcs-1.1.2 release)

The only thing common in all the machines (at least the one with the keyboard 
attached) is I disable Caps Locks key as explained.

.nanorc info
set const
set nowrap
set smooth
set nohelp

Build info:
./configure --enable-nanorc --disable-mouse --disable-color --disable-nls

At work I use win2000 puTTY 0.56 - here at home, Konsole.

Here is the crash debug of work machine1:

Running in debugger:
Program received signal SIGSEGV, Segmentation fault.
0x400dd1e6 in chunk_free (ar_ptr=0x4018ecc0, p=0x80778d0) at malloc.c:3242
3242    malloc.c: No such file or directory.
        in malloc.c
(gdb) bt
#0  0x400dd1e6 in chunk_free (ar_ptr=0x4018ecc0, p=0x80778d0) at malloc.c:3242
#1  0x400dcfc0 in __libc_free (mem=0x80778d8) at malloc.c:3154
#2  0x0805de4b in display_string (
    buf=0x8077b28 
"\t\t\t\t\t\ttesttesttesttesttesttesttesttesttesttesttesttesttesttest\t\t\t\t\t\t\t\t$
    start_col=120, len=112, dollars=120 'x') at winio.c:1885
#3  0x0805ed5c in update_line (fileptr=0x8077b10, index=0) at winio.c:2603
#4  0x0805ef8f in edit_refresh () at winio.c:2834
#5  0x08054bfd in main (argc=2, argv=0xbffffb24) at nano.c:2127
#6  0x4007c336 in __libc_start_main (main=0x80546e0 <main>, argc=2, 
ubp_av=0xbffffb24, init=0x804983c <$
    fini=0x805f694 <_fini>, rtld_fini=0x4000d25c <_dl_fini>, 
stack_end=0xbffffb1c)
    at ../sysdeps/generic/libc-start.c:129



I can't run debug on work machine2, as gdb isn't installed.

Contrary to what I said, I can get it to segfault on any machine by just 
messing about resizing the terminal window with that test file loaded - if I 
leave the terminal the same size when it segfaults, nano will _then_ segfault 
immediately on opening the test file.  Contrary also, I can get nano to 
segfault in xterm too simply by loading that test file and resizing the 
window a few times.

I turned off .nanorc file - same results.  I built just using ./configure - 
same result.

I did build nano with --enable-debug, but that produces so much verbosity, you 
can't see what is happening.

If I remove the extra tabs at the end of the long line, no crash (it was this 
that I saw caused the crash I had ages ago - editing *.php files that were 
produced on a windows box).

If you need any special debug code running, I can do.

Nick
-- 
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb