gmc has ceased working here; the reason has me mystified.
In desperation, and after trying other thigs, I've downloaded
mc-4.5.40.tar.gz and built it with -g to get debugging info into it.
I'm examinging the core file in xxgdb right now, and it's not illuminating
me very much.
Here are some select bits of info:
(xxgdb) where
#0 0x405564b5 in internal_nis_getgrent_r (grp=0x40514a44,
buffer=0x8135388 "staroffice:x:521:", buflen=1024, errnop=0x405138c0) at
../sysdeps/i386/i486/bits/string.h:837
#1 0x405569b6 in _nss_nis_getgrent_r (result=0x40514a44, buffer=0x8135388
"staroffice:x:521:", buflen=1024, errnop=0x405138c0) at
nss_nis/nis-grp.c:153
#2 0x404b0b0a in __getgrent_r (resbuf=0x40514a44, buffer=0x8135388
"staroffice:x:521:", buflen=1024, result=0xbffff924) at
../nss/getXXent_r.c:248
#3 0x404b037f in getgrent () at ../nss/getXXent.c:92
#4 0x80946f5 in init_groups () at ../../gnome/utilunix.c:117
#5 0x80685c0 in main (argc=1, argv=0xbffff9b4) at ../../gnome/main.c:2971
#6 0x404303c3 in __libc_start_main (main=0x8068590 <main>, argc=1,
argv=0xbffff9b4, init=0x8062b9c <_init>, fini=0x80e2fdc <_fini>,
rtld_fini=0x4000bec0 <_dl_fini>, stack_end=0xbffff9ac) at
../sysdeps/generic/libc-start.c:90
#6 I take to be in the runtime initialisation; assume it's okay.
#5 is part of mc. Here's the relevant code.
int
main (int argc, char *argv [])
{
/* We had LC_CTYPE before, LC_ALL includs LC_TYPE as well */
setlocale (LC_ALL, "");
bindtextdomain ("mc", LOCALEDIR);
textdomain ("mc");
mad_init ();
#if 0
/* This is here to debug startup stuff */
{
volatile int i = 0;
printf ("GMC IS WAITING %d\n", getpid ());
while (!i)
;
}
#endif
/* Initialize list of all user group for timur_clr_mode */
init_groups ();
I don't have a very good idea of what all this is about, but it does look
plausible. It gets into init_groups (), so let's look at that:
void init_groups (void)
{
int i;
struct passwd *pwd;
struct group *grp;
user_in_groups *cug, *pug;
pwd = getpwuid (current_user_uid=getuid ());
current_user_gid = (pug = g_new (user_in_groups, 1));
current_user_gid->gid = getgid ();
current_user_gid->next = NULL;
if (pwd == NULL)
return;
setgrent ();
while ((grp = getgrent ()))
for (i = 0; grp->gr_mem[i]; i++)
The stack trace tells me it went down in setgrent (); there are no pinters
passed to it to get wrong....
Here are the locals, only slightly tampered with:
(xxgdb) info locals
i = 0
pwd = (struct passwd *) 0x40514bbc
grp = (struct group *) 0x40514a44
cug = (user_in_groups *) 0x40511a00
pug = (user_in_groups *) 0x8135298
(xxgdb) print *pwd
$6 = {
pw_name = 0x81339a8 "summer",
pw_passwd = 0x81339af "youwish",
pw_uid = 500,
pw_gid = 500,
pw_gecos = 0x81339c5 "John Summerfield,,,,",
pw_dir = 0x81339da "/u02/summer",
pw_shell = 0x81339e6 "/bin/bash"
}
(xxgdb) print *grp
$7 = {
gr_name = 0x8135388 "staroffice:x:521:",
gr_passwd = 0x813538e "fice:x:521:",
gr_gid = 235,
gr_mem = 0x8135398
}
(xxgdb) print *cug
$8 = {
next = 0x405119f8,
gid = 1079056888
}
(xxgdb) print *pug
$9 = {
next = 0x0,
gid = 500
}
(xxgdb)
Now, the info in *grp is clearly junk; the first entry in my group file
contains "staroffice:x:521:"
WHat now?
I don't really want to rebuild glibc to include debugging info(I have
these installed:
[summer@possum summer]$ rpm -qa | grep glibc
glibc-devel-2.1.2-1
compat-glibc-5.2-2.0.7.1
glibc-profile-2.1.2-1
glibc-2.1.2-1
[summer@possum summer]$
).
Hmm
Perhaps I should revert to 2.1.1; I'll do that and let you know if it's
better.
CU
--
Cheers
John Summerfield
http://os2.ami.com.au/os2/ for OS/2 support.
Configuration, networking, combined IBM ftpsites index.
--
To unsubscribe:
mail -s unsubscribe [EMAIL PROTECTED] < /dev/null
