I just recently upgraded to Linux 0.97 patchlevel 2, and after I did, I
found that xinit wasn't working anymore. I finally traced it down to my
.xinitrc file, and to the following specific bug.
If you type the following at bash under 0.97 PL2:
unset SHLVL
exec /bin/sh
The existing shell will core dump without managing to exec the prgram
you ask it to exec.
If you take the same bash binary and run under 0.97 PL 1, the shell will
*not* core dump and exec the request program instead. (This is the
correct behavior).
This behavior can be duplicated using either the MCC 0.96c /bin/bash, or
the /bin/bash on the 0.97 rootimage disk.
I haven't been able to trace it down any further, but maybe I'll have
luck later after I compile my own version of bash and after I get a
working copy of gdb under 0.97PL2.
- Ted
| Andrew Morton | Re: 2.6.24-rc6-mm1 |
| Glauber de Oliveira Costa | [PATCH 8/19] modify write_ldt function |
| Steven Rostedt | Re: Major regression on hackbench with SLUB |
| Satyam Sharma | Re: 2.6.23-rc4-mm1 |
git: | |
| Chris Ortman | [FEATURE REQUEST] git-svn format-patch |
| Bill Lear | Meaning of "fatal: protocol error: bad line length character"? |
| Scott Chacon | Git Community Book |
| Catalin Marinas | Re: [ANNOUNCE] pg - A patch porcelain for GIT |
| Richard Stallman | Real men don't attack straw men |
| frantisek holop | nptd regression in 4.2 |
| Kevin | uvm_mapent_alloc: out of static map entries on 4.3 i386 |
| Vim Visual | GRAPE cluster supercomputer + OpenBSD |
| John Stoffel | Re: [PATCH] LogFS take three |
| hooanon05 | [PATCH 62/67] aufs magic sysrq handler |
| Chris Mason | Re: [PATCH][RFC] fast file mapping for loop |
| Chris Mason | Re: [ANNOUNCE] Btrfs v0.12 released |
