[BusyBox] Patch: sh.c running internal commands

Erik Andersen andersen at lineo.com
Fri Jul 28 16:31:43 UTC 2000


On Fri Jul 28, 2000 at 11:33:53AM -0400, Pavel Roskin wrote:
> Hello, Erik!
> 
> > I actually found that one last night as well (sh.c has been under heavy
> > construction -- I now have if-then-else-fi basically working, and I switched
> > the parser to using getopt) , but I've been on my laptop so I havn't gotten
> > that one into CVS yet.  I'll checkin my work now, and then start integrating
> > the rest of yesterdays patches.
> 
> Two observations:
> 1) "busybox sh" crashes
> 
> gdb) set args sh
> (gdb) run
> Starting program: /home/proskin/src/busybox/busybox sh
> 
> Program received signal SIGSEGV, Segmentation fault.
> 0xff4172c in execvp () at execvp.c:52
> execvp.c:52: No such file or directory.
> (gdb) where
> #0  0xff4172c in execvp () at execvp.c:52
> #1  0x10021738 in delete_module ()
> #2  0x10029734 in swapoff ()
> #3  0x10029880 in swapoff ()
> #4  0x10029734 in swapoff ()
> #5  0xfedf7d4 in __libc_start_main () at
> ../sysdeps/powerpc/elf/libc-start.c:106
> (gdb)
> 
> 2) Your patch is setting applet_name even if
> BB_FEATURE_SH_STANDALONE_SHELL is not defined. I'm not sure it's correct.
> Ken's patch was more accurate (although it produced an easy-to-eliminate
> warning)
> 
> I'll try to elaborate more. It's just a quick message to warn you.

Yup, it is exec'ing stuff when it is not supposed to be.  Noticed that on the
drive into work.   Sorry about that (and welcome to the bleeding edge).  I'll
try to get that fixed right up,

 -Erik

--
Erik B. Andersen   email:  andersen at lineo.com
--This message was written using 73% post-consumer electrons--





More information about the busybox mailing list