Re: Recursive directories.

[CyberPsychotic <fygrave@tigerteam.net>]

~ > 2.5 I noticed that if you do rm -rf AA*, instead of rm -rf
~ > AAAAAAAAAAAAAAAA...AA, it doesn't complain and does the job cleanly.
~ > Another side not: I noticed that when you fire rm in gdb, it doesn't pull
~ > bad address, but rather returns `long file name' error instead.
~ 
~ Not a lib issue (since rm is statically linked) but a kernel issue
~ (new system calls).  You could grab the fts.{c,h} from the current
~ source tree and build an rm on the 2.5 system that should work.
~ 

Oops.. thanks. I should have noticed that this might be a kernel's syscall
issue, since on 2.5 rm gives `Bad address' errno, which is supposed to be
returned, if erroneous pointers are passed to syscall. (if I read manuals
right :-)). I am still unable to cause `rm' on 2.5 to coredump though (which
would mean that rm tried to access some bogus address from within the
userland, sort of different thing IMHO).