"Eric G . Miller" <egm2@jps.net> writes:
I can't reproduce this; I have a directory in /tmp that's owned like
this:
cush:/tmp$ ls -lR foo
foo:
total 1
drwxr-xr-x 3 509 martind 1024 Feb 14 10:16 bar/
foo/bar:
total 1
drwxr-xr-x 2 501 martind 1024 Feb 14 10:16 baz/
foo/bar/baz:
total 0
And I can open up baz just fine. I even went and did an apt-get
source of linuxlogo (and incidentally, it's probably generally a good
idea to not run apt-get source as root, since you don't need to), and
I can view all of those directories just fine, even though most of
them are owned by 501.users. I've also inspected the parts in tkdesk
that use getpw* functions (it's isolated in one file), and it checks
for that null pointer.
Could you send me a core dump, or at least the result of an strace on
tkdesk that shows this crash? (to get a good strace, I suggest you
stop tkdesk and restart it with
strace -f -ostrace.out tkdesksh /usr/bin/X11/tkdesk
and then procede to just opening the browser window to the directory
that causes it to crash and burn)