GNU bug report logs - #9600
some gdbs ignore a world-writable .gdbinit

Previous Next

Full log

View this message in rfc822 format

From: Eli Zaretskii <eliz <at> gnu.org>
To: Glenn Morris <rgm <at> gnu.org>
Cc: 9600 <at> debbugs.gnu.org, rms <at> gnu.org, svenjoac <at> gmx.de
Subject: bug#9600: 24.0.50; .gdbinit does not run
Date: Mon, 26 Sep 2011 04:02:11 -0400

> From: Glenn Morris <rgm <at> gnu.org>
> Cc: rms <at> gnu.org,  9600 <at> debbugs.gnu.org,  svenjoac <at> gmx.de
> Date: Mon, 26 Sep 2011 03:26:18 -0400
> 
> AFAICS, rms still has not said what the permissions of his gdbinit file
> were, beyond "writable". It seems this must mean "world-writable".

Most probably, because this is what I have in my bzr branches.  I
guess that's because of umask, see below.

> I tested with Debian testing's gdb "7.3-debian". A gdbinit file writable
> by me was fine, a world-writable gdbinit was rejected with
> 
>   warning: not using untrusted file ".gdbinit"

Stock GDB does not reject such a .gdbinit.  That was the gist of what
I was saying.

> This seems entirely reasonable to me

It was sufficiently controversial at the time to not make it into the
official GDB.

> Since I hope that everyone building Emacs from source does not use
> `umask 000'

I do, FWIW.  Don't ask me why, it's something I have in my .bash_login
since about forever.

> I don't think any Emacs changes are needed.

No Emacs changes are needed in any case, since RMS can chmod the file,
and bzr will not touch the write mode bits.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.