[quake3] 'Remove Useless setuid code' ???

[Ian Kumlien]

On Sat, 2005-09-03 at 15:41 +0200, vincent at cojot.name wrote:
> Just saw a bunch of 'Remove Useless setuid code' in svn.
> 
> Could we know why these were removed and also more generally what kind of 
> approval/testing does a new patch get before it's accepted in svn?

Well my patch had a small oopsie in it... But thats fixed now =)

> Can we at least see and discuss these patches first on the mailing list?

I assume thats what this ml is for, i mainly did a cleanup thats why i
sent to zakk directly...

> Maybe this kind of 'getuid() == saved_euid)' isn't needed on Linux but it 
> might be useful on other platforms. Can we please take care not to change 
> too much the actual code that worked? If the goal of the project isn't to 
> remain stable and extended the cross-platformability of the engine, Please 
> say so, that's fine with me: I'll just fork a new project where I'll 
> invest time in the non-Linux ports..

I thought that the setuid was to change resolution in the old days or
so? I'll check it if still works on my laptop (since it can't use sdl
due to sdl actually deadlocking X on debian sid). I don't have access to
anything beyond x86/amd64 that could run this so, i can't test.

> When I worked on i-o-q2, many of the patches that got into CVS were tested 
> on Linux/x86 only and immediately found their way into CVS where I had to 
> fix them afterwards while I always backported my Solaris fixes to the 
> generic UNIX tree. In the recent years, I ported the TGE engine to 
> Solaris/Sparc and some things were never fixed because it was/is/will 
> mostly x86-only 'oriented'.

Well, it should be portable imho, but with C you have to make a effort
to make it work on f.ex. amd64 and ia32 and still break it on others
(ppl should know about endian by now imho)

As a side note, perhaps zakk should change the homepage to say that
patches should be sent here?

PS. the first section seems to use root access to be able to access the
sound, so it could really be related to the things we now use groups to
access (or atleast most unixes does this).
DS.
-- 
Ian Kumlien <pomac () vapor ! com> -- http://pomac.netswarm.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
URL: <http://icculus.org/pipermail/quake3/attachments/20050903/7a7859ca/attachment.pgp>