Re: Re: New user

[Marcus Brinkmann]

On Wed, May 21, 2003 at 01:36:18PM -0400, bddebian@cox.net wrote:
> Is there any interest for someone (me?) to maintain a centralized repository 
> for things happening, TODO lists, and so forth?  If so, what should it be?  
> Savannah, Alioth, hurd.gnufans.org???

I have always made the experience that truly useful results are only to be
expected if you really get intimate with an issue and understand it very
well.

For example, for the console, the task list at savannah is pretty good.  For
other things, I wouldn't have a good idea what tasks there are unless I'd
actually read a lot of code and think about its problems.

So, compiling actual task lists rather than list of bugs etc is more a part
of the development process than something independent of it.  Thomas and
Roland can hardly be blamed for the code they wrote, but they can be blamed
a bit for the documentation they didn't wrote ;)

Just as an example.  I couldn't have told you before yesterday, but there is
a race between exec and sigkill that could leave a half-created thread
around.  Now, that can only be fixed by a new mach feature (or equivalent
feature in the Hurd on L4) that associates tasks with an accounting ID.
This is the translation of the following item in the tasks file:

  * Mach needs a facility to find out what task is the parent of
  a given task.

Plus the requirement that sigkill kills all unaccounted childs of a task.

Now, you wouldn't know all this unless you have a very good understanding
about the Hurd and read a lot of code.  The above issue is hidden in the
following comment in hurd/exec/exec.c:

      /* XXX there is a race here for SIGKILLing the process. -roland
         I don't think it matters.  -mib */

So there are several things you can do now:

You can become an expert on the Hurd and actively try to find out such
things as the above yourself.  If you then also have time to document them,
that would be wonderful.

You could become a bit of an expert and follow such discussions on the
mailing list.  The above issue was mentioned yesterday on l4-hurd, but it
was within a long and complicated discussion about how to implement exec on
L4.  The relationshi[ between this design issue and several bugs (like the
sigkill race, and the problem that tasks started by boot are not killed)
could be documented and interlinked, and this would also be useful for
developers so they don't forget about it and see how to prioritize.

Or you could focus on the things that are visible and understandable even
without being an expert on Hurd issues.  These are: bug reports, patches,
missing features, things like that.  We try to address these things, but
sometimes an issue slips threw without being dealt with properly.  For this,
the Savannah interface overs bug reporting tools and patch submitting tools,
etc.  This can be done by anyone with a little understanding and patience.

Maybe this gives you some ideas,
Marcus

-- 
`Rhubarb is no Egyptian god.' GNU      http://www.gnu.org    marcus@gnu.org
Marcus Brinkmann              The Hurd http://www.gnu.org/software/hurd/
Marcus.Brinkmann@ruhr-uni-bochum.de
http://www.marcus-brinkmann.de/