Joerg Sonnenberger wrote:
> On Mon, May 28, 2007 at 11:09:42AM -0400, Johnny C. Lam wrote:
>> Does this patch fix the problem?  It does two things:
>>
>> (1) Move the meat of the bootstrap-depends target entirely within the 
>> lock, which I think is what was originally intended.
>>
>> (2) Avoid creating ${WRKDIR}/.depends when running the bootstrap-depends 
>> target, which is run as part of running the fetch target.  We still 
>> create the .depends file later when we need it.
> 
> Looks good. One note though, as soon as locking is used, this problem
> will reappear. I would suggest the script to actually set a WRKOBJDIR.

Yes, that's true, as the locking code will want to create a lockfile 
under ${WRKDIR} somewhere.  I guess we should have bootstrap-depends use 
slightly different locking code that uses somewhere under ${TMPDIR} 
rather than ${WRKDIR} for the location of the lockfile.  Does that sound 
better?

In the meantime, I'll commit my current patch if you have no objections, 
and we'll fix the locking in a little while.

	Cheers,

	-- Johnny Lam <jlam@pkgsrc.org>