Heh. True, but per-thread might not even be the appropriate partitioning.
The use of pools can provide all the lifetime and scoping controls that we
need, per my other email.
Bill: did you sign up on this list? Dumbass. Just looking to get flooded
with even more stuff? :-) hehe...
Cheers,
-g
On Wed, Aug 16, 2000 at 05:25:09PM -0700, Bill Tutt wrote:
> Thread local storage would do the trick without requiring all of the
> prototypes to change.
> (This is what COM's GetErrorInfo() does.)
>
> FYI,
> Bill
>
> -----Original Message-----
> From: Greg Stein [mailto:gstein@lyra.org]
>
> On Wed, Aug 16, 2000 at 04:48:12PM -0700, Daniel L. Rall wrote:
> >...
> > Would a global pool drawn on by callers of svn_error constructors be out
> > of the question?
>
> Yup. That would presumably be stashed into a global variable.
>
> Bam! Now you just lost your thread safety. :-)
>
> SVN "needs" to be thread-safe. (I put quotes because I'm not sure if that
> has been articulated formally yet; it is my opinion, tho :-)
>
> Cheers,
> -g
>
> --
> Greg Stein, http://www.lyra.org/
--
Greg Stein, http://www.lyra.org/
Received on Sat Oct 21 14:36:06 2006