Re: Patch review completed.

On Friday, October 25, 2002, at 04:49 PM, Karl Fogel wrote:

> 1. Garrett Rooney had a patch reworking how errors and cancellation
> relate. His message implied that he didn't think it was ready
> for committing, though, so I haven't done anything with it. See
> http://subversion.tigris.org/servlets/ReadMsg?list=dev&msgNo=21716
> for more details.

the only reason i stopped working on that patch was that bill tutt
objected based on the use of a global callback/baton and it's
interaction with multithreaded clients. personally, i think
multithreaded clients can make do by using thread specific storage, and
that the performance implications haven't been proven to be a big deal.
  i'd be in favor of going with the technique my patch used for 1.0, and
if it turns out to be a problem we could change to another API later.

of course, if a context parameter gets added to all the API's anyway,
that's another way to go, but personally i prefer the way my current
patch works, with the interaction with SVN_ERR giving us free
cancelation all over the tree.

so basically, if people like the way my cancellation patch works, i'll
update it and commit it, but if not, i'll wait and let someone else
deal with it, since i don't have the motivation to add a context
parameter all over the place.

-garrett

-- 
garrett rooney                    Remember, any design flaw you're
rooneg@electricjellyfish.net      sufficiently snide about becomes
http://electricjellyfish.net/     a feature.       -- Dan Sugalski
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org