Hi all,
I'd sent this out a few weeks back. I got some minor comments back from
elambert and dsal, but not much otherwise. I know we're heading to a
1.4.2 release here shortly. I'd like to try to get some fixes in over
the next couple of releases so we can get to complete pandorabuild
(presuming the contributions are acceptable).
Does this sound like a sane approach? Should I start sending up commits
for more of these small changes, followed by medium changes with tests,
followed by ... more? :)
Thanks,
- Matt
--- Begin Message ---
Hi all,
This is mainly directed to Trond/Dustin since I've discussed it most
with them on IRC, but thought I should socialize this with anyone and
everyone.
I'd started fixing a bunch of things to bring memcached inline with
Monty Taylor's 'pandora build'. I feel this is that this is a
worthwhile effort since it is certainly catching legitimate, though
maybe small, issues. If we remove the small issues, it's that less
likely we'll have big issues later.
Rather than try to drop in a big hunk o' change though, I think I should
probably do it in pieces. There are pieces which are small and obvious
and shouldn't need much more than a code review, and then there are some
that I should either get some confidence with from existing tests or
come up with tests for to establish that confidence. The small fixes
are things we can bring in before all of pandora build itself.
Does this make sense?
I started to break them into pieces. An annotated example is here:
http://github.com/ingenthr/memcached/commit/71c02e9ad3af6331555c5333caf8e38b57f20196
Thoughts?
- Matt
p.s.: sorry this has taken me so long!
--- End Message ---
