dormando at rydia.net
Wed Feb 13 02:38:00 UTC 2008
Thanks for your work on this :)
A couple comments/questions:
- We could probably live without the memory managed lists. I can think
of a couple ways to flatten the loops (right now it loops several times
during setup?) and remove the extra code, but this doesn't necessarily
bug me enough to request it be fixed.
- The code style is closer to mysql/libmemcached than the rest of
memcached. This is actually a little obnoxious.
- The patch no longer applies cleanly. I can clean it up though; will
probably apply it anyway.
- Sad that this doesn't clean up into a more generic "fetching of lists
of possible addresses then applying them in a loop". So there's still no
easy way to interface multiple individual listeners in. Oh well. Another
- Is there an easy way to add some tests for this? Given that it's
internal-only support for ipv6 multiple interface thingers, I'm kinda
doubting it, which is sad.
If any of these are worth addressing, let me know. Otherwise it's going
in with some minor style fixes for 1.2.5.
Brian Aker wrote:
> This corrects the infinite loop on a bind() failure that Tomash found.
> Brian "Krow" Aker, brian at tangent.org
> Seattle, Washington
> http://krow.net/ <-- Me
> http://tangent.org/ <-- Software
> http://exploitseattle.com/ <-- Fun
> You can't grep a dead tree.
More information about the memcached