With apologies to our beloved commitfest-mace-wielding CFM, commitfest
2013-11 intentionally still contains a few open patches. I think that
CF is largely being ignored by most people now that we have CF 2014-01
in progress. If we don't want to do anything about these patches in the
immediate future, I propose we move them to CF 2014-01.
* shared memory message queues This is part of the suite involving dynamic shmem; not sure whether this is a patch that
needsmore review, or is it ready for application, or has it been superceded by later versions in the next commitfest.
Patchauthors please chime in.
* Shave a few instructions from child-process startup sequence Discussion stalled without a conclusion; opinions
divergeon whether this is a useful patch to have. My personal inclination is to drop this patch because it seems
pointless,but if someone feels otherwise I won't object. (The objection that it will break as soon as we decide to
changethe invariant about invalid sockets no longer applies because it has Asserts to that effect.) Do we really care
aboutperformance during process termination? I'd say this is mildly interesting if this code is executed for
non-authenticatedclients.
* Widening application of indices. Was this re-posted in 2014-01?
* fault tolerant DROP IF EXISTS I gave a look and it looks good for application. This wasn't superceded by a future
version,correct?
* SSL: better default ciphersuite I think we should apply this.
--
Álvaro Herrera http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services