Обсуждение: CommitFest 2011-11 Post-Tryptophan Progress Report

With plenty of people in the US returning to work on Monday after a 
modest break last week, I wanted to put out an updated on the active 
CommitFest. I just sent out several pleas for help with specific patches 
where only a limited chunk of the community is really in a good position 
to test the code. Excluding those for the moment, here are some notes on 
the submitted patches that not only don't have a reviewer already, I 
don't know who might take them on yet:

"Command Triggers" by Dimitri Fontaine. This is a pretty big patch (3366 
lines) that won't be easy to slug through. I know Dimitri is still 
making some progress on this just by thinking more about the code, but 
that's not going to keep up for much longer without outside feedback.

"type privileges" by Peter Eisentraut. Another big patch (2537 lines), 
and it needs both some code review and searching for security issues in 
the new feature.

Both "type privledges" and "Command Triggers" are likely to slip well 
into December if we don't get a major reviewer working on them soon. I 
just don't know who that might be in either case.

"cursor calling with named parameters" by Yeb Havinga. This is a 
resubmission of a patch from the last CommitFest, so that previous 
feedback is around to give an idea what should be re-tested in the new 
version. At 702 lines including docs and tests, I'd call this a medium 
sized patch to work on. It'd say it's a decent candidate for someone who 
picked up a patch, got done early, and is enough of a glutton to want to 
do a second patch.

"splitting plpython into smaller parts" by Jan Urbański. Just asked for 
clarification on this one because I'm not quite sure what to do with it.

"unite recovery.conf and postgresql.conf": Since this has a lot more 
consensus issues than code ones, I'm going to look into this more 
myself. The things I submitted actually change around the potential ways 
to satisfy all the use cases here, too.

"Configuration include directory" and "includeifexists in configuration 
file" by me. These are not too hard to look at, and given that this all 
started as hacking on Mangus's code he may get to it eventually anyway. 
I'm not too worried about these two right now.

Hi Greg,

On Monday, November 28, 2011 11:13:52 AM Greg Smith wrote:
> With plenty of people in the US returning to work on Monday after a
> modest break last week, I wanted to put out an updated on the active
> CommitFest. I just sent out several pleas for help with specific patches
> where only a limited chunk of the community is really in a good position
> to test the code. Excluding those for the moment, here are some notes on
> the submitted patches that not only don't have a reviewer already, I
> don't know who might take them on yet:
> 
> "Command Triggers" by Dimitri Fontaine. This is a pretty big patch (3366
> lines) that won't be easy to slug through. I know Dimitri is still
> making some progress on this just by thinking more about the code, but
> that's not going to keep up for much longer without outside feedback.

> Both "type privledges" and "Command Triggers" are likely to slip well
> into December if we don't get a major reviewer working on them soon. I
> just don't know who that might be in either case.
Unless a major reviewer cries, I will - although not a major reviewer - take a 
stab thursday if thats fine with you.


Andres

On 11/28/2011 05:38 PM, Andres Freund wrote:
> Unless a major reviewer cries, I will - although not a major reviewer 
> - take a
> stab thursday if thats fine with you.

I intended to use that term to suggest the review job itself could 
become a major chunk of work, not to try and classify reviewers.  Bad 
wording choice.  If you could take a look at Command Triggers later this 
week I'm sure Dimitri would appreciate your feedback.

On Tuesday, November 29, 2011 02:45:50 AM Greg Smith wrote:
> On 11/28/2011 05:38 PM, Andres Freund wrote:
> > Unless a major reviewer cries, I will - although not a major reviewer
> > - take a
> > stab thursday if thats fine with you.
> 
> I intended to use that term to suggest the review job itself could
> become a major chunk of work, not to try and classify reviewers.  Bad
> wording choice.  If you could take a look at Command Triggers later this
> week I'm sure Dimitri would appreciate your feedback.
Well. I don't think its that bad wording... Imo the patch will soon need input 
from a commiter with experience in the area to solidify the direction...

Thanks,

Andres