Andres Freund <andres@anarazel.de> writes:
> On 2015-08-07 20:16:20 -0400, Noah Misch wrote:
>> On another note, I'm perplexed by the high speed commits from this thread.
>> Commit de6fd1c landed just 191 minutes after you posted the first version of
>> its patch. Now lockdefs.h is committed, right in the middle of discussing it.
> Hm. We'd essentially decided what we're going to do about the inline
> stuff weeks ago, so I don't feel particularly bad pushing it quickly. A
> lot of platform dependent stuff like this is going to have some
> iterations to deal with compilers you don't have access to. The only
> reason I committed the lockdefs.h split relatively quickly is that I
> wanted to get the buildfarm green to see wether there are other problems
> hiding behind the linker error. Which does, so far, not appear to be the
> case.
FWIW, I agree with that: leaving buildfarm members red for any sustained
amount of time is a bad practice. Code cleanliness is something we can
argue about at leisure, but if you have critters that aren't building
then you don't know what might be hiding behind that. We've had bad
experiences in the past with leaving that sort of thing unfixed.
regards, tom lane
