On Wed, Sep 7, 2011 at 7:12 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Bruce Momjian <bruce@momjian.us> writes:
>> Robert Haas wrote:
>>> I was less concerned about the breakage that might be caused by
>>> variables acquiring unintended referents - which should be unlikely
>>> anyway given reasonable variable naming conventions - and more
>>> concerned that the associated refactoring would break recovery. We
>>> have no recovery regression tests; that's not a good thing.
>
>> So we are talking about more than moving files between functions? Yes,
>> it would be risky to restruction functions, but for someone who
>> understand that code, it might be safe.
>
> The pgrminclude-induced bug you just fixed shows a concrete way in which
> moving code from one file to another might silently break it, ie, it
> still compiles despite lack of definition of some symbol it's intended
> to see.
>
> Having said that, I tend to agree that xlog.c is getting so large and
> messy that it needs to be broken up. But I'm not in favor of breaking
> up files just because they're large, eg, ruleutils.c is not in need of
> such treatment. The problem with xlog.c is that it seems to be dealing
> with many more considerations than it originally did.
I agree as well, though we've spawned many new files and directories
in the last 7 years. Almost nothing has gone in there that didn't need
to.
As long as we accept its not a priority, I'll do some work to slide
things away and we can do it over time.
Please lets not waste effort on refactoring efforts in mid dev cycle.
Having this done by someone without good experience is just going to
waste all of our time and sneak bugs into something that does actually
work rather well.
--
Simon Riggs http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services