Re: pgsql: Fix for early log messages during postmaster startup getting lost

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Magnus Hagander <magnus(at)hagander(dot)net>, Neil Conway <neilc(at)samurai(dot)com>
Cc: pgsql-committers(at)postgresql(dot)org
Subject: Re: pgsql: Fix for early log messages during postmaster startup getting lost
Date: 2007-02-12 04:47:01
Message-ID: 22326.1171255621@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-committers

Neil Conway <neilc(at)samurai(dot)com> writes:
> FYI, it is considered good practise to commit a patch at approximately
> (or exactly) the same time on all branches, so tools like cvs2cl will be
> more likely to collect the changes together.

It's also helpful to use exactly the same text for the log messages in
all the branches. Again, this is so that cvs2cl understands they're
the same patch. I tend not to bother with "Backpatch to xxx" comments,
as the CVS log makes that perfectly clear anyway; but if you use them,
they should be the same for all branches committed to.

> Comments like that are fragile (elog.c could change, for example), and
> basically content-free anyway, IMHO. If you need to make SysLoggerPID
> part of postmaster.c's external API, why not just do that, remove the
> comment, and add the extern declaration to postmaster.h?

The counterargument is that he avoided exporting SysLoggerPID to
anything except elog.c. If it's in postmaster.h then who knows what
will start depending on it? But I see Neil's point too; this coding
is definitely a bit fragile.

regards, tom lane

In response to

Responses

Browse pgsql-committers by date

  From Date Subject
Next Message Magnus Hagander 2007-02-12 08:32:20 Re: pgsql: Fix for early log messages during postmaster startup getting lost
Previous Message Neil Conway 2007-02-12 03:45:47 Re: pgsql: Fix for early log messages during postmaster startup getting lost