Re: Errands around AllocateDir()

Tom Lane <tgl@sss.pgh.pa.us>

From: Tom Lane <tgl@sss.pgh.pa.us>
To: Michael Paquier <michael.paquier@gmail.com>
Cc: PostgreSQL mailing lists <pgsql-hackers@postgresql.org>
Date: 2017-12-04T17:05:26Z
Lists: pgsql-hackers
Michael Paquier <michael.paquier@gmail.com> writes:
> I had a close look at all the callers of AllocateDir() and noticed a
> couple of unwelcome things (Tom noticed some of those in the thread
> mentioned above, I found others):

The only part of this that seems likely to be controversial is the
decision to get rid of special-case error messages, eg replace
	"could not open tablespace directory \"%s\": %m"
with the more generic
	"could not open directory \"%s\": %m"

As I said in the previous thread, I don't see anything much wrong
with that; but if anyone doesn't like it, speak now.

Also, I'm inclined to back-patch the exporting of ReadDirExtended,
so that we can rely on it being available if we use it in any
back-patched fixes.  I wouldn't back-patch anything else here that's
not a clear bug fix, but we may need public ReadDirExtended anyway
just to do that much (I didn't really check that yet).

			regards, tom lane


Commits

  1. Clean up assorted messiness around AllocateDir() usage.