Horton

Re: unused_oids script is broken with bsd sed

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

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Peter Eisentraut <peter.eisentraut@2ndquadrant.com>
Cc: Robert Haas <robertmhaas@gmail.com>, John Naylor <jcnaylor@gmail.com>, Stas Kelvich <s.kelvich@postgrespro.ru>, Andreas Karlsson <andreas@proxel.se>, Andrew Dunstan <andrew.dunstan@2ndquadrant.com>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>, David Fetter <david@fetter.org>
Date: 2018-05-04T02:42:41Z
Lists: pgsql-hackers 
Peter Eisentraut <peter.eisentraut@2ndquadrant.com> writes:
> On 5/3/18 15:37, Tom Lane wrote:
>> I took a quick look into this.  It's very easy to do so far as the Perl
>> code is concerned:

> I think in order to introduce warts like that, we have to have really
> great savings.  I haven't seen any actual analysis what the current
> problem is, other than one person expression a suspicion.

Well, the work's already done, and personally I think the code is
cleaner after 9bf28f96c and bad51a49a regardless of whether there's any
performance benefit.  You could call 1f1cd9b5d a wart if you wanted.
But we've done largely the same thing to pgindent, for one, in the past
year or so, and I find that to be a usability improvement, independently
of whether there's any build performance win.  My editor gets cranky
when files change under it for no reason.

			regards, tom lane

Commits

  1. Blindly try to fix MSVC build's use of genbki.pl and Gen_fmgrtab.pl.

  2. Avoid overwriting unchanged output files in genbki.pl and Gen_fmgrtab.pl.

  3. Rearrange makefile rules for running Gen_fmgrtab.pl.

  4. Avoid parsing catalog data twice during BKI file construction.

  5. Fix duplicate_oids and unused_oids so user needn't cd to catalog dir.

  6. Convert unused_oids and duplicate_oids to use Catalog.pm infrastructure.

  7. Make Catalog.pm's representation of toast and index decls more abstract.