Re: script binaries renaming - Mailing list pgsql-patches

From Magnus Hagander
Subject Re: script binaries renaming
Date
Msg-id 468C7ACD.2090005@hagander.net
Whole thread Raw
In response to script binaries renaming  (Zdenek Kotala <Zdenek.Kotala@Sun.COM>)
Responses Re: script binaries renaming  (Peter Eisentraut <peter_e@gmx.net>)
Re: script binaries renaming  (Zdenek Kotala <Zdenek.Kotala@Sun.COM>)
List pgsql-patches
Zdenek Kotala wrote:
> I attach complete patch which renames following binaries
>
> createdb createlang createuser dropdb droplang dropuser clusterdb
> vacuumdb reindexdb
>
> to
>
> pg_createdb pg_createlang pg_createuser pg_dropdb pg_droplang
> pg_dropuser pg_clusterdb pg_vacuumdb pg_reindexdb
>
> Symlinks (or copy on win32) are created for backward compatibility.
>
> This renaming was discussed there:
>
> http://archives.postgresql.org/pgsql-hackers/2007-06/msg00145.php
>
> I create also separate unified patch for documentation.

A couple of comments:

Do we actually need to remove and recreate the documentation files?
Can't we keep the source files and just have it generate the man pages
and other references with the outputname pg_xyz?

Should we be installing symlinks or at least a placeholder manpage for
the old commandnames (maybe a single man page for all the old commands
saying "prefix pg_ to get this command") the same way it's done for the
binaries? (my apologies if it already does this, I don't know exactly
how the man pages are generated from the sgml. I just noticed there was
no makefile modified in the docs area)

I'm also not sure we really should install copies on win32. Given how
seldom these commands are actually used on windows (most people really
use pgadmin to do these things).

Also, your patch does not touch the msvc buildsystem, which will still
be generating files with the old name.

//Magnus


pgsql-patches by date:

Previous
From: Tom Lane
Date:
Subject: Re: SPI-header-files safe for C++-compiler
Next
From: Magnus Hagander
Date:
Subject: Re: Compile error with MSVC