Don't need relhasindexes, relhasrules, or relhastriggers since older DBs have been dropped. This also makes up 8.4 compatible without additional changes required.
Minor fixes for named functions for newer DBD::Pg's. A long time ago DBD::Pg changed from returning an array string to returning an actual array for PostgreSQL arrays. Autodoc did not take this into account.
rbt [Wed, 12 Mar 2008 18:50:19 +0000 (18:50 +0000)]
Leonardo Cezar patches:
One of them to fix a odd behavior when we had views in schema to
schema to extract some reporter.
Previously we got an error on trying to get statistcs from a view,
since views hasn't stats.
The other one to choose which table(s) we want to collect information
with --table=foo,bar,x ... option.
Although it's functional we still get warning when dealing with table
FKs that weren't selected together; among others gotchas ;-).
Unfortunately I didn't find any mailing lists or forum to treat this
issue neither at pgfoundry nor autodoc's website.
--
FKeys still unfixed.
Statistics portion cleaned up by me.
Virtually no testing of input for comma separated values but oh well.
rbt [Wed, 12 Mar 2008 18:21:52 +0000 (18:21 +0000)]
New versions of DBD::Pg like to return an array reference instead of the special text string that PostgreSQL returns.
This might take care of the issue, although only partially tested (FreeBSD ports doesn't have the newer version yet as it breaks a number of applications).
rbt [Tue, 16 May 2006 19:01:27 +0000 (19:01 +0000)]
Assume the strings are in UTF8 instead of in ASCII. This expands on the number of characters that we can work with. While not perfect, it is an improvement.
rbt [Tue, 3 May 2005 02:07:05 +0000 (02:07 +0000)]
Eliminate Dia warnings for newer versions of dia (new elements).
Luckily, Dia will completely ignore these saved values and calculate it's own instead which is convenient because we really have no idea where to place things.
rbt [Sun, 27 Feb 2005 18:08:06 +0000 (18:08 +0000)]
Include argument names as provided by Jacob Smullyan.
Clean up handling of argument types from a schema not in the current path. This means getting rid of the schema prefix added by format_type so when can add our own when appropriate.
rbt [Tue, 29 Jun 2004 16:40:40 +0000 (16:40 +0000)]
From: Bort, Paul
Subject: RE: Patch for Dia templates for Postgresql_AutoDoc
Date: Tue, 22 Jun 2004 15:23:24 -0400
I've used the same method of shortening as the constraint you mentioned, but
I made it a new entry (table_comment_dia) in the hash rather than mangle the
existing one. I tried to follow the existing naming conventions and coding
style. I picked a total of 45 characters because that seems to be about the
right size on my screen. In testing I also found a bug in Dia that causes a
crash if a comment is too long, but that seems to be in the 100+ character
range, so I'm not worried about 45.
Here are the patches for postgresql_autodoc.pl, dia.tmpl, and
zigzag.dia.tmpl.