From: Robert Buels <rmb32@cornell.edu>
Date: Fri, 4 Sep 2009 17:43:17 +0000 (+0000)
Subject: added big block comment explaining Pg sequence discovery strategy
X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=commitdiff_plain;h=46ae368dd29448a47593c7e3d3ec5c18f5c7cc79;p=dbsrgits%2FDBIx-Class-Historic.git

added big block comment explaining Pg sequence discovery strategy
---

diff --git a/lib/DBIx/Class/Storage/DBI/Pg.pm b/lib/DBIx/Class/Storage/DBI/Pg.pm
index 61a8d5e..d13f48a 100644
--- a/lib/DBIx/Class/Storage/DBI/Pg.pm
+++ b/lib/DBIx/Class/Storage/DBI/Pg.pm
@@ -54,6 +54,22 @@ sub _dbh_get_autoinc_seq {
     ( $schema, $table ) = ( $1, $2 );
   }
 
+  # Build and execute a query into the pg_catalog to find the Pg
+  # expression for the default value for this column in this table.
+  # If the table name is schema-qualified, query using that specific
+  # schema name.
+
+  # Otherwise, find the table in the standard Postgres way, using the
+  # search path.  This is done with the pg_catalog.pg_table_is_visible
+  # function, which returns true if a given table is 'visible',
+  # meaning the first table of that name to be found in the search
+  # path.
+
+  # I *think* we can be assured that this query will always find the
+  # correct column according to standard Postgres semantics.
+  #
+  # -- rbuels
+
   my $sqlmaker = $self->sql_maker;
   local $sqlmaker->{bindtype} = 'normal';