Merge 'trunk' into 'ado_mssql'

r7795@pentium (orig r7793):  ribasushi | 2009-10-16 08:28:35 -0400
Fix test to stop failing when DT-support is not present
r7801@pentium (orig r7799):  caelum | 2009-10-18 05:13:29 -0400
 r20728@hlagh (orig r7703):  ribasushi | 2009-09-20 18:51:16 -0400
 Another try at a clean sybase branch
 r20730@hlagh (orig r7705):  ribasushi | 2009-09-20 18:58:09 -0400
 Part one of the sybase work by Caelum (mostly reviewed)
 r20731@hlagh (orig r7706):  ribasushi | 2009-09-20 19:18:40 -0400
 main sybase branch ready
 r21051@hlagh (orig r7797):  caelum | 2009-10-18 04:57:43 -0400
  r20732@hlagh (orig r7707):  ribasushi | 2009-09-20 19:20:00 -0400
  Branch for bulk insert
  r20733@hlagh (orig r7708):  ribasushi | 2009-09-20 20:06:21 -0400
  All sybase bulk-insert code by Caelum
  r20750@hlagh (orig r7725):  caelum | 2009-09-24 02:47:39 -0400
  clean up set_identity stuff
  r20751@hlagh (orig r7726):  caelum | 2009-09-24 05:21:18 -0400
  minor cleanups, test update of blob to NULL
  r20752@hlagh (orig r7727):  caelum | 2009-09-24 08:45:04 -0400
  remove some duplicate code
  r20753@hlagh (orig r7728):  caelum | 2009-09-24 09:57:58 -0400
  fix insert with all defaults
  r20786@hlagh (orig r7732):  caelum | 2009-09-25 21:17:16 -0400
  some cleanups
  r20804@hlagh (orig r7736):  caelum | 2009-09-28 05:31:38 -0400
  minor changes
  r20805@hlagh (orig r7737):  caelum | 2009-09-28 06:25:48 -0400
  fix DT stuff
  r20809@hlagh (orig r7741):  caelum | 2009-09-28 22:25:55 -0400
  removed some dead code, added fix and test for _execute_array_empty
  r20811@hlagh (orig r7743):  caelum | 2009-09-29 13:36:20 -0400
  minor changes after review
  r20812@hlagh (orig r7744):  caelum | 2009-09-29 14:16:03 -0400
  do not clobber $rv from execute_array
  r20813@hlagh (orig r7745):  caelum | 2009-09-29 14:38:14 -0400
  make insert_bulk atomic
  r20815@hlagh (orig r7747):  caelum | 2009-09-29 20:35:26 -0400
  remove _exhaaust_statements
  r20816@hlagh (orig r7748):  caelum | 2009-09-29 21:48:38 -0400
  fix insert_bulk when not using bulk api inside a txn
  r20831@hlagh (orig r7749):  caelum | 2009-09-30 02:53:42 -0400
  added test for populate being atomic
  r20832@hlagh (orig r7750):  caelum | 2009-09-30 03:00:59 -0400
  factor out subclass-specific _execute_array callback
  r20833@hlagh (orig r7751):  caelum | 2009-10-01 11:59:30 -0400
  remove a piece of dead code
  r20840@hlagh (orig r7758):  caelum | 2009-10-03 15:46:56 -0400
  remove _pretty_print
  r20842@hlagh (orig r7760):  caelum | 2009-10-04 16:19:56 -0400
  minor optimization for insert_bulk
  r21050@hlagh (orig r7796):  caelum | 2009-10-18 04:56:54 -0400
  error checking related to literal SQL for insert_bulk

r7810@pentium (orig r7808):  caelum | 2009-10-20 21:10:39 -0400
add test for populate with literal sql mixed with binds, improve error messages
r7813@pentium (orig r7811):  ribasushi | 2009-10-21 10:33:45 -0400
Show what's wrong with the current populate code
r7814@pentium (orig r7812):  caelum | 2009-10-22 05:10:38 -0400
stringify values passed to populate/insert_bulk
r7815@pentium (orig r7813):  ribasushi | 2009-10-22 07:17:41 -0400
Some smoker run the suite for 30 *minutes* - the timeout seems to be too short for them (boggle)
r7816@pentium (orig r7814):  caelum | 2009-10-22 08:41:37 -0400
a few extra tests can never hurt, right? :)
r7817@pentium (orig r7815):  ribasushi | 2009-10-23 04:51:05 -0400
Prevent sqlt from failing silently
r7818@pentium (orig r7816):  ribasushi | 2009-10-23 04:52:49 -0400
{ is_foreign_key_constraint => 0, on_delete => undef } is a valid construct - no need to carp
r7822@pentium (orig r7820):  robkinyon | 2009-10-26 15:11:22 -0400
Fixed bad if-check in columns()
r7830@pentium (orig r7828):  caelum | 2009-10-31 09:01:56 -0400
change repository in meta to point to real svn url rather than svnweb
r7832@pentium (orig r7830):  caelum | 2009-10-31 16:04:39 -0400
pass sqlite_version to SQLT
r7833@pentium (orig r7831):  caelum | 2009-10-31 16:22:37 -0400
fix regex to numify sqlite_version
r7834@pentium (orig r7832):  caelum | 2009-10-31 18:59:19 -0400
work-around disconnect bug with DBD::Pg 2.15.1
r7845@pentium (orig r7843):  ribasushi | 2009-11-04 04:55:51 -0500
 r7817@Thesaurus (orig r7805):  rbuels | 2009-10-21 02:37:28 +0200
 making a branch, here we go again with the pg_unqualified_schema
 r7818@Thesaurus (orig r7806):  rbuels | 2009-10-21 02:38:59 +0200
 more pg unqualified schema tests, which expose a gap in the coverage
 r7819@Thesaurus (orig r7807):  rbuels | 2009-10-21 03:10:38 +0200
 gutted Pg storage driver's sequence discovery to just rely on DBD::Pg's last_insert_id.  this needs testing with older versions of DBD::Pg
 r7821@Thesaurus (orig r7809):  rbuels | 2009-10-21 04:00:39 +0200
 more coverage in Pg sequence-discovery tests.  i think this shows why last_insert_id cannot be used.
 r7822@Thesaurus (orig r7810):  rbuels | 2009-10-21 04:07:05 +0200
 reverted [7807], and just changed code to use the custom pg_catalog query, which is the only thing that works in the pathological case where DBIC is told a different primary key from the primary key that is set on the table in the DB ([7809] added testing for this)
 r7852@Thesaurus (orig r7840):  rbuels | 2009-11-03 18:47:05 +0100
 added Changes line mentioning tweak to Pg auto-inc fix
 r7854@Thesaurus (orig r7842):  ribasushi | 2009-11-04 10:55:35 +0100
 Cleanup exceptions

r7848@pentium (orig r7846):  caelum | 2009-11-06 10:01:30 -0500
transactions for MSSQL over DBD::Sybase
r7851@pentium (orig r7849):  caelum | 2009-11-10 07:16:18 -0500
made commit/rollback when disconnected an exception
r7852@pentium (orig r7850):  robkinyon | 2009-11-10 11:19:57 -0500
Added a note about select
r7853@pentium (orig r7851):  ribasushi | 2009-11-10 12:23:10 -0500
Changes
r7857@pentium (orig r7855):  frew | 2009-11-11 15:56:37 -0500
RT50874
r7858@pentium (orig r7856):  frew | 2009-11-11 17:50:43 -0500
RT50828
r7859@pentium (orig r7857):  frew | 2009-11-11 17:54:15 -0500
clearer test message
r7860@pentium (orig r7858):  frew | 2009-11-11 18:37:27 -0500
some cleanup for $rs->populate
r7862@pentium (orig r7860):  ribasushi | 2009-11-11 19:35:36 -0500
Fix find on resultset with custom result_class
r7863@pentium (orig r7861):  ribasushi | 2009-11-11 19:40:14 -0500
Fix return value of in_storage
r7864@pentium (orig r7862):  ribasushi | 2009-11-11 19:43:48 -0500
Extra FAQ entry
r7865@pentium (orig r7863):  ribasushi | 2009-11-11 20:11:25 -0500
Sanify _determine_driver handling in ::Storage::DBI
r7866@pentium (orig r7864):  ribasushi | 2009-11-11 20:14:37 -0500
Add mysql determine_driver test by Pedro Melo
r7871@pentium (orig r7869):  ribasushi | 2009-11-12 05:10:04 -0500
_cond_for_update_delete is hopelessly broken attempting to introspect SQLA1. Replace with a horrific but effective hack
r7872@pentium (orig r7870):  ribasushi | 2009-11-12 05:15:12 -0500
Clarifying comment
r7874@pentium (orig r7872):  ribasushi | 2009-11-12 18:13:40 -0500
The real fix for the non-introspectable condition bug, mst++
r7875@pentium (orig r7873):  ribasushi | 2009-11-12 18:24:56 -0500
Some cleanup
r7877@pentium (orig r7875):  frew | 2009-11-13 04:01:37 -0500
fix subtle bug with Sybase database type determination
r7882@pentium (orig r7880):  frew | 2009-11-13 18:53:29 -0500
release woo!