Merge 'roundtrip' into 'trunk'

r1509@Thesaurus (orig r1508):  ribasushi | 2009-04-27 17:38:24 +0200
Branch off to work on proper parser/producer roundtrips
r1512@Thesaurus (orig r1511):  ribasushi | 2009-04-28 09:34:49 +0200
Somewhat working global roundtrip test
r1513@Thesaurus (orig r1512):  ribasushi | 2009-04-28 09:36:46 +0200
Add an oversized varchar to schema.xml
r1514@Thesaurus (orig r1513):  ribasushi | 2009-04-28 09:43:29 +0200
Fix mysql roundtrip glitch
r1515@Thesaurus (orig r1514):  ribasushi | 2009-04-28 09:50:34 +0200
Test seems to be finioshed - massive bowl of FAIL
r1516@Thesaurus (orig r1515):  ribasushi | 2009-04-28 10:06:28 +0200
Bail out from failing tests early
r1517@Thesaurus (orig r1516):  ribasushi | 2009-04-28 10:14:59 +0200
Translator/schema dependency test
r1518@Thesaurus (orig r1517):  ribasushi | 2009-04-28 10:33:06 +0200
Todoify test - probably not in this lifetime
r1519@Thesaurus (orig r1518):  ribasushi | 2009-04-29 00:59:41 +0200
Changes to tests to go along r1512
r1527@Thesaurus (orig r1526):  ribasushi | 2009-05-03 04:46:58 +0200
Test XML roundtrip as well (also fail)
r1528@Thesaurus (orig r1527):  ribasushi | 2009-05-03 04:47:30 +0200
re-fix tests after merge
r1531@Thesaurus (orig r1530):  ribasushi | 2009-05-04 08:37:06 +0200
Rewind exhausted globs before attempting a read
(on a side note all this IO::Scalar business is just silly, not touching because a rewrite is imminent anyway)
r1532@Thesaurus (orig r1531):  ribasushi | 2009-05-04 09:10:44 +0200
Do not add xml comment header if no_comments is set
r1533@Thesaurus (orig r1532):  ribasushi | 2009-05-04 09:16:00 +0200
Better debug output
r1534@Thesaurus (orig r1533):  ribasushi | 2009-05-04 09:17:36 +0200
table/field counts are held per-object, not globally
Tests adjusted (no idea why they were initially written this way...)
r1535@Thesaurus (orig r1534):  ribasushi | 2009-05-04 09:30:12 +0200
Concentrate on testing the 'big 3' for now
r1536@Thesaurus (orig r1535):  ribasushi | 2009-05-04 10:13:51 +0200
Variable table and column names? wtf?!
r1537@Thesaurus (orig r1536):  ribasushi | 2009-05-04 10:16:19 +0200
SQLite improvements:
- Support parsing of all DROP clauses
- Support parsing of field-level comments
- When producing do not append table names to constraint/index names - it's useless
r1538@Thesaurus (orig r1537):  ribasushi | 2009-05-04 14:49:16 +0200
Adjust insane tests to pass (the expected returns at times are mind-boggling)
r1539@Thesaurus (orig r1538):  ribasushi | 2009-05-05 09:06:33 +0200
Fix a couple of mistakes in the PG parser
r1541@Thesaurus (orig r1540):  ribasushi | 2009-05-16 08:14:58 +0200
Add disabled YAML roundtrip test
r1551@Thesaurus (orig r1550):  ribasushi | 2009-06-07 10:11:01 +0200
VIEW support for Pg parser, also some cleanups
r1552@Thesaurus (orig r1551):  ribasushi | 2009-06-07 10:18:16 +0200
Add parenthesis into the VIEW definition to make sure the pg parser still can deal with them
r1553@Thesaurus (orig r1552):  ribasushi | 2009-06-07 10:26:06 +0200
The way we generate create view statements is not standards compliant (per RhodiumToad in #postgresql)
r1554@Thesaurus (orig r1553):  ribasushi | 2009-06-07 12:12:08 +0200
Multiple fixes for the SQLServer producer/parser combo
- Rip out braindead parts of mk_name (just like sqlite)
- Disable view/procedure production - they never worked in the first place
- Improvements to the parser to be able to actually parse what the producer spits out (no functional changes)

r1556@Thesaurus (orig r1555):  ribasushi | 2009-06-07 12:19:54 +0200
Make perlpod happy