roundhouse project updates - Google Code

issue 63 (Trailing Comments In Oracle SQL File Cause Error) commented on by trueblu...@gmail.com

2012-02-19T15:47:43Z

I think I may have this one fixed, but I will need to verify.

issue 43 (Enhancement - Command line option to prevent creation of Pro...) changed by trueblu...@gmail.com

2012-02-19T15:46:08Z

Status: Fixed

Labels: Milestone-Release0.8.6

issue 71 (Roundhouse has trouble parsing scripts) Labels changed by trueblu...@gmail.com

2012-02-19T15:44:49Z

Labels: Milestone-Release0.8.6

issue 69 (Roundhouse silently changes the recovery mode to FULL every ...) Labels changed by trueblu...@gmail.com

2012-02-19T15:44:43Z

Labels: Milestone-Release0.8.6

issue 68 (Using ScriptFileVersioner throws exception) Labels changed by trueblu...@gmail.com

2012-02-19T15:44:26Z

Labels: Milestone-Release0.8.6

r381 (adding mpareja's text version resolver full path enhancement...) committed by ferventc...@gmail.com

2012-02-19T15:38:30Z

adding mpareja's text version resolver full path enhancement
(https://github.com/chucknorris/roundhouse/pull/55)

r380 (Adding text version resolver - from Mario Pareja (https://gi...) committed by ferventc...@gmail.com

2012-02-16T05:02:40Z

Adding text version resolver - from Mario Pareja (https://github.com/chucknorris/roundhouse/pull/50)

r379 (bottles updates) committed by ferventc...@gmail.com

2012-02-16T04:57:43Z

bottles updates

r378 (bottles updates) committed by ferventc...@gmail.com

2012-02-16T04:56:27Z

bottles updates

issue 71 (Roundhouse has trouble parsing scripts) Status changed by trueblu...@gmail.com

2012-02-16T04:45:39Z

This is fixed in revision 377

Status: Fixed

r377 (Fix for issue 71 - roundhouse has issues in splitting script...) committed by ferventc...@gmail.com

2012-02-16T04:44:57Z

Fix for issue 71 - roundhouse has issues in splitting scripts where -- is not at the beginning of a
line.

issue 41 (Enhancement - BASELINE - RoundhousE needs a "catch up" mode ...) Status changed by trueblu...@gmail.com

2012-02-08T15:15:03Z

Well that seems a pretty simple fix.

Status: Accepted

issue 41 (Enhancement - BASELINE - RoundhousE needs a "catch up" mode ...) commented on by Roman.Le...@gmail.com

2012-02-08T14:19:28Z

Hi all,
We could not continue waiting for this feature and I've made my small implementation of it (one updated source file is attached). It has no unit tests, sorry for this, but I am not sure it will be used by the developers anyway.

issue 71 (Roundhouse has trouble parsing scripts) commented on by trueblu...@gmail.com

2012-01-27T17:07:00Z

Fixed in https://github.com/chucknorris/roundhouse/commit/a297dd4089a2741ba63b334297eef524b6ed75fc

Will get it back to the SVN repo this weekend.

issue 71 (Roundhouse has trouble parsing scripts) changed by trueblu...@gmail.com

2012-01-27T17:04:23Z

Thanks. I believe I have a fix for you.

Status: Accepted

Labels: Priority-Critical −Priority-Medium

issue 71 (Roundhouse has trouble parsing scripts) reported by matt.hor...@gmail.com

2012-01-27T00:22:15Z

What steps will reproduce the problem?
1. Put this in a script:

INSERT [dbo].[Foo] ([Bar]) VALUES (N'hello--world.
Thanks!')
INSERT [dbo].[Foo] ([Bar]) VALUES (N'Go speed racer, go speed racer, go speed racer go!!!!! ')

2. Have roundhouse execute the script.

What is the expected output? What do you see instead?
The expected output is that the records are inserted correctly. I spent the day stepping through the roundhouse source and determined that the regex used for parsing out sql statements isnt doing the job properly.

Inside of the SqlServerDatabase class (in roundhouse.databases.sqlserver), there is a regex string defined and exposed by the sql_statement_separator_regex_pattern.

This is used in the roundhouse.sqlsplitters.StatementSplitter.split_sql_on_regex_and_remove_empty_statements() method.

After the sql_statement_scrubbed variable is filled in that method (line 18 for me), the result is:

INSERT [dbo].[Foo] ([Bar]) VALUES (N'hello--world.
Thanks!')
INSERT [dbo].[Foo] ([Bar]) VALUES (N'Go speed racer, |{[_REMOVE_]}| speed racer, |{[_REMOVE_]}| speed racer go!!!!! ')

As you can see, some of the text of the record is improperly identified. This causes the sql statement to be split apart improperly, and everything gets discombobulated. Then, when each statement is added to the return_sql_list, you get a bunch of broken sql. For the above example, the list ends up containing:

[0]:
INSERT [dbo].[Foo] ([Bar]) VALUES (N'hello--world.
Thanks!')
INSERT [dbo].[Foo] ([Bar]) VALUES (N'Go speed racer,

[1]:
speed racer,

[2]:
speed racer go!!!!! ')

What version of the product are you using? On what operating system?
HEAD revision on xp sp3.

Please provide any additional information below.
In the case above, the problem seems to be caused by the multi-line INSERT statement in combination with the string '--'. There are likely other problematic parsing scenarios here, since when i tried to execute a big data dump using a sql script generated by management studio, I had many such problems.

r376 (bottles, beaches... enabling package restore hard core...) committed by ferventc...@gmail.com

2012-01-24T23:03:28Z

bottles, beaches... enabling package restore hard core...

r375 (Allow disabling of script output (backups, scripts), log fil...) committed by ferventc...@gmail.com

2012-01-19T13:04:09Z

Allow disabling of script output (backups, scripts), log files are retained

r374 (Use Convert.ToBoolean to parse returned scalar from database...) committed by ferventc...@gmail.com

2012-01-19T13:03:23Z

Use Convert.ToBoolean to parse returned scalar from database when expecting boolean value

issue 43 (Enhancement - Command line option to prevent creation of Pro...) commented on by trueblu...@gmail.com

2012-01-18T14:32:00Z

https://github.com/chucknorris/roundhouse/pull/47

DisableOutput it is. :D