Hi Thomas,
What you are proposing sounds good to me and all the better if you have the cycles available to do it.
It would certainly be good to get Henning's take on this also as I am unsure of his thoughts on the work Martin started in HEAD and of the impact of the migration to Subversion.
Scott
-- Scott Eade Backstage Technologies Pty. Ltd. http://www.backstagetech.com.au
Thomas Fischer wrote:
Hi,
as most of you know, we have at the moment a very unpleasant situation in CVS: There are two branches, HEAD and TORQUE_3_1_BRANCH, both of which contain improvements which are not contained in the other branch. Also, at the moment, TORQUE_3_1_BRANCH is the active development branch, which often causes confusion as patches are suppliead vs HEAD and not vs TORQUE_3_1_BRANCH.
In my opinion, this situation should be dealt with in the near future, as each commit to any of the branches makes it more difficult to resolve the issue.
I have listed the differences between HEAD and TORQUE_3_1_BRANCH in two recent mails to the dev list (see http://nagoya.apache.org/eyebrowse/[EMAIL PROTECTED]&msgNo=4591 for the runtime, and my last mail to the dev list for the generator)
In short, the result are as follows: The runtime and the generator should be considered separately: In the runtime, the TORQUE_3_1_BRANCH is more advanced than HEAD. In TORQUE_3_1_BRANCH, we have partial support of schemata and support for outer joins, which we do not have in HEAD.
In the generator, HEAD is more advanced than the TORQUE_3_1_BRANCH. HEAD contains a major refactoring of the object model, including the replacement of the db.props files by java classes (this mean e.g. that the properties of the database can better be queried at runtime). TORQUE_3_1_BRANCH, in contrast, contains Henning's correctGetters improvement. However, it seems to me that the correctGetters improvement can quite easily be mergead into HEAD, whereas it is much more problematic to move the refactoring of the object model to TORQUE_3_1_BRANCH
So I would propose the following: Take the runtime from TORQUE_3_1_BRANCH and the generator from HEAD, and add the correctGetters to the generator in HEAD, and put this all into the HEAD branch . (Of course, there are some files that need special treatment, such as the postgreSQL templates in the generator, so this should only be seen as the general direction.)
Are there any opinions about this ? I would really appreciate some feedback, especially of people which have recently used the CVS HEAD in their projects.
If the above proposal is accepted, I would volunteer to do most of the work (though I will certainly need some help in some places, e.g. with postgreSQL questions)
Thomas
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
