Re: Large objects
"James Henstridge" <[email protected]>
| Newsgroups | gmane.comp.python.db.psycopg.devel |
|---|---|
| Message-ID | <[email protected]> |
On 05/05/2008, Federico Di Gregorio <fog-NGVKUo/i/[email protected]> wrote: > Il giorno lun, 05/05/2008 alle 22.59 +0800, James Henstridge ha scritto: > > > I had a look at the missing changes on the trunk branch, and have > > merged them to 2_0_x. > > > > There were a few recent changes by Fabio Tranchitella to the > > ZPsycopgDA code that were probably intended for 2_0_x, so I brought > > them across with "bzr replay". > > > > The remaining changes were the large object code. Merging them > > directly resulted in a lot of conflicts due to previous untracked > > merges in subversion. So I did an intermediate merge between trunk > > and the 2_0_x branch at the point of the previous merge, which managed > > to detect most cases of duplicate code. > > > Great work, thank once again. > > > > I think we can consider the old trunk branch dead now, so it might be > > worth renaming things. Perhaps the following: > > > > trunk => lobject-support > > branches/2_0_x => trunk > > > > A symlink from the old 2_0_x location would keep old checkouts working. > > > Ok. The lobject support was quite stable so we can just do the move you > suggested and keep 2.x in trunk. We can always branch a 2.0.6.1 for > short term fixes and release trunk as 2.1 if necessary (i.e., if we add > a lot of unstable features to trunk). Okay. I've moved the branches around now so http://initd.org/bazaar/psycopg/psycopg2/trunk/ contains the most up to date code base. It is trivial to branch off from a particular revision using "bzr branch -r N" if we want to continue work from an older revision. I agree that the lobject code is pretty stable. The test coverage could do with some expansion though (I only covered the basics -- there are a number of error paths that should be covered). James.