Babelers,
I finished the version -1 (pre version zero) of the server output code.
Gabe is now converting it from a stand alone tool I wrote (version -1)
to an integrated part of the server code base (version 0). It generates
all the plug-in fragments and fragments as per as per the wiki-described
output method 1.
Of course, the real test of the output routines is whether the
resulting update site (site.xml, features, plugins, etc) can be
consumed by Eclipse and actually causes Eclipse to change languages.
However, that test will have to wait until we have a realistic set of
translations in the database.
Beside the real life test, the other to do items left on the output
side are:
- M1 (post milestone 0) - add copyright headers to the properties
files, plugins, and features based on the authors who contributed the
translations. Gabe assures me that this data (who) is in the database,
so this should be easy.
- M1 - add mirrorsURL and update site urls to the site.xml and
feature.xml files. I need to know where the download files will live
and then these will be easy.
- M2 - produce multiple versions of plug-ins when multiple versions
exist in the database (e.g., JDT 3.1, 3.2, 3.3, 3.4, ...). First we'll
need to have multiple versions in the database so that I can see what
the importer builds, then I'll need to modify a few of the queries and
put in a fifth loop (the four loops are: for each language, for each
plugin, for each properties file, for each translated string).
- M3 - if we decide to, implement output methods 2 and 3 as
re-packaging (zips of different subsets of the generated update site
plugins/ directory)
As soon as we get M0 done, I'll enter the rest of these to-dos as bugs
and we can track them there. Kit has (probably) already started to add
bugs/enhancements such as the ability to mark strings as "don't
translate" and things like that.
- Bjorn
--
[end of message]
|