Paul W. Frields wrote:
I think this makes sense, but I would leave this up to Stuart. The
nice
thing about a branch is that it's the natural way to get the "good
stuff" folded back into HEAD. I think, though, that for this strategy
to function well, our webtest host needs to be able to build other
branches. As I recall, this came up at last night's FDSCo meeting. Did
anyone have any thoughts on how to accomplish this? The "bulkbuild"
script in our docs-common/bin/ is the starting point, so if someone
wants to submit a patch to the list, it would be appreciated.
The attached patch implements the "-r" option that bulkbuild already possessed
but was not yet implemented in the script. This is not really a final solution
since, with this patch, it builds only the modules that have the specified tag
(e.g., if, say, only the "documentation-guide" and
"fedora-install-guide"
modules were tagged with "FC5", then specifying "-r FC5" when invoking
bulkbuild
would result in successful builds for those two modules; the remaining modules
would fail when pulling from CVS). But, hopefully, this is a step in the right
direction.
Perhaps, ideally, bulkbuild should always result in the build of HEAD and, if
any additional tags for certain modules, those branches should be built as well.
One way (and I'm sure there are more) would be to have another commmand-line
option that specifies the module/branch builds that also need to be performed.
Something like (using -b for "branch build"):
$ bin/bulkbuild -n -b (fedora-install-guide=scott) <-- snip other args -->
In a to-be-modified bulkbuild, this would build HEAD for all the modules and the
"scott" branch of the fedora-install-guide module.
This same/similar info could be read in via a data file as well.
Take care,
Chris