Hi Justin,
Justin Erenkrantz wrote:
> --On Friday, March 12, 2004 11:23 AM -0600 kfogel@collab.net wrote:
>
>> I'm wondering, is there some way we can make it easier to catch such
>> problems earlier? If it would help, we could make it part of the
>> release process to send out a reminder to try building the bindings on
>> the soon-to-be-released branch. If we did that the moment we set the
>> date for a release, then bindings fixes could come in much sooner.
>
>
> Sure. Allow everyone to build it and provide regression tests for the
> bindings. This'll allow the RM to know that their bindings aren't
> duds (i.e. as part of the 'make check' process).
>
> dlr and I are trying to get #1 solved first (get the Java bindings to
> build at all), then we can work on #2. But, this is a slow process
> as, well, we're doing this in the limited amount of time we have.
> (We're trying to schedule a time when the both of us can coordinate in
> real-time on this.)
>
> We have a plan, and we're getting there. Note that the patch you
> committed is rendered moot with the integrated javahl build system.
> Problems like that go away. My plan is that we can get the build
> systems resolved by the time we go to 1.1. Full-on regression testing
> will probably take until 1.2. -- justin
I have created classes for doing regression tests for javahl in the
trunk. They are not complete yet, but they have the scope like the
basics tests of the command line client. Could you help me building a
javahlcheck target in the Makefile ? This would help for 1.1 and later
branches.
Unfortunately this classes can not be backported to 1.0, because they
need an extended interface.
Regards,
Patrick
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Mar 14 12:59:43 2004