On Wed, Jun 20, 2012 at 6:50 PM, Andrew Stitcher <[email protected]> wrote:
> On Wed, 2012-06-20 at 15:34 -0400, Darryl L. Pierce wrote:
>> On Wed, Jun 20, 2012 at 03:24:52PM -0400, Andrew Stitcher wrote:
>> > On Tue, 2012-06-19 at 16:05 -0400, Darryl L. Pierce wrote:
>> > > On Tue, Jun 19, 2012 at 02:11:51PM -0400, Andrew Stitcher wrote:
>> > > > > The big benefit to this would be breaking the Cmake dependencies 
>> > > > > between
>> > > > > the bindings and the cpp build tree. We could build them 
>> > > > > independently,
>> > > > > which is a Good Thing (tm).
>> > > >
>> > > > I don't think you are really breaking any dependencies by moving code
>> > > > around are you? The bindings will still depend on the c++ code where
>> > > > ever it lives in the tree.
>> > >
>> > > What I was thinking was that, by moving the bindings out and then
>> > > versioning the SWIG wrapper code we could build the individual language
>> > > bindings separately rather than having to build all of Qpid to get them.
>> > > I'm still not totally comfortable with the idea of versioning those
>> > > wrappers.
>> >
>> > Could you explain what you mean by "versioning" in this context?
>>
>> Generating a copy of the SWIG wrapper for that language and then commit
>> it in git.
>
> I don't think committing generated code to the repository is a good idea
> (but I could be convinced). Why do want to do this?

For the Java binding, the generated code (compiled into a jar file) is
checked into the java lib dir.
I see no way around it, unless you make the java build (for the cpp
based api implementation) conditional on the cpp build, which is not
going to fly.

Rajith

> Andrew
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to