Re: [Jack-Devel] lib{j, tsch}ack{, 1, 2} [Was: Jack "capsule/container" implementation?]
On Thu, Feb 10, 2011 at 09:27:30AM -0600, Gabriel M. Beddingfield wrote:
>
>
> On Thu, 10 Feb 2011, Paul Davis wrote:
>
> >>OK. Now I get http://pastebin.com/bBG3LQcW
> >>Which has about 1300 lines differing.[1]
> >
> >You've established that these files are different. I think we accept
> >that it would be better if there was a single set of header files used
> >by all implementations of JACK. I'm not sure the best way to approach
> >this goal.
>
> Sorry, I didn't mean that to be argumentative... rather it confirmed
> the assertions that "it's not /that/ bad."
most of that diff is jack2 lacking JACK_OPTIONAL_WEAK_EXPORT stuff.
this stuff already exists in svn.
please dont do this with a year old versions of the codebase.
here is a git diff -w of that set with pretty recent versions:
http://pastebin.com/kcNRrgSc
>
> >>Anyway, the point (which I see has been taken) is that I don't foresee
> >>the libjack1 and libjack2 thing going away as long as Jack1 and Jack2
> >>each provide libjack.
> >
> >I don't agree that this is clear at all. I can build jack1 and jack2
> >on my system. I can install one, run a JACK app, quit, uninstall that
> >version, install the other, and run the same JACK app. Why would
> >anyone sanely call the two libraries libjack1 and libjack2 when there
> >are (and have long been) shared object versioning systems for *nix
>
> ...except that, as soon as they switch to Jack2 Debian had a JACK
> ABI breakage with a few apps.
it was a bug. but this has been discussed to death already.
i am working on a patch, which stops libjack1 from leaking internal
symbols. maybe this nitpicking will stop then.
it would still be nice, if we could get the headers into a deeper sync.
i dont like the idea of a separate repository for the headerfiles
though.
does somebody have a better idea ?
maybe a script which checks the headers, and mails diffs to the ML or
somthing ?
--
torben Hohn
1297436897.23828_0.ltw:2,a <20110211150801.GM5550 at siel dot b>