Re: Jabber protocol porting

4ernov <[email protected]>
Newsgroups gmane.comp.kde.devel.kopete
Message-ID <CAL8gro3kHsxuC3HoZCUiPAXvMCw0nXMtODFzYLLCNwTKH8nydg@mail.gmail.com>
2016-06-07 10:26 GMT+03:00 Pali Rohár <[email protected]>:

> On Monday 06 June 2016 02:51:35 4ernov wrote:
> > Hello,
> >
> > to add some little bit of positive news, I wanted to inform, that I've
> > returned to Kopete porting and fix some issues in the Jabber protocol. I
> > pushed the code to my clone of Kopete repo on GitHub to not push
> > preliminary changes to main porting branch, so here's the code:
> > https://github.com/aclex/kopete/tree/kf5-jabber-iris.
>
> Great! That you very much for doing this.
>
> > Some time ago we discussed with Pali on some issues in the upstream XMPP
> > library, libiris, and ways of experimenting with it to possibly change
> it.
> > I tried to port the protocol with it at first, and it turned out, that
> the
> > actual change to upstream needed is quite trivial and extremely small:
> > https://github.com/psi-im/iris/pull/39.
>
> I just waned to tell you to use qobject_cast, but now I see that you
> already fixed it and upstream merged...
>

Yes, it looked like a trivial, but turned out to be a little bit tricky :)
But yes, the final fix, with Qt4 support preserved with ifdef, is already
merged.

> Hopefully, it would be merged and
> > we can synchronize it to have Qt5-compliant internal tree.
>
> Syncing is not a problem, time to time I'm doing it.
>

Yes, thank you for that. In the Jabber porting branch I've cherry-picked
probably the latest sync from master, so when time will come for another
sync I think I'll push those commits to 'kf5' branch of Kopete repository.

> Switch to QXmpp
> > still remains a sweet option, but it seems more wise to make all this
> > consequently and finish the port first, before changing something else.
>
> That switch would take couple of hours and I think libiris is usable for
> now. Once everything will be ported we can investigate how hard can be
> that switch...
>

Agreed on that.

> Apart from that, most of changes were kde-dev-scripts/kf5-driven, so to my
> > rough testing most of windows I managed to touch work as expected. SSL
> > connectivity doesn't work for some reason, as far, as I see it's
> something
> > with libiris/QCA2-qt5 interaction problems. Still, it works and the
> > functionality loss is less now.
>
> You can try to ask on qca mailing list, if there are some problems with
> switching applications from qt4 to qt5. Maybe somebody would know
> more...
>

I actually managed to fix the build against Qca-qt5, but yes, it seems to
not be used. Will try to sort it out.

> I think, there's a couple not ported protocols remaining in the tree, I'll
> > try to fix it, too. And overall application looks a bit unstable for now,
> > though I use it for everyday chatting for about 2-3 months already (I
> mean,
> > KF5-partially ported version).
>
> If you have code parts which are compatible with both qt4 and qt5 I can
> merge it into kopete master branch (after review). Basically master
> branch needs to be compiled with qt4 until Kopete will be fully ported
> to qt5.
>

Yes, sure, I see. I also think it's too early code-wise to push something
of that to master. Obviously, there're not so many parts which are
compatible both with KDE4/Qt4 and KF5/Qt5, due to changes in KF first of
all, and conditional compilation seems not quite optimal. I think, the way
you've mentioned is preferred: to port in parallel, test the port as
thoroughly as possible and then just switch master.


>
> --
> Pali Rohár
> [email protected]
> _______________________________________________
> kopete-devel mailing list
> [email protected]
> https://mail.kde.org/mailman/listinfo/kopete-devel
>

_______________________________________________
kopete-devel mailing list
[email protected]
https://mail.kde.org/mailman/listinfo/kopete-devel
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.