Re: [PHP-GTK-DEV] History, "Dead" and community

Not for Long:

We're gonna bring this back around, even if it kills us (I hope; this commun=
ity needs some revitalization badly). Bruno is testing some stuff, and Alexa=
ndre is working on an actual implementation. I hope to help both at two poin=
ts in time, particularly mid- to late-July. Thank you for your support, and I=
 appreciate those of you that stayed theough the darker times to help usher m=
e in.=20

--MGage--

> On Jun 22, 2015, at 12:35 PM, Elizabeth M Smith <[email protected]> w=
rote:
>=20
> Hi folks.
>=20
> The mailing list got REALLY busy all of the sudden.
>=20
> I've been having issues with php mailing lists (they REALLY don't like
> my email) and just realized all the stuff going on here.
>=20
> First of all it's great to have people interested in the project.  For
> a LONG time it's been 1 or 2 people... with a few popping up here or
> there to get interested and then moving on quickly (usually when they
> realize how much work needs to be done)
>=20
> And although there's been someone keeping track of the actual code
> itself - the website was completely abandoned (thanks for taking that
> monstrosity on) and the documentation has been similarly abandoned
> (it's in old docbook and uses xsl to generate) - the people who once
> knew how that worked are gone
>=20
> There have actually been a couple of attempts to start up a php-gtk3
> project before.  But it's a BIG undertaking and only having 1 or 2
> people work on it makes it difficult to get off the ground.
>=20
> There's also the fact that a fully introspected version of the
> extension (like pythons gobject) have some pretty deep underlying
> issues (mostly related to the typelibs and the fact that they are
> strongly linked in many aspects to the python gobject wrapping
> implementation) - which is why although there's a prototype of an
> introspected solution it doesn't really work well.
>=20
> If you're interested in some of the code you can find it
> https://github.com/gtkforphp/ - some of it was actually working once
> upon a time.
>=20
> I think it's a good idea to do a "new" gtk for php but I think there
> are some things to keep in mind.
>=20
> 1. We do have an irc channel on freenode #php-gtk - this is a great
> way to connect with the people who still work on the project and know
> the code - not  only the php-gtk code and php internals code, but the
> gnome libraries as well.  Especially when talking technical details
> it's nice to do it in realtime.
> 2. The project has never been completely dead - it just basically only
> had one dev  (who is poor at recruiting) for a long time - which is
> never a good thing
> 3. The php-gtk source code is on php.net and mirrored on github -
> there have been some prs but not a lot - and none from the people now
> currently conversing on the list -  and it's also limited to the
> problem of basically 1 developer to look at and pull them
> 4. Discussions on the future should be all on list - not in private emails=
 :)
>=20
> Some thoughts on a complete rewrite
>=20
> 1. a monolithic extension is not very useful - php-gtk should become
> several small extensions
>=20
> right now php-gtk has wrappings for gtk, gdk, some glib, atk, some
> cairo integration (although most of cairo was pulled into it's own
> extension)  and many others
>=20
> parts of glib, cairo, and other gnome libraries such as clutter make
> sense to use not only in a gui environment, but even in a server
> environment
>=20
> This would even allow for things like a gtk2 and gtk3 extension
>=20
> the extension(s) should probably mirror the underlying libraries
>=20
> glib, gobject, gio, cairo(exists, is getting a php7 fix), pango (also
> exists), atk, gtk2, gtk3, gdk, clutter, gstreamer, etc.
>=20
> 2. If you're rewriting from scratch, a PHP7 only version is probably a goo=
d idea
>=20
> The internal apis of php have changed an incredible amount from 5 -> 7
> and it would be both very ugly and very hard to support both PHP7 and
> PHP 5.6(and higher) in the same code base
>=20
> 3. PHP-gtk has very very few tests... and needs them
>=20
> This is probably as important as the rewrite - although you really
> don't need to test the functionality of the gtk library itself, it IS
> important to make sure the bindings are tight.
>=20
> As far as getting php-gtk commit karma, that's really not that hard
> Submit some PR's - usually 2 or 3 good ones with tests
> Discuss things on list and/or in IRC
> As long as stuff doesn't break the build and works it gets pulled -
> but remember right now there's 1 dev... so expect lag time (maybe a
> week) until there are 2 or 3 or more devs capable of code reviewing,
> building, and testing - because you need those checks
>=20
> If you need help getting an environment for compiling set up on
> windows or linux - I'm happy to help.  Pop into irc.  If you're really
> ambitious someone could update the documentation :)
>=20
> Also php-gtk could really use someone to roll releases.  We actually
> should have had about 3-4 - there's been new functionality added and
> build fixes done and features integrated.. just no release.
>=20
> For example - there is complete gtkbuilder integration in the git
> version of php-gtk (that's the format glade3 outputs btw) - but none
> of that is documented beyond some demo files.
>=20
> Thanks,
> Elizabeth M Smith
>=20
> --=20
> PHP-GTK Development Mailing List (http://gtk.php.net/)
> To unsubscribe, visit: http://www.php.net/unsub.php
>=20