Re: EEL vs. GAL



As I said before, we have only begun to do the work of splitting this
code out of Nautilus.  That continues to involve a lot of careful
unwinding of dependencies and other cruft that tends to creep into
projects.

We haven't even started doing the work of making some of our projects
use this new library, in particular Reef.  That means we really don't
know very precisely how its going to be used by that project.  We have
an idea but only doing the work will bring out the issues.

I've only just started to remove the Eazel services stuff from Nautilus
(because of popular demand!) so I cant say im 100% clear on what the
requirements are yet.  We'll have better idea in the neat future.

You are right that the location and contents of the library are
currently for the convenience of hackers working on Eazel projects. 
However, we are carefully taking pieces of code apart to end up with
something that might be useful outside of Eazel projects.

Unfortunately it is not clear at all what would be useful by third
parties or not.  There are at least 20 good flame fests about some of
the content of Eel as it currently stands.  Some of the stuff is not
controversial at all, of course.  Some of the stuff, like the smooth
widget voodoo, is almost certainly not meant to be long term stuff,
since there is much better stuff on the horizon (xrender)

So for now, I don't know enough yet about the details of how such a
combination would occur.  Of course, in principle, I completely agree
with you about the reasons why combining the libraries or parts of them.

-re

Matthias Warkus wrote:
> 
> Hi,
> 
> I notice that there are now two libraries with the same purpose in
> GNOME CVS: EEL and GAL. Both are free and open, but essentially
> libraries created to optimise code reuse inside a certain company (and
> its associated community of external developers), not inside the GNOME
> community as a whole.
> 
> Am I the only one to think this is suboptimal?
> 
> EEL and GAL should merge. Both intend to be some kind of "libgnomeui
> level 2" -- why not go ahead and just do it?
> 
> At a quick glance, EEL contains following components:
> - smooth text widgets
> - a powerful canvas widget with assorted supporting widgets, such as
>   an ellipsising text field etc.
> - some font management stuff
> - a password dialog
> - a radio button group widget
> - a wrapping table
> - an extended CTree
> - some GTK+ extensions to support it all
> - some GNOME VFS extensions
> - XML utility functions
> 
> GAL contains:
> - a paned widget
> - very powerful tree and table widgets
> - a powerful text widget
> - menu utilities
> - the E Shortcut Bar
> - a reflow widget
> - GtkComboBox and some premade boxes based on it
> - XML utility functions
> - other utility functions
> 
> I surely missed some stuff here, but it looks like these libraries
> could and should be merged into one. The XML utilities, the
> reflow/wrapbox, the tree widgets, etc. overlap at least a bit; and
> IIRC there is ellipsising somewhere in GAL, too.
> 
> People having to agree on what to put in that common library will also
> prevent the wheel from being reinvented too often and keep the library
> from bloating. As it is now, basically Ximian and Eazel seem to put
> into "their" libraries whatever they see fit for further reuse by
> company projects.
> 
> For the future, having two companies agree on that library will
> prevent further corporate players from pulling their own EEL or GAL
> again and make them work towards the goal of improving the common
> library. Code review from more people coming from different corporate
> cultures will also help make the code cleaner and even more reusable.
> 
> On the user side, it would mean less memory usage, one less
> dependency, and less confused users worrying about corporate
> influence. The existence of Ximian and Eazel alone is already the
> cause of much FUD, and I think it's in the best interest of any
> company to keep their profile low in the "neutral" areas of the GNOME
> system core.
> 
> Note that
> - I am not claiming that the GAL/EEL dualism is evil -- it's just not
>   as good as a common library would be.
> - I am not thinking that this is an urgent issue, but it's nothing to
>   be left alone for long, either.
> - I am not proposing to do the merge myself, since I'm neither a
>   Ximian developer nor an Eazel developer.
> 
> I'm just an observer. Call me a nuisance if you want :).
> 
> mawa
> --
> Wenn die Wochentage Länder wären...
> ...dann wäre der Samstag Jamaika.
> 
> _______________________________________________
> gnome-hackers mailing list
> gnome-hackers gnome org
> http://mail.gnome.org/mailman/listinfo/gnome-hackers




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]