Re: libgtk3deprecated (Re: About GTK+ 3.0 and deprecated things)



Am Sat, 19 Jul 2008 12:35:44 -0400
schrieb muppet <scott asofyet org>:

> 
> On Jul 17, 2008, at 2:18 PM, Tim Janik wrote:
> 
> > On Thu, 17 Jul 2008, Martin Meyer wrote:
> >
> >> 3) How long / how much effort would it be do to this? Is it
> >> something that has to be done by experienced GTK hackers or can it
> >> be handled as
> >> a side project?
> >
> > I think any moderately experienced developer could tackle this (e.g.
> > someone who has implemented his own widget at some point).
> >
> > A closing word on the library name, since this'd be an ABI break,
> > such a library only makes sense for Gtk+-3.0 (or 2.99.0) and should
> > probably advertise that it ships deprecated Gtk+ stuff. So the best
> > name is probably libgtk3deprecated for this (there could possibly
> > be a libgtk4deprecated as well at some point).
> 
> Seconded.  And there's a precedent, too -- Gtk2::Deprecated for gtk2- 
> perl, which contains the widgets that were already deprecated as of
> gtk +-2.0.0.
> 
> http://markmail.org/message/qdo73axb72rr5i3k
> 
> The maintenance effort here has been effectively nil -- maybe six or  
> eight hours over four years.  A fair portion of that has been  
> answering help request emails with "this stuff is old and busted,
> port to the new hotness."
> 
> http://markmail.org/message/qdo73axb72rr5i3k#query:gtk2%3A%3Adeprecated 
> +page:1+mid:olfclnxggu3piy2e+state:results

Hey,

While the idea of a library containing frequently used
deprecated code is pretty self-evident, to me this looks like a
path that might actually make the situation worse. Basically I am
concerned that keeping unmaintained code alive, and even introducing a
new, "fresh" library for it, means we will still see it around after
Gtk 3 or 4 is outdated. Of course I see how it is supposed to make life
easier for developers, but still that means using code that contains
bugs and is probably not well tested. To me at least, sticking with Gtk
2 with those applications that won't port to new code, is therefore a
more reasonable idea.

That said, it's just my opinion.

ciao,
    Christian


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