RE: GTK+-2.3.0 released [unstable]

[Murray Cumming Comneon com]

> From: gtk-app-devel-list-admin gnome org  
> you're comparing statically compiled win32 binaries to a 
> shared library 
> binary.  Of course it's going to make a difference what your current 
> gtk+ version is if you have it dynamically linked.  So if that's a
> problem for you, statically link your binaries and distribute 
> your stuff 
> like win32 does, already compiled.

Erm, no. An application that is built against GTK+ 1.2 will continue to
function wether or not GTK+ 2.0, or 2.2, or 2.4 is installed, because GTK+
1.2 is a separate library that installs in parallel. They don't interfere
with each other _at all_. And what's more, although GTK+ 2.0, 2.2, and 2.4
are not parallel-installable, they only add API. API/ABI compatibility is
guaranteed, so installing GTK+ 2.4 will have no effect on alread-installed
applications that were built with GTK+ 2.2.

People get confused about this stuff. The answer is to only complain about
real problems rather than imagined ones - if an application actually stops
working because you installed a new GTK+ version then that would be a real
problem.

> When staying up to date with a library you have to expect certain 
> functions will get updated, deprecated and replaced. The only 
> thing you 
> can do besides the windows way (statically link or have every program 
> bring it's own version of the library in it's own directory) 
> is to not 
> use functions that are on their way to becoming obscoleted.  It's not 
> like this happened over night to decide to remove a function 
> from gtk, 
> you had plenty of time to port the code if you really wanted to.

No, no, no, that's the Windows way. GTK+ and GNOME and most of the linux
world understand API stability issues and we don't fuck with application
developers or break their applications.

Murray Cumming
www.murrayc.com
murrayc usa net