Time for a break...

[Colin Barrett]

On May 21, 2007, at 7:03 PM, Ethan Blanton wrote:

> I firmly believe that we should *not* drop win32 support in libpurple,
> nor should any patches for that compatability be out-of-tree.

On May 21, 2007, at 7:03 PM, Ethan Blanton wrote:

>  (Of course, it might bitrot going forward, and be
> removed, but still...)

You pointed out a pretty serious flaw in your own argument, Ethan.  
Dropping win32 support from Pidgin means there are no *current*  
clients using libpurple that run on win32. I don't see how you expect  
to maintain win32 support without a building, running consumer.

I see this all  the time at my day job working on a large, open- 
source, cross-platform application. The only reason we support  
anything other than Mac OS X, Windows and some versions of Linux is  
the fact that we have a building product that other people can take  
and improve to a working state on other OSes. Dropping support in  
Pidgin itself would hinder that.

In that same vein, the other reason we only have three "tier 1"  
applications is that those are the largest three OSes by user count.  
If Haiku suddenly eclipsed Linux or Mac OS X in terms of number of  
users, you bet we'd add it to tier 1. Windows has a large number of  
users, and effectively abandoning them doesn't make a whole lot of  
sense.

Other than dealing with support requests and having GTK+ suck on  
Windows, does Windows support directly affect Pidgin development? I'm  
not all that familiar with exactly what happens when writing GTK+ and  
GTK+/Windows in the same application.

Abandoning users over a philosophical difference is frankly silly. Of  
course, my entire argument could be swept under the rug if someone  
stepped up to implement a native win32 client. Perhaps it should be  
"he who shouts loudest," which would seem to be you, Ethan ;)

-Colin