kicad-developers team mailing list archive

[Lorenzo Marcantonio <l.marcantonio@xxxxxxxxxxxx>]

On Sun, May 04, 2014 at 03:19:41PM -0500, Dick Hollenbeck wrote:
> I don't think using gtk directly from KiCad is the proper solution.  I would have been
> hours ago if it was that easy.

Never been suggesting that. Only noticing that the code in that class
was doing the wrong thing from the gtk side.

> Using wx is the solution.  But right now its like playing jango on a treadmill.

Don't know that game sorry :P 

> My Ubuntu package repo is at 2.8.12.  So there are a lot of linux users parked on that
> version of wx.  It needs to work back that far.

At work I still have to maintain a GTK+1 application. I know the feeling.

> I may simply need to supply a modern wxEventLoopActivator to the older code.  I got a
> feeling that that's been made smarter lately.

Like the Join function added for the path code? It could work if
something else is not breaking it. That code path (on the gtk layer)
changed quite a bit. Also who knows what's being done on the windows and
mac side, too... never used wx before kicad, I don't like it (liking
even less...)

BTW just finished a rebuild with wx 3.0 (eeschema only) and it works; no
asserts or things, simply works.

IMHO if you find the right technique for 2.8 it should work for 2.9 too.
I'd say that at the moment it's a 3.0 only feature. Useful to know for
other people testing the branch.

-- 
Lorenzo Marcantonio
Logos Srl