On Fri, May 27, 2011 at 4:57 PM, Benjamin Johnson <obeythepenguin@gmail.com>wrote:

On Fri, 27 May 2011 17:22:47 -0400, reza farivar <rf.opensource@gmail.com> wrote:

Thanks Benjamin. I suggest we wait for a day or two before committing to

...

back-porting. As Jorge mentioned, it might not be related to the sockets after all, in which case it might be fixed with his small patch. I also like to see what Manolo has to say on the issue, he seems to be quite active these days (lots of patches committed to FLTK1.3 in the last month).

Well, for what it's worth, there were actually two different things on Windows. First sockets didn't work at all, because of a broken select implementation, then it had 100% CPU usage because it didn't check if FLTK was idling. I sent in a patch for both and never heard back, though then again no one's really maintaining 2.x any more.

Well if sockets didn't work at all on win32 you couldn't get any network traffic, right? So you probably had to fix the socket code anyway. Here, I do have network traffic, it just seems it needs a push to get stuff working. Maybe the event queue has to have more than one element to return something? I'm confused. By the way, unfortunately Jorge's patch (setting timeout to something other than 0.0) didn't solve the problem. So I'm running out of ideas, unless Manolo comes to rescue ;-)

Anyway, good luck, and hope you get a response! :-)

~Benjamin