OpenGL/GLX dual link error in VLC 2.2.0 git/master via MacPorts on OS X

[René J.V. Bertin]

On Saturday February 07 2015 12:41:37 Jean-Baptiste Kempf wrote:

> > Good luck with that.  My experience with that track is that they will just close your ticket and tell you to piss off.
> 
> 
> Where did we ever tell you to "piss off"? Who insulted you? When?

Matter of speaking and a personal impression undoubtedly, but I must say I concur with Jeremy that it's the impression one can get, given how some of my own tickets have been handled up to now.

> Can you point me to a build log?  I'll take a look later.  Too tired now...
I've attached the configure logs to the ticket. I don't have a build log right now; I'd have to rebuild from scratch to make one, but can of course do so. Maybe you can have a look at the configure logs first?
If you got a functional VLC.app with the exact same code then it's likely that something is wrong on my system. Or something was updated since (Xcode, autoconf/automake, whatever) you last tried.

> Not understanding that hacks are not welcome in a massive codebase like
> VLC is beyond me.
> 
> Problems should be correctly fixed, not hacked. We're the ones, in the
> end, that have to maintain it, not you.
> 
> All other platform maintainers understand that, except Macports.

No, I don't think that's true. At MacPorts we have to find and walk a fine line, but our needs are comparable to Fink's, HomeBrew's and a few lesser known others, and together we do represent a large enough movement that our needs cannot just be waved off because we don't do everything the exact Mac way, or (in this case), because we happen to have X11 stuff available.
Specifically, most every time someone submits a patch to MacPorts's trac, there's always someone who steps up and asks if the "fix" can't and shouldn't be incorporated upstream, in a proper fashion.

Please note that I have been asking from the beginning to find a proper fix for this regression (which, again, didn't exist in 2.1.5), or help to figure out what goes wrong, and why. Telling me that I shouldn't have the issue in the first place, while true in itself, isn't very helpful...

Of course I can also find a more hackish fix, like the fix I used to build against Qt4 instead of Qt5, but giving us the impression that we're on our own if we don't follow your build recipe exactly (because we can't) will only discourage us from sharing proper patches.

R.