[Scummvm-devel] ScummVM 1.1.0 has been tagged

[Max Horn]

Am 01.04.2010 um 12:16 schrieb Filippos Karapetis:

> I don't think there is time for this before the release (which has already been tagged), or am I wrong?

We have a vote running on postponing the release "until all major regressions are fixed"; I thought you were replying to that very email? :-)


> 
> In any case, perhaps the best course of action would be to introduce the old code again, with the new
> one ifdef'ed out? Or is that not possible? I merely made a suggestion, and I'm not familiar with the engine
> myself... this isn't bugfixing per se, rather reverting commits which aren't "production-ready".

This is most definitely bug fixing -- it removes a bug, doesn't it? :-)

Or are you saying that we should take the parallaction engine from say 0.11.x or 0.12.x, paste it over the current parallaction engine and then adapt the old code to work against our new common/backends/sound/graphics code? This would *probably* fix all known regressions. It would also would undo all other bug fixes and improvements since then. And also might introduce some new bugs due to the rushed adaption of old code to the new infrastructure code... 

This sounds rather crazy to me, but certainly would be doable, although it strikes me as being not at all easier than just using the same time to properly analyze the bug, bisect which revisions introduced the regression ("git bisect" is super duper helpful for this, by the way), then analyzing how this revision causes the error, and then developing a proper fix...

Maybe you meant something else anyway? In any case, "somebody" has to do it (and be careful not to break far more than is fixed). Again, we need a volunteer ;)


Bye,
Max