There's a limit to how far we can go in supporting stuff that depends on bugs. Almost every bug fix potentially breaks something, but we can't put every detail behind a quest rule check. It's more work (now and indefinitely), it's more overhead, and it's more time we're not working on other things. Also, there's a limit on how many quest rules we can add before the quest format has to change, and we're very close to that limit.
You can make a quest that crashes 2.50.0, but not later versions. Obviously, we're not going to support that. Generally, if a quest works right in the newest version and has issues in older ones, I'm okay with that.
Fine with me. You could create a batch file or launcher that runs
zelda-w.exe -standalone "MyQuest.qst"
Then you could leave out the instructions and most of the support files. -notitle might be useful, too.
If it were minor, I wouldn't care. It doesn't break anything that I've made, but the number of quests that use bitmap drawing for this purpose that will break, isn't insignificant. I believe that 'Lost Kingdom of the Banana Blood God', will break, in 2,50.2, because of this. Anything with scripted darkness, in general, will become unplayable; and there are quite a few of those.
It's the 'unplayable' part, that matters.
It may just be best to revert, if that's the only sane option, and leave the bugfix out there in the source for the o-s project, and later fixing. Let it be fixed, when a quest file format change won't be problematic.
With regard to QRs, that's why I'd use a version check when loading the quest file, to handle this sort of thing.
If every bug fix potentially breaks something, maybe don't fix bugs unless they're serious? I mean, is it not a bit silly to risk breaking stuff for minor fixes?
For major stuff, yeah, you have to take the risk. But I would think the minor stuff isn't worth it, really.
But what I find most disturbing here, is that you're starting to talk about breaking people's quests as if it is no big deal. You may not mean it that way, but I certainly HOPE you don't.
Eventually, everyone reaches a point, where they become weary of trying to back-support every minor detail. The perpetual arguments about passwords; and the continual onslaught of feature requests, while many bugs persist, and are abused to make games, only worsens this sensation of anguish; that leads to a certain level of apathy in development processes.
Thanks for notifying me about the bug in InvertedCircle(). I'll post an update (with a notice concerning it) to stdExtra post-haste!
Modify it, with an additional PutPixel() before anything else, that places an invisible (colour) pixel on layer 6, at 0,0; before drawing to the bitmap. That should prevent it from being a problem with 2.50.0/1 compilations, and will maintain compatibility with 2.50.2. .
Edited by ZoriaRPG, 26 October 2015 - 09:34 PM.