| Re: [AD] windows resizing regression |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
On 8 Nov 2010, at 23:32 , Trent Gamblin wrote:
> You can pull up as many mailing list posts as you want, but Elias and I discussed this
> and every other issue for hours on end when we were creating the graphics API. I
> already stated the reason al_acknowledge_resize was initially used in combination
> with al_resize_display, and it's evidenced in the examples. The _initial_ reason for
> having it is as you pointed out. Let's not start an internet p%*@ing match here.
Ok, suggestion: take a deep breath and count to 10.
Many things have changed substantially between when they were first proposed/drafted and the current implementation (originally A5 was supposed to have a state-free API). So there is honestly no point in saying things like "when we designed it back in the day, this is what we came up with, therefore it doesn't matter what you say, I'm right and you're wrong." That's not helping anyone and I don't think anyone is interested in having a discussion like that.
It's also completely irrelevant.
Whether al_acknowledge_resize() was initially intended to be used in combination with al_resize_display() or not is not important. What's relevant is whether that's still the intention. As I said, my understanding is and always has been that it isn't (which, going by the date Peter quoted, is accurate, since I didn't start following A5 development again until some time in late 2008), which is consistent with the current documentation (which is still a bit sparse and not always unambiguous or accurate) as well as the OS X implementation.
Clearly something should be changed so behaviour is the same across platforms. That's the really important point.
I would say that the behaviour I outlined above is reasonable, and it agrees with the documentation as well as the current OS X and X11 implementations (both call the acknowledge_resize function from their al_resize_display() handlers and neither explicitly generates a resize event, although comments in the X11 source file say that X11 will generate one; presumably this gets passed through to the user through Allegro's event system - comments suggest this is undesirable). Quite why you seem to take this as a personal attack I don't know, because really it isn't.
Evert