Opened 20 years ago
Closed 20 years ago
Last modified 5 years ago
#642 closed defect (invalid)
ALL: Mac midi drivers not working anymor
|Reported by:||fingolfin||Owned by:||SF/jamieson630|
|Priority:||high||Component:||Port: Mac OS X|
Since today's imuse changes, CoreAudio & Quicktime midi drivers on OS X don't work anymore. They did work just fine yesterday. The Adlib does seem to work OK, though.
Ticket imported from: #656018. Ticket imported from: bugs/642.
Change History (11)
comment:1 by , 20 years ago
comment:2 by , 20 years ago
OK it compiles fine again here, but the problem still persists. What makes it annoying that stuff that syncs with iMuse doesn't work, e.g. the opening credits in MI2 don't start playing.
comment:3 by , 20 years ago
In what way do they "not work"? Does music simply not play, but without any warnings/errors? Or are you getting "Cannot find sound" errors to boot?
Also, are you playing with the Mac-version resources, or some other version?
No other port is reporting any problems whatsoever -- the latest architectural changes are working under -ewindws and -eseq. Furthermore, none of the architectural changes directly involved the Mac-specific drivers; the changes are in the common areas and would theoretically break all drivers at once.
Before I go digging around for a problem potentially in code that's not even included in my own builds, could you verify for me that 0.3.0b is still playing music fine on your box? Just so we can isolate the problem to CVS ScummVM, and not to some other box-specific problem.... Thanks.
comment:4 by , 20 years ago
1) "not work" means that music doesn't play (with CoreAudio and QuickTime drivers), it means that I can't hear anything, and as was stated in my report, that syncing to iMuse effects doesn't take place (i.e. the MI2 intro just hangs forever, because it is cued to certain iMuse events I assume).
2) The "Mac-Version" resources for music fall into two catgories: not supported by us at all, or MIDI. So I don't see how this is relevant :-) But the problem occurs with PC versions just as well as if with Mac versions.
3) Every backend is different, so it is very well possible for a change in a common area to only change 1-2 ports. E.g. the Mac backend is big endian, the ones you list aren't. And windows/seq hardly are an exhaustive list, what about ALSA and morphos I wonder, or runing it on Linux/PPC etc. :-)
4) Yes it works with 0.3.0b, and also with CVS from before your changes. So I don't think it's my machine being broken.
5) I just tried Simon and see that it doesn't play music anymore, either. So it's probably not the changes to imuse, but rather some of the changes in sound/.
comment:5 by , 20 years ago
|Priority:||normal → high|
comment:6 by , 20 years ago
Kirben has also confirmed this in mingw builds. That's something I can at least test with a little more. Hopefully we can put this to bed quickly now.
comment:7 by , 20 years ago
Okay, music problem in mingw vaporized into thin air after a few tests, so that's out. However, another problem that came up was an SDL segfault anytime the user tries to F5 into the GUI. The crash was still present even after removing iMuse from the system entirely. Whatever is causing that does not sem to be iMuse-related.
Fingolfin, can you tell me if you have a problem with SDL segfaults when trying to get into the GUI on Mac? Like the music problem, this bug does not seem to be manifesting itself on VC++ or Linux gcc builds.
comment:8 by , 20 years ago
Music still not working here.
No crash upon F5 here. I don't see how those two could be related, anyway. If the F5 crash is reproducible, a bug report including backtrace would be highly appreciated.
comment:9 by , 20 years ago
Okay, music problem was apparently a phantom related to dependency tracking failures in the two compilers reporting problems. Closing.
comment:10 by , 20 years ago
|Status:||new → closed|
comment:11 by , 5 years ago
|Component:||→ Port: Mac OS X|
I have to correct myself, this wasn't latest CVS. Latest CVS doesn't even compile here.