winegstreamer and Devil May Cry HD Collection Cutscene Lockups

[jjbarr]

The Devil May Cry HD Collection's cutscenes being unplayable is a known and documented issue with Wine. However, it manifests differently in tkg than it does in earlier versions of proton, seeming to lock the application up (likely due to improved MF support...), and judging by the logs making at least moderately more progress. You can find the complete log file of my attempt to run the application here here:
steam-631510.log, but given that the issue is linked to cutscenes and the lines immediately precede what appears to be the final SEH trace, I suspect that the most relevant lines are

89711.380:015c:01cc:fixme:mfplat:mf_media_type_from_caps Unrecognized video format video/x-ms-asf
89711.472:015c:01cc:fixme:mfplat:mf_media_type_from_caps Unrecognized format WMV3
89711.473:015c:01d4:fixme:mfplat:mf_media_type_from_caps Unrecognized format WMV3
89711.473:015c:01cc:fixme:mfplat:mf_media_type_from_caps Unrecognized format WMV3
89711.473:015c:01cc:fixme:mfplat:mf_media_type_from_caps Unrecognized format WMV3

mf_media_type_from_caps can be found inside winegstreamer, and this is odd because winegstreamer does explicitly support WMV3 video. The fixmes seem to specifically be coming from this section, save for the "unrecognized video format" fixme, which is caused by the highlighted strcmp and all subsequent (because we're not decoding mpegs) returning inequality.

video/x-ms-asf is the container format for WMV. I have no idea what we're supposed to be doing with it, but it doesn't seem to cause any issues. The only indication that this might be a problem is that the application eventually segfaults, and this does cause a null return (although this could be coincidental).

Since there is explicitly code for handling WMV3, it seems strange that it would be considered unrecognized. Which figures, because it very much is being recognized: that fixme is actually spurious, and being generated for a known format due to the lack of an else in Line 926. So that's a dead end as well.

Honestly, don't really have a clue what might be going wrong at this point, and it's possible the only answer is to wait for more MF work to be done. I've logged run with trace+mfplat, in case that sheds any more light on the subject.

I have no idea whether or not this is really a good place to submit this bug (which likely manifests on upstream wine-staging, although I haven't tested it), but I figured it was best to start with the version of Proton I'm actually running.

[gabriele2000]

Any progress?

[ghost]

oof