libxmp 4.6.3

[AliceLR]

Making a new pin for the next bugfix release of libxmp. I don't really have a release date in mind for this aside from "sooner than 18 months from now" 🙂

I added a few more issues to the 4.6.2 milestone. Aside from those, nice-to-haves would be:

• Finish fixing the Desktop Tracker loader (revisit dmf_load.c, gtk_load.c and dtt_load.c #624). My patch has this mostly complete aside from the packing format and effects testing.
• Address some number of modules from the OpenMPT test backlog (Broken OpenMPT test modules. #497).

[sezero]

We have four small patches in since 4.6.1, one being the bad cve fix to miniz reverted. See PR/830 for the proposed changelog.

@AliceLR: If you don't have any patches scheduled, I suggest that we make a 4.6.2 patch release this month and move the remaining items in 4.6.2 milestone to 4.6.3.

[AliceLR]

I kicked most issues down to 4.6.3 since they're non-urgent. There's one thing I would like to check into first before 4.6.2 is cut: OctaMED effect 1Fxx, since the change in 4.6.1 might have affected some modules to play differently incorrect than before. This is a delay/retrigger so relatively easy to implement in libxmp semantics, I just never got to it.

[sezero]

OK

[AliceLR]

Update: still working on this. Not only was I correct about 1Fxx being wrong, but I also got MED command FF3 wrong. Patches for this and MED3 song files should be ready tonight or tomorrow.

edit: and to properly regression test this, I have to fix MED3 and MED4 time factor and tempo... :/

[AliceLR]

Something came up with libfuzzer, don't tag 4.6.2 just yet.

[sezero]

Something came up with libfuzzer, don't tag 4.6.2 just yet.

Sure

[sezero]

OK, 4.6.2 tagged and and released, both at sourceforge and at github.