SCI: Update Sound:vol in processUpdateCues

[sluicebox]

Sound:vol is updated when updating cues starting in SCI_VERSION_1_MIDDLE.

Fixes bug #10244. SQ4 localized floppy versions depend on this when getting
in the orange ship, they fade music and wait for the volume to reach zero.

Confirmed against asm that this is not in SQ4 floppy (early) but is in
LSL1VGA (middle) and SQ4 localized floppies and SQ4CD (late).

[m-kiewitz]

Please don't commit yet, I have to verify various interpreters to make sure that everything checks out.

[bluegr]

Any news on this?

[m-kiewitz]

Any news on this?

Still on it.
Very busy at work, had to work through one whole weekend this month already :/

[Kawa-oneechan]

Hang in there!

[bluegr]

Ping: Any news on this?

[m-kiewitz]

Ping: Any news on this?

Not yet. Sorry.

[sluicebox]

If it helps, the update cues function can be quickly found in most DOS SCI1 interpreters by the unique instruction:

05 7f 00 add ax, $7f

and in mac interpreters:

D0 7C 00 7F add.w #$7F,d0

The function ends with three consecutive calls to the set selector function prior to the change, afterwards there's a fourth for volume.

[sluicebox]

I'm going to merge this next week unless there's a reason not to.

• We already do this for SCI32 a few lines above in processUpdateCues()
• I'm seeing this in every interpreter I inspect
• Depressingly, I just saw that csnover wrote this same fix in 2017 minus the version guard: csnover@00be923

That last one is so depressing it's a wonder I don't click merge right now.

[m-kiewitz]

* I'm seeing this in every interpreter I inspect

My point is making sure that it's in every single SCI1 MIDDLE interpreter (Leusire Suit Larry 5 Demo, Conquests of the Longbow Demo, Leusire Suit Larry VGA Remake (2.0+2.1+EGA), , Astro Chicken 2, King's Quest 5 Japanese English FM-Towns) and definitely not in any SCI1 EARLY interpreter (King's Quest 5 Floppy versions, Leisure Suit Larry Demo, Mixed-Up Fairy Tales EGA+VGA, XMAS Card 1990 EGA+VGA, Space Quest 4 EGA+1.052+1.1 Floppy, Mixed-Up Mother Goose CD and Mixed-Up Mother Goose FM-Towns).

If you have checked all of these, then it's okay to commit.
If you have checked some, please tell me which you have checked.
Sometimes a change wasn't done exactly during one of our SCI version splits, and thus such a change can cause an obscure error. That's what I want to avoid at all costs.

If for example the change was done for the last few games that we consider SCI1 EARLY, then we would have to check on top if the game is accessing that selector at any point and making sure that not setting it will cause no issues. Because if it would, then we would have to add that specific game id too for your fix.

That last one is so depressing it's a wonder I don't click merge right now.

I don't understand. What's do depressing about a fix written in 2017?

[m-kiewitz]

@sluicebox
Can you just please tell me which of the games you checked, that I previously mentioned?

[sluicebox]

I keep forgetting about this. Here are versions I've verified from ye olde notes:

Other than the Longbow demo, it cuts cleanly along the SCI_VERSION lines, and Longbow demo doesn't care either way since nothing in it depends on the vol property, which I tested.

NO		QFG2 1.105 (final)
NO		KQ5 0.000.051		SCI1 Early	
NO		Mixed-up Fairy Tales	SCI1 Early
NO		SQ4 1.052			SCI1 Early	
NO		XMAS1990			SCI1 Early	
NO		JONES Floppy			SCI1 Early	
YES		JONES CD			SCI1 Middle	
NO		LONGBOW DEMO		SCI1 Middle	
YES		LSL1VGA 2.0			SCI1 Middle	
YES		LSL1VGA 2.0 EGA		SCI1 Middle	
YES		LSL1VGA 2.1			SCI1 Middle	
YES		Ms Astro Chicken		SCI1 Middle	
YES		SQ4 Localized Floppy	SCI1 Late
YES		BRAIN DEMO			SCI1 Late	
YES		BRAIN 1.000			SCI1 Late	
YES		SQ4CD				SCI1 Late