Opened 21 years ago
Closed 20 years ago
#1055 closed defect (fixed)
DOTT: Actor layering glitch (hack-in-cvs)
| Reported by: | eriktorbjorn | Owned by: | cyxx |
|---|---|---|---|
| Priority: | low | Component: | Engine: SCUMM |
| Version: | Keywords: | ||
| Cc: | Game: | Day of the Tentacle |
Description
I played through Day of the Tentacle with a July 20 CVS snapshot. It works almost perfectly as far as I can tell, but there are a few minor issues I noticed:
1. At the end, when the kids are running away from purple tentacle, there's a brief sequence on the second floor where they are gliding instead of walking. I think this used to work, but I can't say for sure when that was.
I think it may have something to do with startWalkAnim() not being called from actorWalkStep(), but I don't know how to fix it.
2. Shortly after issue 1, the music changes. When it does, some notes are usually (always?) left hanging. The same thing sometimes happen when I push nurse Edna out of her room, but that's harder for me to reproduce.
3. If you try to replace the taped-up Dr. Fred with an unpainted mummy, one of the IRS agents will get suspicious and enter the room. When he does, he is drawn behind the mummy, even though he's standing in front of the bed.
This does not happen with the original interpreter when I run it with DOSbox.
Ticket imported from: #775097. Ticket imported from: bugs/1055.
Attachments (7)
Change History (28)
by , 21 years ago
| Attachment: | tentacle.s22 added |
|---|
comment:2 by , 21 years ago
I also forgot to mention that the music issues happen to me regardless of which music backend I use. (I used the ALSA backend at the time, but I also tried with the AdLib one.)
comment:3 by , 21 years ago
Gnaw, gnaw, erik. You should know better than to report three bugs in a single tracker item.
Also, can you reproduce any of those with 0.4.1 (read: are they regressions?)
comment:4 by , 21 years ago
> Gnaw, gnaw, erik. You should know better than to report > three bugs in a single tracker item.
Yeah, yeah, that was lazy of me. But some savegames applied to more than one bug.
> Also, can you reproduce any of those with 0.4.1 (read: are > they regressions?)
The mummy layering bug happens in 0.4.1 as well, so if it's a regression it's old.
I haven't been able to reproduce the hanging notes in Edna's room in 0.4.1, so that one may be a regression. I haven't played the game to the end with 0.4.1 so I don't know about the other hanging notes.
I'm pretty sure the walk animation bug is a regression. I don't remember it happening the first time I got this far in the game, but that was probably between 0.2.0 and 0.3.0. I don't know how it looks in 0.4.1.
comment:5 by , 21 years ago
| Component: | → Engine: SCUMM |
|---|---|
| Game: | → Day of the Tentacle |
| Summary: | DOTT: Issues with July 20 CVS snapshot → DOTT: Actor layering glitch |
comment:6 by , 21 years ago
I've filed separate bug reports for 1 and 2, so this bug report now only concerns the third issue.
comment:7 by , 21 years ago
There are various things which could have caused this. E.g. the changes made to the draw order / sorting. Or possibly a walkbox issue.
comment:8 by , 21 years ago
Or our draw order rules could be wrong, but surely that would have been noticed in other cases already...?
As far as I can see, Ted (the mummy) is actor 8 and the IRS agent is actor 7. The both have the same layer value (0) and Y coordinate (95), which means that the IRS agent is drawn first since he has the lower actor number.
It wasn't until today I realized that forceClip doesn't seem to affect the draw order at all. That does surprise me a bit...
comment:9 by , 21 years ago
I compared to DOSbox again, and as far as I can see ScummVM draws Ted and the IRS agent in the exact same positions as the original interpreter did. It's just the layering that's wrong.
comment:11 by , 21 years ago
I've played through DOTT with a January 23 CVS snapshot, and I can confirm that the glitch is still there. I'm attaching an up-to-date savegame, in case the older ones have problems.
comment:13 by , 20 years ago
Maybe it would be a good idea to have a look at the corresponding disasm. Maybe in DOTT, forceClip is used for actor ordering after all? Or maybe the actor layer should be set in more places than we currnetly do?
comment:14 by , 20 years ago
actor.layer isn't used until Scumm7 onwards. There is no sign of actor.forceClip in the actor comparsion of DOTT disasm. either.
comment:15 by , 20 years ago
| Summary: | DOTT: Actor layering glitch → DOTT: Actor layering glitch (hack-in-cvs) |
|---|
comment:16 by , 20 years ago
| Priority: | normal → low |
|---|
comment:17 by , 20 years ago
Since this is cosmetical and we have a workaround in CVS: Lowering priority.
comment:18 by , 20 years ago
Back to some old theory... to quote erik: >Or our draw order rules could be wrong, but surely that >would have been noticed in other cases already...?
Maybe, maybe not -- the sceneario isn't *that* common, I think. But now we do have a second report, and that report, too, seems to indicate that there's a problem: bug #1093867.
In particular, I now suspect that maybe for some (all?) of the newer SCUMM games, we should draw high numbered actors first (not the other way around). The attached patch does just that and fixes both issues.
A brief look at FT disasm doesn't seem to confirm this, though :-/
by , 20 years ago
| Attachment: | actor-draw-order.patch added |
|---|
comment:19 by , 20 years ago
Attached disasm. of processActors function from Day of the Tentacle if you want to check it too.
comment:20 by , 20 years ago
| Owner: | set to |
|---|---|
| Resolution: | → fixed |
| Status: | new → closed |
Save game in nurse Edna's room