#987 LOOM: Some bits of dialogue are repeated

[Dan Fabulich]

In the attached save, double-click on Rusty. Bobbin
will say "Strange place to take a nap" twice. This
happens at random throughout the game, but this was the
easiest place to reproduce it.

Win2K, ScummVM 0.5.0pre-cvs Built on Jul 10 2003 20:29:19.

[Dan Fabulich]

Saved game right in front of Rusty

[Torbjörn Andersson]

Logged In: YES
user_id=577918

I don't know about this particular case, but in at least one
case I looked at a while ago the script actually printed the
message twice. Probably because if it was only printed once
it'd time out too soon. When I tried the same scene with the
original interpreter, the text wasn't removed in between the
two printings. In ScummVM it was.

It's still a bug, of course. I just wanted to point out a
possible explanation. :-)

[Max Horn]

Logged In: YES
user_id=12935

The "Strange place to take a nap" is also printed twice. The script
basically has a small loop like this:
for(i = 0; i<2; i++) {
print(...)
waitForMessage();
}

Now, the script prints the text once, then reaches the
waitForMessage(), which puts the script to sleep until the message
is removed from the screen. Then once the message times out,
the following happens: it times out, which leads to stopTalk to be
called, which redraws the screen area the text was in. *then* at
the end of the scumm loop iteration, the screen is redrawn. Then
in the next iteration of the scumm loop, scripts are processed. Our
sleeping script resumes, and prints the message again. As a
result, there is (at least) one frame without the text visible.

The only way to "fix" this would be, AFAICT, to change the order in
which message/script/redraw processing are done. However, the
current order definitely is right for the majority of games, and I'd
be surprised to learn that Loom is different here... maybe there is
some other subtle difference we are not properly covering yet.

[Max Horn]

Logged In: YES
user_id=12935

Actually, I rechecked some disassembly, and the order *is*
different for V1-V3 games (well at least for Loom and Zak -
somebody should verify this in Indy).

[Dan Fabulich]

Logged In: YES
user_id=611886

Cool. I can verify this in Indy when I get home. How would
this bug manifest itself in general? What kind of
regressions should I be looking for?