[Scummvm-devel] ScummVM 0.13.0: Release status 2009-01-19 -- Branch

[Max Horn]

Am 19.01.2009 um 16:40 schrieb Eugene Sandulenko:
[...]
> Still we have one release critical bug, and overall list of nice to  
> get
> fixed bugs did not shrink much...

Actually, personally, I would consider the matter of the unresolved  
GMM issues release critical. Like, how to trigger it on ports without  
a keyboard,  and what to do about games that use F6 for something  
else. But maybe that's just me.

>
>
> Release critical:
> #2489508 DIG: slower animations on 0.12

To elaborate a bit on our only (?) release critical bug: It goes back  
to patch #2060517 by salty-horse, possibly accepted prematurely :-( by  
myself: <https://sourceforge.net/tracker2/index.php?func=detail&aid=2060517&group_id=37116&atid=418822 
 >.

In the past, there were some discrepancies how the SCUMM engine  
interpreted / displayed the "talk speed": Internally it was used to  
set the talk *delay*, but to the user and in the config file (I  
think), we displayed it as *talk speed*. One suggestes that higher  
numbers mean slower, the other that higher numbers mean faster  
subtitles.

THis patch was supposed to resolved that, by changing the talkspeed to  
really act as talk *speed*. Some of the problems this has subsequently  
cursed (in the above mentioned bug reports and others):
* existing config file entries are of course now interpreted  
differently. So if you set the subtitle speed to extra-fast in the  
past, then it now is extra-slow and vice-versa
* the default value for the talkspeed was not adjusted, resp. our  
interpretation for it. Consequently, we now default to a different  
(bad) subtitle speed, which is the primary cause for the above  
mentioned bug report and its siblings

Note that the talkspeed setting was never really well-defined.  
Different engines may use it differently than the SCUMM-engine, too.  
In particular, the default value for the talkspeed is 60 for all  
engines, and some engines rely on this (in the end: arbitrary!) value  
(it is 60 for historical reasons, SCUMM needed it that way). So, we  
can't just change the default value to make SCUMM happy, because that  
would break other engines.


To fix this mess, I see two way:
1) We revert the patch completely. This will slightly annoy people who  
used SVN builds, but at least all those users who didn't won't be  
bitten, and also new users are fine. The supposed benefits of the  
patch are lost, too.

1b) As in 1, but in addition, we simply adjust the GUI to say "talk  
delay" instead of "talk speed".

2) We devise a strategy to resolve the issues while keeping the patch.  
However, I don't really see a good way to do that right now.


Clearly 1 is much easier and quicker to implement. I propose that  
unless we finde a volunteer in 1-2 who promises to take care of the  
issue RSN, then we should just go with 1.


Bye,
Ma