{S} Interim maintainer afk

[camlost]

The thread is here: http://angband.oook.cz/forum/showthread.php?t=1483

Unfortunately, the code changes are no longer available, and I don't think I ever dled the files. Does anyone out there have the code changes?

[Bostock]

Also, S seems -- seems -- to run into trouble with UAC on Vista, since it never remembers my name and setting.

BTW I wouldn't worry about this particular bit - seems to have been a corrupt, erm, install, inasmuch as one can speak of installs here. Folder, I guess.

Thanks for looking into the extended fonts issue, and sorry that all I could do was draw attention to the issue, rather than providing help with the solution.

Bostock

[camlost]

Thanks for looking into the extended fonts issue, and sorry that all I could do was draw attention to the issue, rather than providing help with the solution.

Have you tried pure ascii mode? = (options) ! (display) b) Extended Fonts -> Pure Ascii

[Bostock]

Pure ASCII mode works fine, but... it's pure ASCII mode. :-)

[camlost]

Pure ASCII mode works fine, but... it's pure ASCII mode. :-)

Yes, but an easy workaround does change the priority of the fix.

[Bostock]

Yes, but an easy workaround does change the priority of the fix.

In that case, it might have been better to ask about Tiles+Text mode, since the person playing Extended ASCII probably is dissatisfied with pure ASCII. Or at least that's my situation. :-)

(Tiles+Text works fine.)

[Bostock]

The thread is here: http://angband.oook.cz/forum/showthread.php?t=1483

Unfortunately, the code changes are no longer available, and I don't think I ever dled the files. Does anyone out there have the code changes?

Does the file on the last page of the thread...

http://angband.oook.cz/forum/showthread.php?t=1483&page=5

...not have the code changes then? This file is still downloadable.

[camlost]

Does the file on the last page of the thread...

http://angband.oook.cz/forum/showthread.php?t=1483&page=5

...not have the code changes then? This file is still downloadable.

It's a compiled version. I guess I could grab the modified font files from it at least though.

[RogerN]

The compiled version in the above thread uses bitmaps instead of native font files. The bitmaps were generated directly from the Sangband .FON files.

I will try to dig up the code changes, but I think I may have lost them. I didn't consider my changes good enough to be worth keeping around. Although the performance was fine on my development system, I found out that it lagged pretty badly on systems without a decent GPU. It was playable but somewhat frustrating. A better solution would be fix the .FON files to avoid those "special" indexes.

I think the performance issues were being caused the usage of raster operations for combining the font bitmaps with the foreground color. I'm guessing that this operation is accelerated if you have the right GPU but is handled (very slowly) in software for other systems.

Unfortunately I couldn't think of another way to respect the various foreground colors without making multiple copies of the font images (which is horribly ineffecient) or switching to OpenGL, which was more work than I wanted to do. The OpenGL route might have even worse performance for all I know.

[camlost]

The compiled version in the above thread uses bitmaps instead of native font files. The bitmaps were generated directly from the Sangband .FON files.

I will try to dig up the code changes, but I think I may have lost them. I didn't consider my changes good enough to be worth keeping around. Although the performance was fine on my development system, I found out that it lagged pretty badly on systems without a decent GPU. It was playable but somewhat frustrating. A better solution would be fix the .FON files to avoid those "special" indexes.

I think the performance issues were being caused the usage of raster operations for combining the font bitmaps with the foreground color. I'm guessing that this operation is accelerated if you have the right GPU but is handled (very slowly) in software for other systems.

Unfortunately I couldn't think of another way to respect the various foreground colors without making multiple copies of the font images (which is horribly ineffecient) or switching to OpenGL, which was more work than I wanted to do. The OpenGL route might have even worse performance for all I know.

Thanks for the info. Do you know for certain that this is a problem with "special" indices (aka 127, the wall character)? That was my guess, but I didn't find any support for that guess. Also, wouldn't this be a problem for other Angband variants? It looked like vanilla uses chr(127) for wall characters also, but I haven't heard of them having any problems.

[RogerN]

Thanks for the info. Do you know for certain that this is a problem with "special" indices (aka 127, the wall character)? That was my guess, but I didn't find any support for that guess. Also, wouldn't this be a problem for other Angband variants? It looked like vanilla uses chr(127) for wall characters also, but I haven't heard of them having any problems.

Unfortunately I haven't been able to find the code changes. They're probably lost.

I do not know the specifics of the original issue, only that uninstalling the Asian language pack (on XP) fixed the problem for me. When I posted the original "fix" I did not have access to a Vista machine.

[Bostock]

RogerN, camlost, thanks for your continuing work on this.

Wouldn't decent GPUs be a fairly common fixture on Vista machines? I know the GPU on mine has stood up decently well on the occasions when I've asked a bit of it.

OTOH if it makes sense to implement Vista-friendly Extended ASCII at all, it practically needs to be a separate display type from traditional Extended ASCII... an additional complication.