Release GPSP For Pandora... From Notaz!


I'll ask on behalf of the Wiz/Caanoo users, could you do a build for for the GPH handhelds? The current version of gpSP on the Caanoo isn't that great, especially the sound issues.
Probably, what's painful with Caanoo is that it's sound runs at 44200Hz or similar instead of standard 44100, so sound sync is tricky. I had it worked out for Wiz years ago (it has the same problem), but I don't remember the details.
 
This works really well! Thanks Notaz. Only issue I experience was that I got a random restart of the game when going down a pipe. It just crashed and went to the GBA Bios screen and then started the game again. I had done a savestate prior and could not replicate the issue.


The filter really does look sublime on the Pandy screen, and it's smooth as silk. (At least on my 1Ghz, I have sold the old 256Mb unit)
 
hmm, i have recently changed my pndXmodmap to make the game-buttons behave as mousebuttons. I think this was no issue with the last gpsp, but now I can't use the menu.
 
I like the look of pixel art with Scale3x applied to it - much better than no filter or blur.


Are there any other pixel art scaling algorithms that could be tried? (of course they have to be fast enough to be applied in real time, unless there's a way to preprocess and cache these things) On IRC someone (I forgot who) showed me this, which looks pretty cool: (for comparison, Scale2x is the same as EPX)


http://research.microsoft.com/en-us/um/people/kopf/pixelart/supplementary/multi_comparison.html


One minor thing: I'm using pandafe, and when exiting the emulator I get some error message about not being able to open /dev/tty or something. Could be a problem with pandafe, I don't know.
 
One minor thing: I'm using pandafe, and when exiting the emulator I get some error message about not being able to open /dev/tty or something. Could be a problem with pandafe, I don't know.
I just happened to notice this today as well. It seems to be intermittent, sometimes but not always returning an exit code of 256. It may involve running some certain other program first that creates a later side effect when running gpsp.


In any case, the quick solution is to set the expected exit code for the gpsp program definition to 256, which will suppress the error dialog. I also just happened to have released a new version of pandafe today, so I went ahead and included an updated gpsp program definition as a convenience for anyone else who might encounter this.
 
Did a little research into whats going on here.


First, the error message _wb_ refered to is "open /dev/tty: No such device or address". This message comes from trying to write to /dev/tty (the terminal connected to the current process), in order to send some message.


gpsp.sh yields this error message at the end of its output, whether its run from pandafe or via pnd_run.sh, unless run from a terminal (again, via either pandafe or pnd_run.sh). When run from a terminal the actual error message is revealed as:


"KDSETMODE_KD_TEXT: invalid argument"


Interestingly, when pandafe is run from a terminal, launching gpsp.sh returns exit code 0, which explains why the problem looked intermittent to me (I sometimes, but not always, test from a terminal). When its not run a terminal and thus /dev/tty is not hooked up, the no such device error causes an non-zero exit code.


This is arguably a bug in pnd_run.sh -- non-zero exit code if the pnd (quite reasonably) writes to /dev/tty. Its certainly a bug in pandafe given the loudness of the error dialog message when its encountered. I'll do something further on my end to make sure /dev/tty is available to spawned processes. Edit: new build on repo deals with this launching pandafe with "screen -D -m bin/pandafe $@". This makes /dev/tty available to spawned apps (and allows messages sent there to be captured with stdout/stderr). Nice.
 
Last edited by a moderator:
Back
Top