Please add the skip dialog button (□) like the psp/pc build
According to PSL1ght's readme, the sysexit callback is not hooked up properly (the system call I guess), so when trying to exit a game, it will instead reset the PS3.
Lousyphreak, is there anyway to make your code migration to PSL1ght optional in case somebody doesn't want all these regressions? Because it would really suck if somebody had to resort to a fork just because of the penchant to be 'legitimate' (when this is still decidedly a grey area anyway given that the closed platform proprietor, Sony, does not want you either linking to PRXes in flash memory or developing unlicensed at all. Either way, Sony just freeloaded off GCC and GNU - so why anyone feels bad about us using the SDK, I dunno - it's not like they put that much effort into their documentation or the samples for that matter - licensed developers are really getting shafted hard given the shitty condition they even released 3.41 in).
I dunno, I'm just looking a this from a pragmatic point of view. I have yet to research PSL1ght or migrate SNES9x to it - but if I were presented with any drawbacks vs. the official SDK, then the decision is clear to me - functionality comes first rather than perceived legitimacy.
Distributing binaries made with the Sony SDK can still get you into trouble, it might be a low possibility, but it is there nonetheless.
One of my main reasons of osing the PSL1GHT SDK is that this might open the possibility to integrate the port right into the official ScummVM tree, including official releases like for any other platform, downloadable straight from the ScummVM website.
PSL1GHT is by the way almost usable: the missing callback is one of the big problems and the second problem is that the framebuffer is much too slow right now.
The callback will probably be fixed tomorrow, and RSX support is progressing fast too (phires tree can already ask the RSX to clear the buffers :P ).
As for progress on the PSL1GHT port: it's almost usable, after loosing 3 days to a bug with the linker, I got the base system including graphics running in about 2 hours. I think I might be able to play my first game today ;)
Way to make progress with it. As long as it doesn't crash the PS3 I'm happy to move forward.
I'm just looking at this from a pragmatic point of view - if the switch to PSL1GHT at the moment results in obvious downgrades in terms of functionality and performance, then it would seem to me that most would want to be able to keep compiling with the official SDK for the moment.
I know I would. Sony itself is not being legitimate with their SDK in any case - I've said this a dozen times, and there might be some technicality they can fall back on, but really, using opensource software as the backbone of your entire SDK and then NOT POSTING toolchain patches to a public repository of any kind is so NOT DONE. You can't say - 'oh I made a couple of patches to this GPL software, now the patches only get distributed to licensed developers who signed an NDA with me and only on an https site that requires an SSL certificates'. GPL stands for GNU General Public License - note 'General Public' in there. This flies in the face of the entire principle behind the GPL and why it even exists as a license. Instead of everyone worrying about 'oh I can get in trouble for using the SDK', how come nobody has yet to make a big deal about this? It's as bad as the Linksys prior example - they tried something similar with their WRT router firmware in the beginning, and eventually had to obey to the license they're SUPPOSED to be bound to.
Sony has been caught before in numerous GPL violations - so it's not like they give a shit about the law or any kind of rules bound by licenses, so why should you freak out over it?
You might need some (many?) #ifdef's and #define's, and you should base the code on the PSL1ght SDK as there are many functions which are in the Sony SDK but not yet in PSL1GHT.
I am currently busy writing a SPU scaler/blitter which will make drawing much faster than it currently is. It will be mainly aimed at EMUs or programs requiring upscaling, as there will be some MAME filters builtin. And all that while using almost no CPU :P
(I made the first tests yesterday evening and the results look VERY promising ;) )
ScummVM is already working close to the state it is with the Sony SDK, but the display is 1:1 mapped to the center of the screen, so its mostly experimental at this time ;)
In case you're going to drop the official SDK entirely, mind if somebody just makes a fork and tries to base it as much on the mainline repo (the PSL1ght one in this case) as possible? That is as long as there are still discernable differences between the two.
No, not at all, thats why it's posted on github, and not released binary only ;)
It would be nice though if the forking was done also on github so I might be able to merge useful changes back
as for SPU blitting: I got a working sample which needs below 3ms to clear a 1920x1080 screen.
I am hanging around #ps3dev on efnet if there are questions :)
OK, I have a working PSL1GHT env setup (I think) so I'm having a go at compiling this. Unfortunately, it errors out pretty quickly with the following:
backends/platform/ps3/ps3-main.cpp:5:22: fatal error: lv2/time.h: No such file or directory
make: *** [backends/platform/ps3/ps3-main.o] Error 1
i cant use graphic filters at all (super eagle, x2 sai and rest of them), cant fidn em on list
am i missing something pretty obvious or thats not yet supported
and thx to everyone that are involved in this shizzle, gj!