With JIT almost finished for our E-UAE, some of E-UAE's other problems are now painfully obvious, and we really really REALLY need a developer(s) work on it. Would it be worth someone creating one or more bounties? I think the two biggest issues are:
1. E-UAE doesn't use Compositing for scaling graphics, so a 640x480 window is a little slow, and anything larger is unusably slow. Using Compositing to do the scaling shouldn't be that hard (assuming E-UAE doesn't do anything too wierd), and would make a real difference on modern large monitors (and for Sam440 users whose monitors don't support a 640x512 screenmode).
2. E-UAE-Amigfx is the fastest version of E-UAE (hence why RunInUAE uses it by default), but it's Picasso96 emulation is broken, so we can't have Workbench 3.1 running in a lovely 16-bit screenmode at native resolutions. Picassion96 emulation surely must have worked in the distant past(?), hence I guess it was just broken by a code change at some point... and so it might be relatively simple to fix (if you understand how the Picasso96 emulation works - I don't have a clue). It *does* reportedly work in the SDL build of E-UAE, so checking where it differs from E-UAE-Amigfx (for P96) might be a good start.
So, what do people think? Would a bounty be the way to go?
EDIT: Just to be clear, I think we should have several smaller bounties, rather than one big one. So one bounty for Composited scaling, another bounty for Picasso96 Emulation, etc.
Classic AmigaOS ROMs and the Workbench disks are no problem at all now that they are included with every copy of AmigaOS. We just need some people to help work on UAE and help improve it so that the classic experience is nicer.
Add to that a long stading bug in all latest builds flying around : on exit our uae builds (that include all previous ones, and that on which Rachy works) it do memguard hit. So usually you can't notice or have crash right now, but after you do run/exit uae about 10-20-30 times, you crashes in end (and, i somehow sure that memguard crashes are symptoms because of which lately real crash come).
I report it to Rachy of course too, but i fear its out of his work to fix all kind of bugs. If anyone in interest, i can provide all the info i found (in few words , it crashes in freeding some sound resources). Maybe even easy to fix.
P96 not working in the amigfx driver is not a bug. It has never been implemented at all. It wouldn't be particularly difficult to do, someone just has to mimic it from the sdl driver.
@BSzili & SpotUp I'm not against a bounty for that (port new UAE from scratch), but it would probably be a lot more work than the two bounties I've already suggested. Maybe better to leave that kind of bigger harder bounty until our existing port works relatively well?
P96 not working in the amigfx driver is not a bug. It has never been implemented all. It wouldn't be particularly difficult to do, someone just has to mimic it from the sdl driver.
I think this is more about trying to attract a developer in the first place
Yes, thats why its important that its a reliable developer like i said and not somebody that quits after 4 weeks. You could probably attract some developers but alot of projects just dies after some time because they are not serious or dont understand how much time it takes. I would happily donate if i knew things would be done.
(it might become very important when SMP/AMP becomes available as some SW will start to fail or otherwise load only core0 (or so). People could then for example run Lighwave on some core and screamernet on the rest of the available cores.)
- Kimmo --------------------------PowerPC-Advantage------------------------ "PowerPC Operating Systems can use a microkernel architecture with all it�s advantages yet without the cost of slow context switches." - N. Blachford
By coincidence, I had sent you a PM on this site shortly before you made this thread. I notice you still haven't read it, so I thought I'd point you to it.
As for P96 support in amigfx, I think Thore Sittly (aka Thunder on the e-uae JIT blog) was planning on writing something for this.
I'd donate for sure though I must admit I don't use e-uae that much anymore on my X1000 since I got my Amiga 1000 and Amiga 1200 back up and running sweetly in my mancave
Yes, thats why its important that its a reliable developer like i said and not somebody that quits after 4 weeks. You could probably attract some developers but alot of projects just dies after some time because they are not serious or dont understand how much time it takes. I would happily donate if i knew things would be done.
IMHO you have it backwards (kind of). You get a large bounty to ATTRACT someone good/capable (hopefully). If they fail to achieve the bounty requirements, then someone else can have a go. i.e. If we get someone "unreliable" then we loose nothing but some time.
We'd just need to make sure the bounty requirements are clearly defined, and don't allow someone to continue working on it forever without making any progress. (I'm not a fan of 100% hard deadlines, but you do need some flexibility.)
IMHO (Amiga) bounty terms are sometimes not well thought-out. I'd prefer to see open public discussion of proposed bounty terms, rather than closed private discussions between a small number of people (as that often seems to lead to obvious problems being overlooked or downplayed).
Compositing for scaling graphics would be very usefull(and other improvements), if a bounty is needed to make it happen then so be it. But i would prefer to donate to the developer directly when i see that someone is doing some good work. And that reminds me that i should hand out some money tonight because it's been a while since last time :)