Current Progress

[Liazon]

hey, you're trying to improve it on speed, which is already amazing in itself. Don't worry too much about the intermediaries cuz I'm sure it's gonna be all cool when it's all done :biggrin:

[Iambian]

In observing the way people use the sprite routine, I might as well write a "special" case of when the sprite that is being displayed is aligned. Most of the BASIC games out there use aligned sprites, so I believe this would speed up rendering by a massive amount. If you don't believe me, just look at the sprite routine in the source (_PFN.z80).

EDIT: On second thought, the routine is working pretty good now that I've accelerated the part of the sprite command that buffers the sprite. With that, I found an almost instant speed boost.

[Iambian]

For this update, just about the only thing I did was copy the sprite buffering routine to RAM so that it can work independent of any page changes, and "fix" the horizontal sprite flipping routine.

I found the sprite buffering routine to be horribly inefficient when I rediscovered the contents of the routines I was using to do the job. Hopefully, there's a speed gain, even when copying your standard 8*8 pixel sprite in.

I've also included and fixed the sprite flipping routine so that it actually works. I've verified it against an 8*8 pixel sprite (1 byte wide), but it should theoretically work on larger sprites (I haven't tested it yet).

Bugs still exist, mainly... I haven't bothered fixing the line drawing routine to function, if it ever did. There's also a bug that prevents destructible object from being destroyed (though you can still walk through them) from the "Metroid II: The Last Chozo Expansion Set" provided by DJ Omnimaga. I believe this is a Celtic III bug since it worked fine with xLIB, though I haven't exactly tested on whether or not it was because CIII-style errors were being put through regardless of the xLIB-compatibility settings or what. Maybe something weird is happening with the Ans variable, but I'm not entirely sure. Regardless of what the cause is, there's a problem that needs to be addressed and I'm not exactly sure how to address it other than to fix the problem I just mentioned and cross my fingers.

The plan to special-case the sprite routine also fell through for the moment. That's not included.

So, in all, compatibility with xLIB is improving more and more.

I do see light at the end of this tunnel...

[vuurrobin]

still a bad signature error

[Iambian]

Yeah. Bad signatures.
I'll have that solved when I'm ready to release a "final". Since it's still in the development stage, I'd prefer to keep things sitting nice in an emulator, such as PTI. Yeah, you heard me right. I don't think my application is ready for the world of actual calculator usage yet.

On another topic, I'm still tweaked over Metroid II not working correctly under Celtic III. I mean, the game is completely playable, but some of the nice effects aren't there, and I'd like to know why before I pull out any more of my hair. (the hair, by the way, is already thinning, so *that* might not be the best idea...)

EDIT: I've gotten an effects pack that revolves around xLIB's abilities, and Celtic III fails miserably at simulating xLIB. Seems I have more work to do about it.

[elfprince13]

is it just the speed of execution?

[nitacku]

[tr1p1ea]

[DJ Omnimaga]

it may be a challenge to have Nitacku's program work perfectly in Celtic, as most effects in it were created by abusing some xLIB exploits, bugs and glitches. Sometimes those are the 3 hardest things to recreate by ourselves, as normally we are used to make sure there aren't any.

I'M still unsure about what's wrong with Metroid II though. Is the version of Celtic on UTI the one you're testing Metroid II with atm? I may have to take a look at it to see what you exactly mean. I think I have a savefile on my PC where I was near the secret boss (which was only in the Expansion Set) so I can go everywhere in the game and check everything.

Also make sure you keep that copy of Metroid backed up on your computer because it may be more challenging to find it online than making sure Celtic works exactly like xLIB now.

[elfprince13]

i have metroid on my laptop if you accidentally delete it and need a new testing copy.

[tr1p1ea]

Kev and I were talking about this bug in Metroid and it might have something to do with your createPIC function? Are you copying from the buffer or the LCD? Does it overwrite an existing PIC? Could possibly be the recallPIC function as well.

[Iambian]

I've an update almost in response to the bug report posted in the ...uh... "appropriate thread".

I'm not entirely sure how this will perform, but I'm sure it'll get somewhere.

Details about what I changed are in ... one of the other pinned topics here.

Oh, and just FYI, I've also updated Celtic III to read off the revision codes similar to what you'd see on the filename of the .zip file. If you can tell, the title screen's version line for this will say "Flashapp version 3.09(K)" instead of the promised 3.10, which I've been holding off of actually giving until I can get xLIB support working just about as good as I think it should, or until I need to figure out what comes after "Z".

EDIT: I forgot to mention that the command pair that does indexed file data lookup remains untested, so if you choose to test it, ensure that the system you perform the test on is backed up sufficient enough to survive a RAM reset (or worse, but that's unlikely)

[vuurrobin]

well, the create pic function (real(9,...) works so it creates the correct pic and stores the content from the graphbuffer into the pic correctly. I don't know if its exactly like xlib, since xlib itself seems to fail at creating a pic correctly (it creates a completely white pic) (yes I downloaded the absolutely latest version from ticalc.org).

I also don't have the metroid expansion pack (or that effect pack, for that matters) so I can't see if the bugs are gone or not.


good job :lol:


(late) edit: is 3.10 not the same as 3.1 :confused:

[Iambian]

Quick update. I'd elaborate, but my battery's dying on my lappy, so I'll cut right to it.

I tried to fix the indexed file lookup routines. I updated the sprite routine to support inverted sprites as well, embedded as part of the xLIB commandset. I've tried to make the CreatePic routine behave more like xLIB, where it would update the screen as well as store the image to the file. I don't know whether or not the screen is read BEFORE or AFTER the update, but I also added in a function where it doesn't update the screen if you wanted to store it.

Read the ReadME for more about that.

Anyway...

EDIT: The indexed line routines don't work out very well. Gimmie a little longer to reupload the file. Since people have already downloaded this, I'll have to upload with an upped version code.

[tifreak8x]

Hmm... Was given an 8xg file (pokemon) to look at, though it needed xlib, so I sent celtic instead, and seems to be displaying ok :)

I can set the 8xg up on my server if you want it to test.. it even has animated tiles.. and such..

[Iambian]

A small update. I'm not entirely sure what I've changed in here, but I know I've fixed a lot of stuff, added a lot of stuff, and (hopefully not) broken a few things, too.

From the most recent changes (dating at the most 1 hour old), I added support for Omnicalc's real(20) [sprite] and real(33) [ExecAsm], though none will output correctly (that is, won't output the 1 or zero as a result of running the function). These functions are not in the documentation.

I also allowed the compression options for images in the image database to work, but since I changed the format of the database, I don't know whether or not it'll work at all. Attempting this on previous versions will not work. This isn't documented, either.

Some problems were not corrected. Time for everyone to guess which they were.

I is sleepy. I requires sleep. (improper English used for amusement)

[vuurrobin]

from what I heard, omnicalcs sprite function sucks (compared to xlib), and I thought that executing hex code was already inplemented :confused: . seeing that you don't have that much memory left, I think it would be best when adding a new function to think if it would be usefull for basic programmers, whether basic can do itself fine, and if the function isn't already added in some way.


nice to hear that compressing images is added . that will be really usefull when a project has a lot of pictures.

[DJ Omnimaga]

[TheStorm]

lol nice DJ

[vuurrobin]

ah, now I get it.