Amiblitz3 new version

For people who can interest, i've seen, and downloaded, the new version of Amiblitz. The last one was the 3.12. The new one is 3.33 and published on 29 April.

Have fun and thanks to their developpers.

Nic

@unimon

a link pls!! and also a changelog would be great..

@afxgroup

http://amiblitz3.amiforce.de/index.php/downloads/category/1-.html

changelog wasnt published up till now afaik

@wawa

New Features:
- added source code of Amiblitz3/PED/Debugger and available BlitzLibs/Residents and Tools
- hosted on SVN at Sourceforge (http://sourceforge.net/projects/amiblitz3)
- auto layout of IDE windows
- added .ab2 format support for included files
- added CLI Mode (non-GUI), not finished yet

Bugfixes/Improvements:
- cleanup of directory structure of distribution
- OS4 Debugger fix
- huge speedup (~30x) of loading/saving .ab2 files
- PED refresh optimizations
- GUI beautifications, e.g. source-tabs, real italic/bold fonts
- dynamic Blitz3 Assign supporting multiple instances of Amiblitz3
- made version string RKM compliant
- more security checks for libraries
- improved english texts and compiler messages
- localization of most messages (no Wizard windows yet)
- internal code beautifications, e.g. Common Data is defined in the same file, more plain text labels/constants etc.
- fixed Workbench color-trash bug
- ... and as usual, lots of minor bugfixes

@Wanderer

Not to impose on your heavy workload, and the fact that you have some nice gear out there,

was wondering how the RPG is coming along?

and the music program?

Cheers.

@Gebrochen

RPG is progressing every now and then, here are some experiments with alpha blending of textures:



HD-Rec is currently slowed down, since there are lots of other things to do. But I guess I should make a 24bit release, you are right. And then: it will be free and open source, as soon as i have converted all Amiblitz3 Inlcudes to .ab2 files (they are plain ASCII and work well wth SVN, unlike .bb2 files).

@Wanderer

Awesome GFX :-O

Looks great :) and keep up with the HD-REC 24bit.

@unimon

#1
Doesn't work here on the A1-XE. Freezes on any IO input. Moved it to EUAE and get file name errors. A very impressives looking program.

@Snuffy

What kind of file name errors?

It is supposed to run under OS4, only the Debugger is currently broken. If it makes problems, please report.

You say you can start the IDE, but once it is open and you click the mouse somewere it freezes?
Sounds like a stormwizzard problem. Double check the libraries in Libs/ with the ones you have in Libs:

@Wanderer



What kind of file name errors?
Don't know. I tried this out:

Copied AB3 to EUAE (0.8.26)
1. Icon Start in AIABr10.6/ScalosWB Applications
2. Chose in PED the source file - dbl_Simple_01.bb2
3.Chose Compile & Run
4. Output Pass-0 compile window; Output Pass-1 window; build...run...
5. Graphics output was dead slow of 3 balls falling and bouncing. Even in 8bit mode.
6. Way too slow ; press ESC key
7. Returned to PED
8. Quit
AB3 worked correctly, but exceedingly slow in EUAE.
------

It is supposed to run under OS4, only the Debugger is currently broken. If it makes problems, please report.

My AB3 results in OS4.1 (A1-XE/G4/800Mz/256MBram:Radeon 9250/1152x854/32bit):
1. Icon Start
2. Chose in PED the source file - dbl_Simple_01.bb2; loaded nicely; appears to a prg. about 3 bouncing balls.
3.Chose Compile & Run
4. Output Pass-0 compile window; Output Pass-1 window
5. GrimReaper: Amiblitz3 (Crashed)
Stack: 0x6b82f004 - 0x6b838c3c, pointer @ 0x6b838be6 (Cookie Damaged)
Signals: SigRec 0x00000010, SigWait 0x00000100
6. Saved Crashlog
7. Tried to close PED window or any that would.
8. Tried "Continue Program" in GR window
9. Graphics output was three gray/silver bouncing balls! OK...
10. Watched awhile and then clicked LHB mouse.
11.PED output window:
"Program terminated.
Press <enter> to return to Amiblitz3..."
12. A small 'request' window was in the middle of PED output window:
"This action is not allowed while executing a program.
Should PED do a PANIC program kill?"
13. Chose 'Yes', and AB3 closed .
-----------------------
Other source files tried:
3. dbl_Simple_04_100Hz.bb2 OK
4. dbl_Simple_07_AHI.bb2 Error - No sound
"Compile error in line 4 ... can't open include file!"
5. dbl_Simple_10_control.bb2 GrimReaper - No recovery
Mouse controls & keyboard appeared to work fine.
6. dbl_Simple_13_fade.bb2 No error, but to slow.
7. dbl_Fireworks.bb2 Nothing
Asked for #particles and quit to PED.
8. dbl_Movingwaters.bb2 GrimReaper - Recovered. No gfx
"Compile Error in line 37: Varibales/constants must be declared...
-------
You say you can start the IDE, but once it is open and you click the mouse somewere it freezes?
It seems different source programs have different results.

Sounds like a stormwizzard problem. Double check the libraries in Libs/ with the ones you have in Libs:
/Media/AB3_Release330/Libs> ls
Ted.library v1.0 id3tag.library v5.6 ptplay.library v2.5
blademp3.library v1.0 jpeg.library v6.2 wizard.library v51.3
disassembler.library v40.5 lh.library v1.8 zlib.library v3.2

@Snuffy
afaik euae has no jit. all other amiga 68k apps should be slow on your system as well whatever it is.

as for the rest from my experience panic program kill requester appears when you try to interact with ab3 while it is executing compiled app. that is all right.

as for grs you get, i cant tell anything, maybe i can fire up os4 on my classic later on to check if i can confirm your problems with current version of ab3.

@Snuffy

So it seems it works correctly on EUAE, just slow, but I guess that holds for all software.

The bouncing ball demos however can be tweaked to be much faster (using selective refresh), but the way it is implemented is to satisfy GFX cards running at their real speed. It runns fine on e.g. 68060 with Cybervision 64 3D. The demos are not suppoosed to run under lower specs. For lower specs => use BlitzBasic2 instructions.


Quote:
hm... would need further investigation.


Quote:
Doing a PANIC kill is not the right thing to do. Amiblitz3 asks you this if you try to edit something in PED while the program is still running. This is because the program could be stuck in a deadlock and you even can't save your work. But during regular usage, this should be no option, only if you are in PANIC!

I corrected the other thing in sourcecode.
(it's damn hard to keep all the 1000th of files up to date!)

> It seems different source programs have different results.
Well, that's what different source programs are supposed to do...

Actually I see only one bug, the Grim Reaper on some progs.

When asking for the libs, I wanted to see the libs in LIBS:, because they are preferred over the local ones .
Copying them accross to LIBS: might help.

It is correct that the simple demo give a reaper hit, while the 100Hz demo runns smoothly without any trouble?
That would give us a hint what might be wrong, since the two progs are almost identical.

@Snuffy

Quote:

You're stack is damaged, which implies an overflow at some stage. Although the pointer seems to be inbounds at the point of crash. Up the stack and see if you still get errors. Don't waste time fiddling with it, bump it to 1000000 to be almost 100% sure of avoiding any overflow. You can work out what it actually needs later.

Currently your stack frame is only 39992 bytes. That's not very big by OS4 standards. If that's your default stack you must have a system on verge of crashing all the time....

@broadblues

No, i just saw that this is the Stack set in Amiblitz3 Icon. That used to be 500,000 to be on the safe side. Someone must have lowered this. I will increase it in the next release. However, those 40K are enough on Classic, but might not be engough on OS4 or MOS.

@Wanderer

Yes I would say 40K is too small for OS4. I just tried the archive as described by Snuffy with no crash. My global stack is 72K. This will have overidden the 40k in the icon. So it need more stack but not a lot more

BTW there is a slight glitch in the editor, any bold text in the syntax highlighting is typically mising the last character. Not quite enough room allowed to render it? Perhaps a sublte diiference in font handling with OS4

@broadblues

Last whole character or last pixel column? (as it is in italic fonts)

@Wanderer

Quote:

Somewhere inbetween I'd say

Here's a grab



It's the blue bold text that's wrong the green italic is okay

What you can't see in the grab is that the bold text is slightly out of sync with the cursor. ie it's not the same width as the plain.

@Wanderer

Quote:

Typically you double the stack size on the PPC. So if 68k needs 40K stack you use 80K stack on the PPC. Ambient takes care of it automatically (it doubles the stack size).

If 500k stack is needed then there is a design problem or some other task is trashing your stack.

I've seen this glitch too - is it a case of the character widths being miscalculated, maybe only for bold? By using courier fixed or a similar font the tokens are rendered correctly (but it's an ugly font!) I'll check when I get home which font I was using in the earlier version, but it didn't have this problem...

Good work though, each update brings along some nice new features :)