SpartaDOS X v4.22

[MrFish]

Is there any documentation, or at least general information, regarding the differences (from a user standpoint or otherwise) between v4.22 (FTe) and v4.21 (ICD)?

 

Also, were there any bugfixes/optimizations?

 

The only information I was able to dig up so far is this quote (translated from Polish) from someone who goes by the name of "Pecus" on Atari.Area forums (posted in 2003):

"...In version 4.22, support for 65c816 has been added (part of internal procedures can use it, the system itself detects and sets up an additional presence flag) and further optimizations are made, again without changing the internal rules of system procedures."

 

[MrFish]

 

1 hour ago, Kyle22 said:

Sorry, I didn't see the request, but you have it now.

 

Nice, thanks!

 

But the log file only covers changes through version 4.21.

 

SpartaDOS X update log
----------------------

Date	  Fixed	Version	Description of bug or Feature added
----	  -----	-------	---------------------------------------------------
 9-12-88   -	4.17	First release of SpartaDOS X
10-29-88  Yes	4.18	XIO 49 does not work in LIB_CIO ver of CIO handler
          ?		ARC crashes when no filename (only extension)
          No		FIND CAR:*.* shows matches as D33:fname.ext
10-29-88  Yes	4.18	FPUTS terminates on 0 - is a problem with ACTION
10-18-88  Yes	4.18	Made the cart latch a CONSTANT in KERNAL.
10-29-88  Yes	4.18	Does not work on 1200XL (probably $c4c9 OS call)
10-26-88   -	4.18	XEP80.SYS driver written (v0.1)
10-29-88  Yes	4.18	JIFFY.SYS crashes on year 2000
10-29-88  Yes	4.18	ARC crashes with 0 length files (in verbose)
10-29-88  Yes	4.18	ARC does funny stuff if no date on orig file
10-29-88   -	4.18	Expanded CAR: by adding end of BANK 0 to regions
10-29-88   -	4.18	Changed 'FATAL ERROR... CALL ICD' message to
			  'DISK CORRUPTED: RUN CLEANUP'
 1-06-89  Yes	4.19	ATARIDOS.SYS driver is slow because of positioning
 1-06-89  Yes	4.19	Append on AtariDOS diskettes gives range error
 1-05-89  Yes	4.19	TD ON/OFF OS vector crashes (call to OS routine)
	  No		File length in ARC wrong on AtariDOS diskettes
 1-16-89  Yes	4.19	Delete of protected files in MENU fails
 1-06-89  Yes	4.19	If first batch files calls 2nd non-existent batch file, system crashes
 2-02-89  Yes	4.20	No EOF returned on directory listing if SP<$80
 2-05-89  Yes	4.20	PUT through CIO puts extra EOL char (in LIB_CIO)
 2-16-89  Yes	4.20	Now works with AW+ (CP not set in X.COM)
 2-16-89  Yes	4.20	DUMP restores LMARGIN correctly
 2-16-89  Yes	4.20	FORMAT sets and restores CHBASE also
 2-16-89  Yes	4.20	Binary files of 0 length will not crash anymore
 6-08-89  Yes	4.21	High byte of sector count was bad on AtariDOS
			  diskettes using ATARIDOS.SYS.
 7-10-89  Yes	4.21	Closing a directory that failed open crashes
	  No		Filesize in ATARIDOS.SYS uses 128/256 b/s factor
			  rather than 125/253 - probably will not change
			  due to large cost in code and it appears to not
			  effect operation.


Date	  Done	Version	Features desired
----	  -----	-------	---------------------------------------------------
 1-16-89  Yes	4.19	Batch files using the PATH varaible
 1-16-89  Yes	4.19	X command using PATH variable
	  No		Add a NUL: device
 1-16-89  Yes	4.19	Drive letter in MENU
	  No		Make batch files end on EOF (no return added?)
 1-20-89  Yes	4.20	High Speed support for XF551 3/12" drives

 

Edited April 19, 2020 by MrFish

[Kyle22]

Sorry, that's all I have. I really miss the source to the 3.x versions and my never released 3.23 which had useful utilities in CAR: I lost my TeleVideo 816 system and my backup tapes.

 

3.23 was small cleanups to the code and enhancements to the INDUS.SYS Z80 code for track buffering SuperSync. That was my main goal. It was a small bug in the timing.

 

 

Unless those tapes are found, this is very unlikely to be recovered. I think it may have made it's way into the warehouse. AFAIK, the contents of the warehouse were sold at auction.

 

 

[MrFish]

No big deal. I was just going to put this version up on my site and wanted to have a little more information about it if possible.

 

It does seem to detect an 816 when present in Altirra. So, I'll just assume that to be the main change -- and take the information on Atari.Area into account. If there was anything else, it was probably just small bug fixes, since the version number only changed by a small amount anyway.

 

So, I take it you didn't work on 4.22 yourself.

 

[_The Doctor__]

Some branding was changed,

4.23 support was added for 65c816 and some of the internal procedures changed to use it, but not all. No 816 High Ram

It detects and sets up a flag so that any other flag aware modules can run 65c816 optimized code, without the flag set it is business as usual.

A number of touch up and fixes were done afaik.

I was not in house, merely a consultant. Lots of phone time. I can guarantee a sh*t ton of FTe's costs were phone bills...

4.23 had indus fixes, ask kyle

kyle on the other hand was probably tired of dealing with MIO's at the time... 😮

[MrFish]

Alright, thanks. It's worth adding to my site just for the 816 changes, even if partial.

 

[drac030]

It as been quite long time ago, but I do not remember any particular 65C816 support in the 4.22 code. Maybe it was added later, to the version which did not manage to get sold.

 

This particular message, "Sweet 16 not Detected", is displayed by RAMDISK.SYS. This was about the only program which did contain some 65C816 support, i.e. it detected the CPU the usual way then, IIRC, used the move block instructions to copy the data to and fro, with interrupts disabled.

 

[drac030]

36 minutes ago, drac030 said:

This was about the only program which did contain some 65C816 support, i.e. it detected the CPU the usual way then, IIRC, used the move block instructions to copy the data to and fro, with interrupts disabled.

I actually checked this... and it is not even doing that. It makes CPU detection just to display the message, there are no other differences (when you fool the test on vanilla 6502, the ramdisk still works without problems).

[_The Doctor__]

yes, much like the sdx project is now, it was a work in progress... but I'm talking about 4.23 , crap... I did type 4.22 at the top and end with 4.23 at the bottom. changes occurred in between.  F*ck it can be read as if I am referring just to 4.22 and not the revision going on towards 4.23. I need to edit/clarify that.

Edited April 19, 2020 by _The Doctor__

[MrFish]

16 minutes ago, _The Doctor__ said:

but I'm talking about 4.23

Does anybody still have a copy of the v4.23 cartridge, Doc?

 

Even if it was just minor bug fixes in v4.22, I'll still add it to my site, since it was the last commercial release before DLT started working on it.

 

[MrFish]

2 hours ago, drac030 said:

I actually checked this... and it is not even doing that. It makes CPU detection just to display the message, there are no other differences (when you fool the test on vanilla 6502, the ramdisk still works without problems).

Thanks for the info, I was hoping you might speak up here.

 

I was starting to examine the v4.22 source files by date last night, and noticed that very few sources had been changed after the 7/10/1989 release date for v4.21. So, I was starting to wonder about the possibility of any significant changes for 816 being present.

 

Edited April 19, 2020 by MrFish

[drac030]

The version we reverse engineered was the version we had here, i.e. 4.20. The differences to 4.22 (which was compared later) were rather scarce, and it seems they in fact were made in 4.21. I remember that the BMI branch at $0807 was one of them (which is mentioned in the log as being added in 4.21), there was also some difference in SPARTA.SYS, nature of which I do not remember at the moment.

 

EDIT: it is possible that the sources you have are sources of SpartaDOS X 4.21. See command.asm and its copyright notice, or try to find the message above in ramdisk.asm.

Edited April 19, 2020 by drac030

[MrFish]

18 minutes ago, drac030 said:

EDIT: it is possible that the sources you have are sources of SpartaDOS X 4.21. See command.asm and its copyright notice, or try to find the message above in ramdisk.asm.

Yeah, I was wondering that myself, and I can see now that the last date for changes here are 3/12/95 (file date stamps), compared to the v4.22 notice of 11/5/95.

 

I was searching for the "Sweet 16" message already, and it's not in there. So, this is likely some version in between v4.21 and v4.22, with few changes from v4.21.

 

[_The Doctor__]

The 2 people that would still have the image of 4.23 are Mike & Kyle.

[MrFish]

11 minutes ago, _The Doctor__ said:

The 2 people that would still have the image of 4.23 are Mike & Kyle.

Alright... that would be an interesting artifact...

 

[jaybird3rd]

2 hours ago, _The Doctor__ said:

yes, much like the sdx project is now, it was a work in progress... but I'm talking about 4.23 , crap... I did type 4.22 at the top and end with 4.23 at the bottom. changes occurred in between.  F*ck it can be read as if I am referring just to 4.22 and not the revision going on towards 4.23. I need to edit/clarify that.

Fixed!

[drac030]

From what has been written above I think 4.23 is already here, it is called 4.49, where many things were done which ICD apparently planned (even though we did not know that they were planning these):

 

	No		FIND CAR:*.* shows matches as D33:fname.ext
	No		Filesize in ATARIDOS.SYS uses 128/256 b/s factor
			rather than 125/253 - probably will not change
			due to large cost in code and it appears to not
			effect operation.
	No		Add a NUL: device
	No		Make batch files end on EOF (no return added?)

All these are "Yes" now (the last one was already done in 4.20 unless I am missing the point).

 

Besides, 65C816 support added, the flag which is to be used by programs wanting to run optimized code is there (COMTAB2+0), an editor has been added to CAR:, INDUSX.SYS provides track buffer on RAM-Charged Indus GT drives and so on.

[Kyle22]

4.23 was never released. I don't think Mike ever had a copy. It was just something I was tinkering around with. I'm afraid it's long gone.

 

4.22 was the last FTe release.

 

[_The Doctor__]

Now I am confused because during discussion, it was supposed to be revised for that other big project and support it fully... As I understood it.

But the it doesn't matter as it's all in there now anyway.  drac030 had the list and checked it off. looks like the information was in many places and was rounded up.

 

Kyle, It looks like a version conflict that would have rolled into 4.23.... should have been 4.22i for indus... but it's not going to matter as it's lost on some old tape or hard drive in the move from California and the cartridge was lost during successive moves.

 

Edited April 19, 2020 by _The Doctor__

[MrFish]

3 hours ago, drac030 said:

From what has been written above I think 4.23 is already here, it is called 4.49...

Yes, of course, no loss in terms of planned functionality. I was just concerned with it as an artifact, as I mentioned above. As Kyle said, it doesn't exist anyway.

 

Moving on from there, does anyone have a copy of the initial release, SDX v4.17 ? I've heard it talked about before, and see it in the log notes (above); but no copy has surface yet, that I know about. From what I understand, most people who had earlier versions, like 4.17, were able to upgrade them by sending in to ICD; subsequently it would make these earlier versions more difficult to come across.