Minerva4Q68 v1.70 released

[Peter]

Hi Jan,

Thanks to Mark Swift, I fixed another long-standing bug that affects 68020+ processors. On these processors, the SSP would get a wrong value of $28482 instead of $28480 because the RTE instruction works differently. I'm not able to test this because I have no 68020+ equipment to test, but I'm curious about the result.

Can this fix be tested on a SuperGoldCard or is it patched there by the boot ROM?
Every improvement toward 68020+ is good news, maybe the QL60 can benefit.

I've also succeeded in extracting the Q68 SD card driver from the SMSQ/E source, but so far I haven't been able to get it working on Minerva.

This sounds very difficult. Wolfgang must have had similar issues when he originally did that. A somewhat easier approach might be to use the QL-SD driver and replace only the low level routines.

Thanks for your impressive work!
Peter

[janbredenbeek]

Hi Jan,

Thanks to Mark Swift, I fixed another long-standing bug that affects 68020+ processors. On these processors, the SSP would get a wrong value of $28482 instead of $28480 because the RTE instruction works differently. I'm not able to test this because I have no 68020+ equipment to test, but I'm curious about the result.

Can this fix be tested on a SuperGoldCard or is it patched there by the boot ROM?
Every improvement toward 68020+ is good news, maybe the QL60 can benefit.

Mark has reported this, see https://github.com/janbredenbeek/Minerva4Q68/issues/2. AFAIK it's not fixed by the SGC ROM.

I've also succeeded in extracting the Q68 SD card driver from the SMSQ/E source, but so far I haven't been able to get it working on Minerva.

This sounds very difficult. Wolfgang must have had similar issues when he originally did that. A somewhat easier approach might be to use the QL-SD driver and replace only the low level routines.

The SMSQ/E driver relies on the THING system, which is, by default, not present in Minerva. It might also rely on other SMSQ/E-specific features.
Porting the QL-SD driver might be an option; it contains some BASIC commands like WIN_DRIVE, WIN_USE and WIN_CHECK which are not present in the current driver. On the other hand, CARD_INIT is missing which is something of a surprise as QL-SD supported two SD drives like the Q68?

[ql_freak]

Hi Jan,

I have backported the HISTORY device from SMSQ/E and implemented it in Minerva's command line editor.

Arrggghhhh... Why did you do that!

The history device has been written by Boris Jakubith from Berlin (a friend of mine) and it's on Dilwyns page (search for history).

This version is IMHO better than the reimplemented version in SMSQ/E. There is also an explanation, how to use it from Assembler (also I have not yet tried it/experimented with it).

So if you are using anything other than SMSQ/E, I recommend to use Boris history device, it's the original. I even will try to find a (simple) way to remove the history device from SMSQ/E and reload the real one.

Or are there any advantages in your version?

Cheers Peter

[BSJR]

...
Porting the QL-SD driver might be an option; it contains some BASIC commands like WIN_DRIVE, WIN_USE and WIN_CHECK which are not present in the current driver. On the other hand, CARD_INIT is missing which is something of a surprise as QL-SD supported two SD drives like the Q68?

As QL-SD was also missing some of the useful Q68 commands, M.Kilgus has made a separate BIN that contains more of them: qlsd_utils101.zip with CARD_RENF, CARD_CREATE & CARD_DIR$.
https://www.kilgus.net/ql/ql-sd/

BSJR

[tofro]

Hi Jan,

I have backported the HISTORY device from SMSQ/E and implemented it in Minerva's command line editor.

Arrggghhhh... Why did you do that!

The history device has been written by Boris Jakubith from Berlin (a friend of mine) and it's on Dilwyns page (search for history).

This version is IMHO better than the reimplemented version in SMSQ/E. There is also an explanation, how to use it from Assembler (also I have not yet tried it/experimented with it).

So if you are using anything other than SMSQ/E, I recommend to use Boris history device, it's the original. I even will try to find a (simple) way to remove the history device from SMSQ/E and reload the real one.

Or are there any advantages in your version?

Cheers Peter

The "original" (at least in some versions) positions the cursor of a re-called line to the beginning of the line, which is a nuisance. And why exactly do you think "it's better"?

[RalfR]

The "original" (at least in some versions) positions the cursor of a re-called line to the beginning of the line, which is a nuisance. And why exactly do you think "it's better"?

I think he also means the advanced options of Boris' "History," which are not available under SMSQ/E.

The only thing that always bothered me about the "Berlin clique" was their sometimes nasty grumbling about SMSQ/E and PE in general. Hans-Peter and Boris's software was mostly quite usable.

[Popopo]

Nice job!
I have read other comments that already existed an alternative for it.

But it was never the reason to stop us. The joy of develop your OWN solution is so great!
So congratulations for this goal reached, and another little piece of art in our retro world.

Also cause you did it, now I know different options to use log commands
Just left a sys var that remember the last accessed unit

Thanks!

[janbredenbeek]

Hi Peter,

I have backported the HISTORY device from SMSQ/E and implemented it in Minerva's command line editor.

Arrggghhhh... Why did you do that!
The history device has been written by Boris Jakubith from Berlin (a friend of mine) and it's on Dilwyns page (search for history).

I'm aware of Boris's HISTORY driver. I have observed several problems with it, especially in conjunction with Qemulator and sQLux (to the point that it renders the system unusable). My guess is that these problems are caused by the fact that it nests itself deeply into the CON driver, which is already tinkered with by low-level emulation code and software like the Pointer Interface. It's hard to tell which is at fault since the source code isn't available.

Given that it's written 30 years ago, it would certainly have been great software at the time. However, the QL scene has evolved since, with new emulators and the source code of Minerva and SMSQ/E becoming available. This enabled me to implement command line history in the command interpreter itself, without having to tinker with device drivers and the potential problems this causes.

I have tested my version of Minerva on several platforms and haven't seen any problems yet - it's a clean implementation and uses code known to be stable. Also, the source code is available so if any problems emerge they can be tracked down and fixed by anyone with sufficient skills.

And last but not least, like Popopo said - being able to write your own software solutions is perhaps most of the fun of our hobby!

Jan

[janbredenbeek]

Hi Bob,

...
Porting the QL-SD driver might be an option; it contains some BASIC commands like WIN_DRIVE, WIN_USE and WIN_CHECK which are not present in the current driver. On the other hand, CARD_INIT is missing which is something of a surprise as QL-SD supported two SD drives like the Q68?

As QL-SD was also missing some of the useful Q68 commands, M.Kilgus has made a separate BIN that contains more of them: qlsd_utils101.zip with CARD_RENF, CARD_CREATE & CARD_DIR$.
https://www.kilgus.net/ql/ql-sd/

I've found some useful information in the QL-SD documentation. Also, the source code of the QL-SD driver is included in SMSQ/E.

Nevertheless, there is still a significant difference in the low-level routines between QL-SD and Q68, and I guess replacing them will not be straightforward. Also, I don't have the source code of the current driver.

I'll take a look at the QL-SD utilities - if they can be ported this would be a great enhancement to the current driver, which lacks most of the commands that SMSQ/E has.
(a FAT device would also be useful, but whether it's worth to sacrifice 13K of ROM space is a moot point...)