Qlib v3.45 error under QDOS

[BSJR]

I compiled a new version of BMP with the latest Qlib v3.45 on QPC2v5 with SMSQ/E v3.38.
As it was also meant to run under QDOS I did some tests there and ran into a problem.
The Q_ERR_ON keyword produced a "bad name" error, even when the loaded Qlib_run was also the latest version.

Testing the same _obj under Minerva went without hiccups.
A Qlib v3.36 compiled version also worked fine under QDOS, with both the old and the new Qlib_run.

Could there be a problem with the updated Qlib_ext - which adds the Q_ERR_XX keywords to the _obj - and QDOS?
Maybe it expects a 680x0 CPU and not a 68008?

BSJR

[dilwyn]

In QLIB 3.45 the Q_ERR_ON/OFF keywords are now in an extensions file called QERR_bin, linked in with $$asmb=win1_QERR_bin,0,10
in case that's the issue you're having.

[pjw]

In QLIB 3.45 the Q_ERR_ON/OFF keywords are now in an extensions file called QERR_bin, linked in with $$asmb=win1_QERR_bin,0,10
in case that's the issue you're having.

Did that fix it, Bob?

[BSJR]

In QLIB 3.45 the Q_ERR_ON/OFF keywords are now in an extensions file called QERR_bin, linked in with $$asmb=win1_QERR_bin,0,10
in case that's the issue you're having.

OK. thanks.
Oops, I never noticed the Q_ERR commands had jumped bins.
So I compiled a new built with the correct _bin.
The _obj still runs on SMSQ/E and Minerva but again fails on QDOS.
This time no error message but it freezes the machine only at a later point.
An RJOB is needed but as everything has come to a snails pace, it's hard to get there.

For the SMSQ/E & Minerva tests Qlib_sys was loaded while the QDOS machine had Qlib_run.
But Booting with the _sys there too made no difference.

Back to the drawing board.

[dilwyn]

The only QDOS setup I have here to test is QemuLator with JS ROM.

A very quick test with this shows that Q_ERR_ON seems to work on that system at least, both interpreted and compiled:

Code: Select all

100 REMark $$asmb=win1_bootup_qerr_bin,0,10
110 CLS : CLS #0
120 DATA_USE '' : REMark prevent DIR using data default drive
130 INPUT #0,'DIR drive > ';ip$
140 IF ip$ = '' THEN STOP
150 Q_ERR_ON 'dir'
160 DIR ip$
170 fer = Q_ERR
180 Q_ERR_OFF 'dir'
190 IF fer <> 0 THEN PRINT #0,'Oops, error ';fer
200 INPUT #0,'Finished, press ENTER > ';ip$

Obviously, it's only a quick test on a QDOS emulator, not a real BBQL.

[BSJR]

The only QDOS setup I have here to test is QemuLator with JS ROM.

A very quick test with this shows that Q_ERR_ON seems to work on that system at least, both interpreted and compiled:

Obviously, it's only a quick test on a QDOS emulator, not a real BBQL.

My compiled test was done on a BBQL with SGC and Marcel's Minerva/QL-SD (v1.07).
The compiled JS-QDOS test was done on an Aurora with SGC and Qubide v2.
After the fix with Qerr_bin there has been no indication why this freeze occurs but it's probably not related to Q_ERR_XX anymore.
At the moment I can only do an S*Basic test on QPC2/SMSQ/E.

BMP tested well with some Bmp and Pic files but stops working when processing a Scr file.
At one point it was not just a slow down but a complete stand still, a hard reset was needed.
Why this happens and only on the JS system is the big question.

We keep on digging deeper,
BSJR

[desin]

The QBasic Thing works only under SMSQE

Qdos needs
Loading the program
Liberate dev_program
ex QLib dev_program
Compile

Greetings from Switzerland
Markus

[Derek_Stewart]

Hi,

I stopped using TK2 v232, 236, there are issues, tried with sQLux, Q-emulator, starting up with no boot file in monitor mode, changing to TV mode with WTV does not clear the border. When WMAN is loaded, entering WTV or WMON works correctly. So back to TK2 v2.23 as I sometime like not install the Pointer Environment or Extended Environment.

Another issue is I have setup Minerva v1.98, 896K ram

I have a Double Density FLP image with QLC compiler contained in the image.
I used use Tooklit 2 v2.12 in ROMPORT, which the DIR header of FLP image gives:

QLC
822/1440 sectors

All work correct. But use TK2 2.32 in ROMPORT, gives DIR header:

/home/derek/QLC.img
2466/4460 sectors

Why does Tooklt 2 v2.32 give 3x the sector count?

Is there a problem with Toolkit 2 v2.32?