sQLux v1.0 (The Xmas Turkey)

[Exile]

I just miss my keypad under Windows. Any chance?

I've just pushed a change to master that enables use of the keypad. I've tested it with a GB keyboard, but fingers crossed it should work for all keyboard types.

[ql_freak]

Where can I find "german_bin"? as KBD = DE in the *.ini-file does not work on my normal notebook. On the GPD it's no problem, it has an US keyboard.

Minerva German_bin etc.

Thank you Ralf. I have LRESPRed it in sQLux with the "Minerva_1.98a1.bin" ROM, but the keyboard is still EN (don't know if US or UK/GB). When I use "Minerva_1g98a1.bin" I have a (nearly) German keyboard, x/y okay, umlauts okay, but other chars like _, ?, ... not. But then all messages are in German (aargghhh).

[XorA]

Where can I find "german_bin"? as KBD = DE in the *.ini-file does not work on my normal notebook. On the GPD it's no problem, it has an US keyboard.

Minerva German_bin etc.

Thank you Ralf. I have LRESPRed it in sQLux with the "Minerva_1.98a1.bin" ROM, but the keyboard is still EN (don't know if US or UK/GB). When I use "Minerva_1g98a1.bin" I have a (nearly) German keyboard, x/y okay, umlauts okay, but other chars like _, ?, ... not. But then all messages are in German (aargghhh).

You will need to have patience, tofro updated the keyboard code from his original. But it is still in review.

[XorA]

...
BTW: sQLux seems to have the same problem as UQLX (with WSL2 on Windows 10):

You can't create directories with MAKE_DIR.

You get bad parameter and afterwards a file exists with the name of the directory you wanted to create...

With this do you mean at all?

Or with qdos-like drives?

I'll note the MAKE_DIR function in uqlx source is empty so I'm guessing never finished!

I'm using only "qdos-fs", i.e. *.win-files with UQLX (and currently with sQLux), cause with UQLX on WSL (don't know if this has changed with WSL2) it is/was not possible, to access files on NTFS formatted partitions (I have no more FAT partitions since Windows 10).

I have attached my ".uqlxrc" file from UQLX, where you can find a lot of comments, what I have detected on problems with UQLX on Windows 10 with WSL (and now WSL2, but still using Xming, not the inbuild X-capabilities). Some of the problems (open overwrite, rename) seems to have been corrected. I just want to test overwrite and rename on UQLX and found following strange behaviour:

Code: Select all

DATA_USE'RAM1_sub1':OPEN_NEW#4,'test_txt':PRINT#4,'Test file without terminating LF';:CLOSE#4

If you type WSTAT in SB you see the file "sub1_test_txt", length 32 bytes (file type is 0).

Now open Files of Qpac2 and switch to RAM1_.

RAM1_ is empty. After further investigation I have found, that even if you copy files with Qpac2 Files to RAM1_ Files will always show an empty directory on RAM1_ while DIR, WSTAT, ... will show the files.

On QPC II all is fine.

So there is an obvious reason this doesnt work in the uqlx codebase

Code: Select all

	case 77:
		*reg = -15;
		break;

Im going to take a guess that -15 is BAD_PARAMETER.

So it seems to be implemented for qdos-like/ directories but not .win/.img files.

[ql_freak]

3) if old-fashioned text user interface job is active, there is the text cursor active, and no way to move the mouse. I have to switch jobs to an extended environment one to invoke the mouse cursor.
I thought that this is normal behavior.

Wrong.

It has even been possible to raise the mouse cursor from SuperBASIC with a "keyboard" mouse (i have had built the interface from [Robert? Bessler] myself, was my first mouse). I've wrote a driver (or better an interrupt routine in the 50 Hz loop) which sends an interrupt, so that the mouse cursor appears. I've tried it on UQLX and sQLux, but it does not work (brings the mouse up), it seems both have no keyboard, but a real mouse. (The program is named "click_cde" or "clickQIMII_cde". I think I have given it to a few people, so if you find it, you may try - on sQLux it makes a lot of noise, if you are typing).

p.s.: The QIMULATE command, starts a "program" which set's the mouse hotkey (MOUSE_STUFF), which IMHO was not available in Extended Environment without a QIMI installed. It seems the mouse hotkey is hard coded to:

<ALT>+<.>

It works here on uQLx on Windows (WSL2 with Xming) but unfortunately not on sQLux. On uQLx I have had defined:

ERT hot_wake('.', 'Wake')

If I now press '.', Wake will be started. But with QIMULATE you can use the middle mouse button (left+rigtht together [as on my BBQL} does unfortunately not work) to wake Wake.

p.p.s.: Source Code: It is the one program where I cannot find (even 30 years before!) the source code anymore.

[ql_freak]

prt_gen v2.01 is the last version that produces a K cursor for me.

Thanks. I have tested it now myself. With Ptr_gen 2.01 the cursor comes up if you move the mouse (EVEN IN NONE-POINTER PROGRAMS), with higher versions (I have tested 2.05 and 2.02) NOT. I haven't had time to test it on UQLX (on Windows Subsystem for Linux 2 [WSL2]) unfortunately.

That's BAD :‑( This is, because 2.01 (and I think also 2.02) does not yet have the PE_BGON command (when you have entered this, windows in the background are not burried [suspended]). So if you want real multitasking even with pointer environment, you must LRESPR pie or pext – and I don't know, if any of these extensions will work on UQLX or sQLux(!). I always have only used pie (it was the first - you need at least Minerva, so that this will work), but I have now found a side from h.p. for pext (inspired by pie):

https://phpr.tripod.com/index-ql.html

EDIT: Look for "Hidden Window Refresh" or so.

wman version seems to be irrelevant for that!

Not here(!): When I tried Ptr_gen 2.01 with the current wman I've got error "Too old" at boot.

[RalfR]

Thanks. I have tested it now myself. With Ptr_gen 2.01 the cursor comes up if you move the mouse (EVEN IN NONE-POINTER PROGRAMS), with higher versions (I have tested 2.05 and 2.02) NOT.

That's consistent with my experience. According to Update_txt, PG_BGON/OFF doesn't work under QDOS (according to Marcel).

Now I wanted to try WMOV under sQLux, but even with the current PE versions, I only get an ERR.NI in #0 when I LRESPR ptrmenr_cde. It's all pretty strange.

[XorA]

Unfortunately the ptr_gen discussion is a bit circular unless someone knows what changed on the ptr_gen side.

On the MAKE_DIR discussion I actually have the code to MAKE_DIR work, compared it with the results of the same command on QPC2 and I get byte for byte identical results. If someone wants to test I can upload a branch to either uqlx or sQLux repos with the change.

[ql_freak]

That's consistent with my experience. According to Update_txt, PG_BGON/OFF doesn't work under QDOS (according to Marcel).

Then Marcel is wrong, from my Std-Bootfile for uQLx and sQLux:

Code: Select all

2000 :
2010 REMark now let them all go:
2020 HOT_GO
2030 PE_BGON: REMark Allow output to windows (programs) while they are (partly) overlayed (in background) from other windows/programs. You can switch it off with PE_BGOFF
2100 :

Of course Minerva is required, as it's screen/console routines supported SD.LINEL (which is documented in Technical Guide, 18.9, page 169) from the beginning. The Sinclair ROMs seem to have SD.LINEL hard coded (not checking/using SD.LINEL in the Screen Driver Data Block Definition).

In fact there were Tools for Minerva which allowed output to background windows long before SMS supported it.

[RalfR]

In fact there were Tools for Minerva which allowed output to background windows long before SMS supported it.

Please do not confuse this, continuing to write in buried windows is only a matter of the pointer interface.