Nibl-e

[Phil__G]

I wonder, has anyone successfully assembled the NIBL-E.asm source with AS? or with a different assembler?
It produces so many errors, but AS itself is so complicated there may be something that I've missed, I have previously tried a couple of times to get into AS but its like swimming through treacle so I gave up until now.
NIBL-E is a substantial listing and converting to SBASM will take a while (without guaranteed success) so I'd like to get a clean run with AS if poss
Ronalds readme file suggests it should assemble ok...
Karens PAGE2.SYS assembled fine, first time, no errors, but the syntax Karen used is very different to that used in NIBL-E.ASM, its all a bit weird
Cheers
Phil

[audiokit]

Hello Phil. The numerous errors are a result of the missing processor name in top of the NIBLE.asm code. Simply put in top following line

CPU SC/MP

There still are some errors, all related to address jumps and I guess that's because they go beyond the page. Perhaps someone else knows how to surpass this error and fix the addresses manually afterwards?
Benny

[audiokit]

If I assemble the zip file on Ronald's site, I get approx. the same result
error #1370: jump distance too big.
I enclose the text document with the 31 errors (of which a lot are of the #1370 type.
As I said in my previous post I guess this is because the jumps are surpassing the page.
I hope this can be a start to get the errors fixed.
Regards, Benny

[Phil__G]

Hi Benny, no its not that, I had already added the SC/MP line after forgetting it a couple of times on the command line.
The errors dont go to stdout so dos cant even pipe them to a file for examination but looking at the lst its almost every pseudo-op.
The source NIBL-E.ASM seems a different syntax to NIBl.ASM
Ronalds 'build' file suggests that NIBLE assembles just the same as all the NIBL sources, yet the syntax is very different. I'm not convinced.
One reason I'm struggling is that I find the AS documentation overwhelming but then I'm a bit more dopey than usual with Covid at the moment. I'll email Ronald & see if he managed it. There are two options I can think of - reformat for AS using Karen-like syntax, or rejig it for sbasm.
But I think I might leave the source for now and instead do an overlay patch to fix Geco, set the baud rate and fix the error messages mixup.
But first I need to build a 32k board (the slow-boat is on its way)
Cheers
Phil

[Phil__G]

Had a chat via email with Michael Haardt and he put me onto yet another version of AS (the fourth version of AS I'd tried) and finally, all is well. With a few minor tweaks, zero errors and zero warnings Progress!
Thank you Michael!

[Realtime]

Would you be able to share that link, Phil? I've been trying AS V1.42 but it's going over my head - incredibly versatile but I'm just a poor HW engineer and I'm losing the will with that one.


I also tried to set up the Java based Dev65 which has an SC/MP overlay, but again ran out of steam.

Wrt converting NIBL-E to SBASM I started with "how hard can that be?". On the first pass I quickly resolved about 300 errors. That then opened up about 500 errors on the second pass . I resolved some but it's mainly down to finding equivalence between SBASM and AS, and the wide use of Macros.

[Phil__G]

This is the link that Michael posted, I had no success at all with the previous three versions I tried.
Whilst I very much respect that for the experts this is the holy grail of assemblers,
I really dont like it, its too complicated and too weird, everything about it is different to what you'd expect!
And intermediate '.p' files? whats all that about?
http://john.ccac.rwth-aachen.de:8000...32/aswcurr.zip
As a means to an end I can manage, but I cant imagine starting a new project with it, I'm not clever enough for this one

[ortek_service]

Originally Posted by Realtime

Would you be able to share that link, Phil? I've been trying AS V1.42 but it's going over my head - incredibly versatile but I'm just a poor HW engineer and I'm losing the will with that one.
>>
>>
Wrt converting NIBL-E to SBASM I started with "how hard can that be?". On the first pass I quickly resolved about 300 errors. That then opened up about 500 errors on the second pass . I resolved some but it's mainly down to finding equivalence between SBASM and AS, and the wide use of Macros.

Originally Posted by Phil__G

This is the link that Michael posted, I had no success at all with the previous three versions I tried.
Whilst I very much respect that for the experts this is the holy grail of assemblers,
I really dont like it, its too complicated and too weird, everything about it is different to what you'd expect!
And intermediate '.p' files? whats all that about?
http://john.ccac.rwth-aachen.de:8000...32/aswcurr.zip
As a means to an end I can manage, but I cant imagine starting a new project with it, I'm not clever enough for this one

It seems that is also the same V1.42 (Dec 2023), that Realtime was struggling with, from looking into the pdf manual for it - Nearly 500pages! but not much specifically covering the SC/MP with the huge amount of other often-obscure processors it supports, so maybe best to search for 'SC/MP' to find specific info on using that with it.

Unusually, it seems the manual lists the changes (in versions since v1.31), and that SC/MP support has only recently been added in previous v1.41(r6?) - With no changes to SC/MP implementation listed in the latest v1.42?
However, at the end of v1.42 changes, it says to refer to (more conventional these days) 'Changelog' file
- Although I can't see any of these in the zip file, Phil linked to.
(But maybe this is hosted on github etc. where you should be able to compare versions / look at full commit history, as well as view Changelog files).


Regarding .p files, it seems this section in the manual may be helpful (although not a lot, as I'm still unclear exactly what these contain, without looking at an example). But I don't think they are 'intermediate' (e.g. pre-linker, but then only relevant for multiple source files / extra library files used and usually more associated with Compilers than usually simpler single source file assembler) files as such.

6.1 PLIST
PLIST is the simplest one of the five programs supplied: its purpose is simply
to list all records that are stored in a code file. As the program does not do
very much, calling is quite simple:
PLIST <file name>
The file name will automatically be extended with the extension P if it doesn’t
already have one.
CAUTION! At this place, no wildcards are allowed! If there is a necessity
to list several files with one command, use the following ”mini batch”:
for %n in (*.p) do plist %n


PLIST prints the code file’s contents in a table style, whereby exactly one line
will be printed per record. The individual rows have the following meanings:

? code type: the processor family the code has been generated for.
? start address: absolute memory address that expresses the load destination
for the code.
? length: length of this code chunk in bytes.
? end address: last address of this code chunk. This address is calculated
as start address+length-1.

All outputs are in hexadecimal notation.

Finally, PLIST will print a copyright remark (if there is one in the file),
together with a summaric code length.

Simply said, PLIST is a sort of DIR for code files. One can use it to examine
a file’s contents before one continues to process it.

[Phil__G]

Hi Owen, sorry I meant 'intermediate' in the non-computer sense, as in a peculiar 'inbetween' stage you have to go through. I've no problem with it, and the P2HEX and PLIST etc conversion utilities work fine, the question was more like 'why?'

[ortek_service]

Hi Phil, Yes that's what I'd originally understood you'd meant - recalling some old Acorn BCPL compilers that produced 'G-code' as an intermediate file to proper machine code.

Although reading through the manual, it seemed that the .p files were just a kind of supplementary symbol table / full assembly + machine code 'Output list' file that it produced (which you then needed to use PLIST to view for some reason) - Not that you had to manually run PLIST to convert .p files to machine-code, and it doesn't do that by default!

Maybe as 'AS' essentially consists of several command-line programs, then you could write a small batch program to chain all the required ones together, so only needing to type one command to do the assembling.
I recall having to do something similar with old DOS-based IAR compilers on the 68HC11, and I also wrote ones to do all the ISP via PCBUG11, to load command files into that to save having to type quite a bit of configuration each time. So nearly a 'one-click' toolchain firmware updater.

[ortek_service]

Originally Posted by Phil__G

I wonder, has anyone successfully assembled the NIBL-E.asm source with AS? or with a different assembler?
>>
>>
Phil

I just wondered if you'd / anyone else had tried SBASM, which always seemed to be the main (/ first? - free open source at least?) PC Cross-assembler to support the SC/MP?

I do note that the new version 3 of SBASM, does now allow a space between operands, if that was a major problem with source code originally written for other assemblers - See details at the end of this page on compatibility with other systems: https://www.sbprojects.net/sbasm/scmp.php

[Realtime]

Originally Posted by ortek_service

I just wondered if you'd / anyone else had tried SBASM,

See #26 for my initial try with SBASM. I'm not suggesting it's not possible, just that it put me in my place fairly quickly .

[Phil__G]

I'm getting a good, clean assembly with AS but I've no hardware so Kris is helping me fine-tune the serial for 2400 and 8 bits to match KB+, with the KB+ geco and putc values we're still getting the occasional bad character. Its a bit of an awkward situation with the USA time difference and with Kris having to go to work (!) but until I have my 32k SC/MP it all has to be done 'blind'. One unexpected consequence of going 8bit was that all the error messages have the top bit of the last character set - so SNTX is S, N, T, X+80h which shows on Teraterm but was easily fixed in PUTC. I've taken out the redundant paper-tape reader 'start' so the mods haven't made the overall code any larger. I'll post the whole kaboosh as soon as its ready - shouldnt be long (FLW )
Cheers
Phil

[ortek_service]

Originally Posted by Realtime

See #26 for my initial try with SBASM. I'm not suggesting it's not possible, just that it put me in my place fairly quickly .

Thanks - I hadn't remembered that you'd also tried SBASM.
- Was that the latest version 3? (that apparently now handles having a space between operands, in case that was causing many errors).

Maybe there's some other tweaks that could be made to SBASM, to avoid many other errors you got, if there were a similar error (I think it was this? that Phil / Graham had already done some fixes regarding errors on branch/jumps that were actually within range).

[ChrisOddy]

[QUOTE=Phil__G;What state is your NIBL-E in now Chris, are you at 1200, do you have the bit7 set bug on echo, if you break out of a listing does it say BRK or ERR? Is your prompt '<' or '>' ?
Ta v much
Phil[/QUOTE]

I'm running NIBL-E on the MK14E at 1200
Yes I do get the bit7 error, setting PUTTY to 7-bits fixes it
I get 'BRK' if pressing Ctrl-C during a LIST
The prompt is '<'

Chris

[audiokit]

Hello Chris,
I had some private mails with Phil in the last few days. I am using a SC/MP computer as shown on Ronald's website (with the necessary mods) and both NIBL-E and FPBasic are working (switchable) on 2400 N 8 with Phil's modifications. I need to do a mod on the wire wrapped board to have RAM/ROM border at A11 instead of the current A9 and then also KITbug+ should run from the same eprom.
In the NIBL-E part you see the statement LINK #DFC1 which forces a jump to FPBasic.
I am very pleased with it, great job Phil!
Benny

[audiokit]

Yesterday late I received the last update from Phil (SCMP_ALL.hex) allowing for a menu at boot of the SC/MP. Then you can chose between NIBL-E, FP-NIBL (FPbasic) and Kitbug+

All works in the same baudrate (2400) and all without the 8 bit echo error so you can leave the terminal parameters unchanged.

This addition really makes it a workable computer with a lot of possibilities!

Again a great job Phil.
Benny

[Realtime]

Well done Phil and Benny - that sounds fantastic. I've been wanting to get NIBL-E up and running on the Low Cost Development System (LCDS), as NIBL was one of the original applications available for the LCDS. It was supplied as eight 512x8 ROMS / EPROMS hosted in the LCDS PROM Application Card, costing $595 in 1977 - equivalent to about $3000 now . This photo, from the Cambridge Centre for Computing History, may well show that exact thing. I wonder if they have ever powered it up?


I doubt I'll be lucky enough to come across one of those cards but I can program it into my home grown RAM/ROM module developed to demonstrate KITBUG+


Looking forward to the official release

[Phil__G]

Kris has announced his PCB based on Ronald Dekker's design, I've just ordered a batch so will have a few to give away, this is a really nice 100x100mm board with a 4Mhz SC/MP II, a 32k RAM (24k available), a 64k eeprom and an 8255 with 8bit I/O ports. It also has everything brought out to an 'INSbus' connector and theres an LED that can optionally be linked to an IO port. Nice
Benny's wire-wrap version is almost identical and runs Karens PAGE2.SYS ok so I expect Kris's PCB will too, so there are lots of software options with this board.
Here's Kris's FB announcement
PCBway link here
...and some photos attached:

[Realtime]

NIBL-E is now running on the LCDS. It was quite thrilling to see a BASIC listing rather than assembly code! Currently using the Erich Küster version with the bit 7 error, so needing a bit of fiddling in the TeraTerm setup. The photos show the Calendar program in action.