Generating an abort

Was a printed PRM5a ever issued? I have the PDF.

I have the printed one (along with the updated Indexes), so at least one was issued :-)

I think it came with a special bundle offer together with the AcornC/C++ DDE Version 5 stuff, directly from Acorn.

Thanks Steffen – that’d explain it as the last AcornC/C++ DDE we bought back then was Version 4 (and some of our code wouldn’t compile with cc409 so we had to keep using cc405).

(exception registers are stored is zero-initialised on startup. Mode 0 = USR26.)

Thanks for that info – is there a way of setting that to USR 32 – without resorting to ‘GO 0’?

However, given the choice between pirating content that isn’t even sold any more, and that content being lost forever, I’d say the latter is preferable.

I think it depends an awful lot on context, which is why loaded words like “pirate” aren’t helpful.

One that is affecting us right now is that a fair few software projects either don’t have any sort of distribution indicator attached, or have a strict “do not redistribute”.
Some of my own older software falls into both categories.
Why?
The reasons are twofold. Back when BBS uploads were the primary way of passing software around, one would drop it into the filebase and others could download it. So I’d generally call Arcade and The Digital Databank and upload, and frankly copyright wasn’t a big issue. It was rather implied that by making the file available was, well, making it available. Copyright and EULAs was all that rubbish on commercial software that was ignored by everybody (and in many cases flat out wrong – no reverse compilation is not “illegal”).
And, secondly, while it would be nice to have everybody have access to the latest version of something, in reality that didn’t happen. Sometimes a BBS would pick up and older version and then never update it. Plus, I used to be aware of a so-called PD library that used to not bother updating their stuff and charging rather a lot for making the discs, which is a bit galling if one makes a program for free and somebody else tries to coin it for supplying this. Not to mention all of the side effects of people discussing the use and problems of a program that’s years out of date. Better then, to forbid redistribution and point everybody at one place where the latest version lives.

The side effects of this are being felt nowadays where failure to include a statement of copyright and intent technically means the software cannot be used. Note that the Berne Convention makes things automatically copyright, so “public domain” is a rather cloudy term and the only legal way to make something PD is to have a proper copyright notice clearly stating this. A note saying “This is PD” is insufficient.

Also, when developers move on and older things get left by the wayside, the only way to keep a project alive whilst respecting the “do not distribute” rule is…to basically ignore it. This is what I’m doing with Zap and DeskLib ¹. While the rule made sense during active development in order to stop a proliferation of different incarnations ², when the thing is no longer supported it’s either “bend the rules or let it die”. With the sources and such being available, I can’t imagine “let it die” being the preferred outcome.

As for commercial software, even there it can be a tricky situation. There is still a use for older software such as Windows 3.x or WordPerfect 5.1, especially on things like DOSBox that can run these sorts of systems on entirely different architectures.
As these packages are no longer sold, your three options are:
1, the internet
2, trying to find a cheap copy on eBay where the discs are actually readable
3, there are plenty of original unopened copies on eBay, note the “rare vintage” prices (I’m pretty sure Microsoft didn’t charge that much!).

Of course, you could always do without, but that’s not quite the point. ;-)

¹ Actually my version of DeskLib predates the GCCSDK one. I started making updates and fixes around 2000 or so, having decided to stick with DeskLib instead of moving to Desk. I made a half hearted 32 bit attempt, and then did the job properly around 2012 when I needed it for my own software. Since 2015 or so I’ve been adding bits, and I make it available partly for anybody who wants to use it and partly because the GCCSDK version is, as of version 3.xx, ELF only which is great for GCC but no use for Norcroft/DDE.

² Arguably this is already happening with Zap, which is why “Zap Ultimate” is in the Store. Better a full install than a patch from here and a patch from there and… I’m not aware that the plan to adopt Zap actually went anywhere, sadly.

Thanks for that info – is there a way of setting that to USR 32 – without resorting to ‘GO 0’?

Well, you could OS_EnterOS to go to SVC mode, then switch to USR32. Note that you’ll need to disable interrupts and such as, well, older (26 bit) RISC OS won’t know what to do with a 32 bit environment. ;-)

Short example here: http://heyrick.eu/assembler/psr.html

That’s interesting – but the OS seems to be in 32bit mode – mrs r0,cpsr gives &10 – just *showregs gives 26 user mode – it would be nice to reset it without ‘GO 0’ :-)

As a note – since people are using the Emulators for development – VRPc can have two banks of memory 256Mb each – it doesn’t seem to work – is that the emulator or RO5?

older software such as Windows 3.x […] no longer sold

Not the best example, as 3.11 is actually still available from Microsoft on request :)

Does early versions of Windows work with DosBox on RO -ie internet & Sound?

A a note about a Prog switching to 26bit unnoticed from 32bit – would it be possible for a bit of code in a module to check for this and switch back.

Something in a bit call after code.

Mrs R0,cpsr
Tst R0,#&10
Moveq PC,r14
Orr r0,#&10
Mov PC,r14

Always think ‘Tst’ test is back to front.

I think this thread has become confused with this thread https://www.riscosopen.org/forum/forums/5/topics/16551 I replied in the wrong thread, and people have carried on from there.