Recent Low Vector ROM for Pi

Perhaps it would make sense for OS_Byte &A6-&AF to return an error, at least for high vector ROM builds.

And while we’re there, officially deprecate it! ;-)

Perhaps it would make sense for OS_Byte &A6-&AF to return an error, at least for high vector ROM builds.

All the unallocated OS_Byte above &A5 should probably return an error. I’m not sure it would help trap use of &A6 though as XOS_Byte would suppress the error and it’s unlikely any code that’s using undocumented calls is going to be doing any error checking.

There’s a fix for this localised desecration of an ancient API here

The MOSVarFix module ensures OS_Byte,166 always returns the correct result, even if someone tries really hard to break it.

The MOSVarFix module

Link doesn’t work with NetSurf – be handy if it did!

Interesting set of patches. Do you have a file of source mods for those who would like to patch this at source level when building a custom ROM?

I already have my own set of patches:

• Bump up the IIC bus speed
• Make CET and CEST known to UK territory
• Adjust click on Switcher opens config
• Switch F12 for ShF12 (kitten proofing)
• Change default boot video to 1280×1024 (as is my monitor)
• Something else, I forget. ;-)

I ought to automate these patches…

BTW the whole “it’s not documented” whinge is completely bogus as I’ve just discovered OS_Byte,166 is used by SoundChannels (amongst others, I imagine).

Which means, because somebody broke it, there’s some extra code in SoundChannels to try ReadSysInfo,6,71 first and then fall back to OS_Byte,166 if that fails. Everyone else who uses OS_Byte,166 is forced to do the same all because somebody couldn’t be bothered to keep 166 working.

Zero justification for that. It’s just vandalism. Properly cross about it now.

Link doesn’t work with NetSurf

The link in the posting above? The link on the page to the zip?

It’s all cobbled together from Google Sites and Google Drive (because you can’t put zips onto Sites), and so it’s ALL a bit weird. Not to mention that the usual method of domain spoofing utterly failed with Sites (frame redirection produces a blank page).

There is an old saying:

• You can have it Cheap
• You can have it Fast
• You can have it Right
• Choose two

I chose the first two. ;-)

Rick asked

Do you have a file of source mods

I have decided to not work on the sources. I have never worked from them, so it’s a whole extra step to then go and work on them. Others who are more familiar with the structure, conventions and style of the source are better placed to modify and test them than I.

Interesting set of patches

Well there’s a lot more where that came from. I’m going to start releasing stuff soon.

Should I announce things in the forum, or just let people find things?

John said

All the unallocated OS_Byte above &A5 should probably return an error.

No, that’s absolutely the wrong approach.

Take 188-190 for example. These control various aspects of the analogue to digital converter on the Beeb. On a standard RiscPC they will do nothing but update the MOS variables… unless you have a BBC IO Podule fitted, in which case they work the same as they always did.

We’re not short of OS_Bytes – they’re a legacy interface which will not be extended – so we don’t need to retire any of them to make room for new ones. For the sake of the 88 bytes of RAM and 8 instructions required to implement OS_Bytes &A8-&FF, I say leave them alone.