Hi Daniel,
I’ve been following your work on bringing the UCSD p-System to the ZX Spectrum Next, and I wondered whether you might be interested in a related experiment.
I’ve recently been working on adding the original UCSD/Sage BASIC and FORTRAN IV.0 compilers to the Z80 p-System supplied with RomWBW.
We have reached an interesting point: both BASIC R4 and FORTRAN compile successfully, but even trivial compiled programs fail at runtime in REALOPS:
Unimplemented instruction
Segment REALOPS Proc# 14 Offset# 73
The system reports SIZEOF(REAL) = 4.
After investigating the original SofTech adaptable-system documentation, we discovered that the RomWBW SYSTEM.INTERP appears to have been built from INTERP.Z without one of the floating-point packages. We recovered an original FP4.CODE and attempted to rebuild SYSTEM.INTERP with:
INTERP.Z + FP4 + RSP + BIOS.CRP + INTER.X + TERTBOOT
The linker and COMPRESS stages complete, but our resulting interpreter does not boot, so clearly there is still something we are missing.
I’ve documented the complete investigation here, including a bootable p-System image containing BASIC R4, FORTRAN, their libraries, FP4.CODE, and our experimental interpreter:
wwarthen/RomWBW#706
Since you’ve recently worked on the Z80 side of the p-System for the Spectrum Next, I thought this might be right up your street.
I’d be particularly interested to know whether you’ve encountered the FP0/FP2/FP4 interpreter configuration in your work — and, perhaps even more interestingly, whether BASIC and FORTRAN could eventually be incorporated into your Spectrum Next p-System distribution as well.
Everything needed to reproduce what we’ve done is attached to the RomWBW issue.
Any thoughts would be greatly appreciated.
Best regards,
Giancarlo
Hi Daniel,
I’ve been following your work on bringing the UCSD p-System to the ZX Spectrum Next, and I wondered whether you might be interested in a related experiment.
I’ve recently been working on adding the original UCSD/Sage BASIC and FORTRAN IV.0 compilers to the Z80 p-System supplied with RomWBW.
We have reached an interesting point: both BASIC R4 and FORTRAN compile successfully, but even trivial compiled programs fail at runtime in REALOPS:
Unimplemented instruction
Segment REALOPS Proc# 14 Offset# 73
The system reports SIZEOF(REAL) = 4.
After investigating the original SofTech adaptable-system documentation, we discovered that the RomWBW SYSTEM.INTERP appears to have been built from INTERP.Z without one of the floating-point packages. We recovered an original FP4.CODE and attempted to rebuild SYSTEM.INTERP with:
INTERP.Z + FP4 + RSP + BIOS.CRP + INTER.X + TERTBOOT
The linker and COMPRESS stages complete, but our resulting interpreter does not boot, so clearly there is still something we are missing.
I’ve documented the complete investigation here, including a bootable p-System image containing BASIC R4, FORTRAN, their libraries, FP4.CODE, and our experimental interpreter:
wwarthen/RomWBW#706
Since you’ve recently worked on the Z80 side of the p-System for the Spectrum Next, I thought this might be right up your street.
I’d be particularly interested to know whether you’ve encountered the FP0/FP2/FP4 interpreter configuration in your work — and, perhaps even more interestingly, whether BASIC and FORTRAN could eventually be incorporated into your Spectrum Next p-System distribution as well.
Everything needed to reproduce what we’ve done is attached to the RomWBW issue.
Any thoughts would be greatly appreciated.
Best regards,
Giancarlo