in Xbasic kun memory is constrained by the requirement to have both the compiled and tokenized version side by side.
If you take into account that the basic area is itself, at most about 28000 bytes, it's easy to understand that xbasic limitation is somewhat to much heavy.
Nestor Basic (XBasic+), not have this limitations.
You able using almost whole RAM for code, and even VRAM.
Is the basic source code held in vram? If it is the limitation is still here
Is the basic source code held in vram? If it is the limitation is still here
No problems, simple - not use it.
Is the basic source code held in vram? If it is the limitation is still here
No problems, simple - not use it.
Maybe it would help to explain why the original code remains in memory (I thought maybe for editing, saving)? And whether or not it is actually a limitation. I don’t know if it is, can the compiled output be greater than the available ~28kB? If it can’t it is not really a limitation anymore (as long as the uncompiled code is stored in other RAM/VRAM).
I don’t know if it is, can the compiled output be greater than the available ~28kB? If it can’t it is not really a limitation anymore (as long as the uncompiled code is stored in other RAM/VRAM).
Compiled frame can't be greater 10kb. Uncompiled code might be stored in any RAM/VRAM, with special call syntax and interslot calls (also with special calls for VRAM snippets).
Any news about this project ?
i've seen that on IBM PC the amount of free ram for basic is more higher than on msx, maybe due to the ability to address more memory on x86 CPUs. I suspect that an entire segment is reserved to BASIC PROGRAM.
This may be an additional difficulty when porting to msx..... would be nice to use the slot mechanism to emulate the x86 segments....