* write intval size into PBC header
Leo, I know this is a first cut at freeze/thaw, and I'm happy you've done it. Let me make some comments to you and Dan.
I'm pretty sure Dan and I discussed this when I was reworking bytecode to be portable last year, but I think at the time we decided there were other things (like freezing PMCs) that needed to be done. Now that it is in, its time to re-light this discussion until we get a document out of it. :)
A) I didn't look at your commit to see the reason for storing the INTVAL size into the bytecode header, so I won't question it yet, but since p6i has argued this for a couple of years, I want to restate some things.
Parrot currently assumes INTVAL size == OPCODE size because both get configured as the same integral type, although you can choose to override it, it is not supported to choose INTVAL > OPCODE, though the inverse is. So storing it in the header is probably redundant, unless we change the rules.
While we may have platforms that have void */size_t > than INT, we should not have any where void */size_t < INT.
An int should always pack into 1 opcode_t, if it doesn't we have configured
wrong for INTVAL (someone chose a 64-bit type on a 32-bit platform),
and the fact that by configure, it is even possible, says we have the
typedef hierarchy wrong. I still feel that Parrot_Int should be dependant on Parrot_Opcode.
B) For bytecodes to be portable we have to have some rules about standard sizes. Yes, we can unpack a 64-bit bytecode on a 32-bit CPU, but INTVAL constants (non-PMC) get truncated.
Since it is NEVER portable for someone to write code using the Ix registers with values > than 32-bit, we have to make a rule. Any use of integers > than 32-bit should use a PMC type, not an integral type.
The fact that:
set I0, 0 LOOP: inc I0 lt I0, <MAXINT CONSTANT>, LOOP
Runs differently between platforms is enough reason to make a least common denominator MAXINT the Parrot standard.
The summary of this:
Parrot must use a PMC Int type to store portable data, so the INTVAL should be stored as part of the PMC stream, not the bytecode header, as we may support multiple sizes.
-Melvin
