Tom Hughes wrote:
>>#17578
> Applied.
First of all, thank you for comitting these. I hate 3-way rediff's ;-)
>>#17193 necessary for imcc to write out PBC
> Applied. Like you I don't like it much but there aren't any other
> obviously better ways.
Yes, seems so.
> I missed that when it went through originally, but I don't like
> your first two suggestions as they have nasty speed and/or memory
> overhead issues. I don't really understand what you mean by the
> third one so I can't comment much on that...
imcc (when preparing the compiled code to run) produces a packed key
structure (the bytecode of a _kc constant). This bytecode gets unpacked
by the standard packfile.c, which generates the linked key structure for
runtime.
For the reverse thing in packout.c I could provide above bytecode
directly via some callback function:
opcode_t (*packout_keys_helper)(Interp interp,
opcode_t key_const_nr,
opcode_t * byte_stream);
If packout.c find's a non null function pointer, it calls this function,
which fills byte_stream with the appropriate bytes. This function could
live in interpreters PackFile structure.
This solution could be used as an additional shortcut for packout users,
that - like imcc - know, how to produce this bytestream. So the current
hack should stay in place (for the general case) and above would save
the linear search in the Constant_Table.
leo
>
> Tom
>
>
