Binary bytecode files
It's an obvious performance improvement to use a binary format for the bytecode.
If we're going to support 64-bit systems, we need to allow for 64-bit values in the data section (see #2 (closed)). We don't want to make all constants 64 bits however, so we would have to think of a way to deal with that. For instance, we could place all the 64-bit values first, and prepend it with the number of 64-bit values. But this is a rather minor detail.
There is also the issue of 64-bit constants in the code section, e.g. for pushI
.
Edited by Clean Importer