| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
matching is now significant because register no. fields are re-used as additional function codes in operations with fewer operands so I pulled out all decode clauses to the beginning of file for easier rearranging. Old encodings can co-exist with new encodings as the only overlap is for recently added instructions which already use the new scheme. Eventually the old encodings will go away, however, and the opcode space may be reclaimed."
|
|
generation for dubious casting enumeration to int.
|
|
|
|
this and use shallow embedding conversion?
|
|
wrong variable name (iK vs. ik). Spotted via compare_analyses.
|
|
and not obsolete interp_interface one.
|
|
|
|
|
|
|
|
|
|
things, SF and CP bugfixing
|
|
|
|
|
|
|
|
|
|
|
|
|
|
comparison fails. This is specified in manual and also helps RMEM with locked writes.
|
|
|
|
|
|
|
|
|
|
|
|
etc/regfp.sail.
|
|
|
|
know the possible destination of an instruction for memory indirect jumps (the register name is not used).
|
|
indirect jumps.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
and Set membership for free
|
|
|
|
|
|
|
|
|
|
fixed some compare functions;
|
|
|
|
|
|
|
|
fixed the interpreter nias analysis;
|
|
|
|
|
