I'm about ready to commit the fixes to the current assembler, which will finally make proper keyed aggregate access ready. Looking at the test suite, I see some places where we use raw opcode names, such as 'ne_n_nc_ic'. It appears to be random, and I'd like to know if keeping the ability to access raw opcode names is of any use? I personally don't think so, beyond the dubious benefit of being able to validate that you meant the opcode you typed in.
I'm tempted to simply rewrite the tests properly (t/op/ifunless, t/op/number, t/op/stacks, and t/op/time) to use the proper opcode. Thoughts, anyone? -- Jeff <[EMAIL PROTECTED]>