Opcode F4 (PEA) Export mislabel

[Thaidakar16]

PEA is weird... It only accepts an immediate value, but that immediate value is written as $1234 instead of #$1234. This is what ASAR expects, and for the most part, the export gets it:
PEA.W $0030
PEA.W $0003

however, if there happens to be pointer or a label at that location, the exporter treats it like any other address, and replaces it with a label:

PEA.W PTR16_80FFFF
PEA.W CODE_808C21

It should be PEA.W $FFFF and PEA.W $8C21, but because there is a pointer at $80FFFF and a label at $808C21, it incorrectly replaces them with a label.

[binary1230]

Yea!

I've noticed this recently.

There's kind of a sledgehammer solution I added recently that works fine though I struggle to think of a more elegant way to detect this. I could also make PEA instructions explicitly NOT use or generate a label too, not sure which is the better default.

So, Diz has new "special directives" that let you type stuff in the comments field on that line, and modify the label generation.

For PEA, what I've been doing is typing

!!n ; (this disables generating a label on this line)

if you add "!!n" as a comment, it'll skip placing the label there, which is what you want for a lot of PEA usage because like you said, it's often not really referring to that memory location, it's just doing a false positive matching/generating a label.

so, like this:

[binary1230]

these directives are also a bit of a hack for "I need this and just didn't feel like programming more stuff into the GUI that day". Need to re-implement in a more discoverable way and not hack up comments fields

[binary1230]

a bit more info about the new directives is tucked away at the end of the help section:
https://github.com/DizTools/DiztinGUIsh/blob/master/Diz.App.Winforms/dist/docs/HELP.md#experimental-comment-special-directives

[Thaidakar16]

cool! thanks!
It is genuinely useful if the code pushes a literal subroutine to the stack, but the stack is often used like a scratch pad, so it is a hard thing to accommodate!