Deprecate LDIO

[Rangi42]

I already expect the results of our polls will be to deprecate ldio [...] but not ld [$ff00+...], so went ahead and implemented it in what should be an efficient way.

(In the test case, I know ld [$ff11], a and ld a, [$ff11] assemble to different opcodes than the rest; they're just there for completeness, and they illustrate b0495b3.)

[Rangi42]

Frankly it might be simpler to just remove, since that would let people do a one-line DEF ldio EQUS "ldh" "fix".

(The poll has zero votes for using ldio so far, and before today there were barely any mentions of it in gbdev, and most of those were wondering whether it's even supported by RGBDS or not.)

[vulcandth]

Unpopular opinion... Too much deprecation. I say just remove it, and let the assembler errors notify people they need to update their code. It's a simple replace anyways.. I know lots of people who ignore those deprecation warnings until a release causes it to stop building... And then they might just stay on the old version anyways.

Prepares for incoming rant :P

[Rangi42]

Just to note, re: allowing multiple alternatives in general, we already do so for ld [hli/hld] aka ld [hl+/hl-] aka ldi/ldd [hl]. But unlike ldio, none of those is an obvious candidate for deprecation. (Even ldi/ldd get some use; they were explicitly requested in #73.)

[Rangi42]

So far 13 votes to deprecate ldio vs just one to keep it; but 7 votes to keep ld [$ff00+c] vs 9 to deprecate.

[ISSOtm]

Yeah, ldio is uncontroversial enough, so let's act it out. We'll see about the other syntax in a separate thread.

[ISSOtm]

Unpopular opinion... Too much deprecation. I say just remove it, and let the assembler errors notify people they need to update their code. It's a simple replace anyways.. I know lots of people who ignore those deprecation warnings until a release causes it to stop building... And then they might just stay on the old version anyways.

Oops, that was actually a good point, I think. Hmm, well, the main reason we're using deprecation warnings is to provide a nicer error message, so people who do upgrade are less confused and lost than getting a generic syntax error.