| 1 Aug 2021 |
 Midnight#9871 | Is this what a lookup table is, or is that something else? | 21:27:16 |
 Eievui#2759 | Yes. | 21:27:24 |
| Midnight#9871 | How does this actually work, though? How can you easily access any item in the table other than the first? | 21:32:05 |
| Eievui#2759 | By adding a to .table, we can offset the table to point to the corresponding entry. | 21:32:46 |
| Eievui#2759 | That's what this does:
add a, LOW(.table)
ld l, a
adc a, HIGH(.table)
sub a, l
ld h, a
| 21:33:00 |
| Eievui#2759 | hl is now .table + a | 21:33:11 |
| Midnight#9871 | Hmm... I'm not sure if something like that would be too long for what I'm trying to do. | 21:34:56 |
| Midnight#9871 | When using an LYC interrupt, how many cycles do you have to do something before it moves on to the next scanline? | 21:35:21 |
| Eievui#2759 | It probably would be. Like I said, all of this is unnecessary because those values are sequential. | 21:35:23 |
| Eievui#2759 | https://gbdev.io/pandocs/pixel_fifo.html?highlight=FIFO#pixel-fifo | 21:35:37 |
| Midnight#9871 | I'm not working on the memset stuff right now. | 21:35:38 |
| Eievui#2759 | An entire line is 114 cycles, VRAM time per line varies. | 21:36:17 |
| Eievui#2759 | It's best to assume you only have 16, because that's the minimum. | 21:36:43 |
| Eievui#2759 | (of VRAM time) | 21:36:48 |
| Midnight#9871 | 16's not a lot. | 21:37:54 |
| Eievui#2759 | What do you need to do in that time? | 21:38:20 |
| Midnight#9871 | Do something to figure out which line I'm on, in order to know what to do for that line, and then do something (probably mostly scrolling). | 21:38:56 |
| Midnight#9871 | And then after that, set stuff up for the next interrupt. | 21:39:15 |
| Eievui#2759 | So you only need to load a single value during HBlank | 21:39:22 |
| Midnight#9871 | ? | 21:39:34 |
| Eievui#2759 | None of that requires VRAM access. | 21:39:48 |
| Midnight#9871 | The scrolling does. | 21:39:59 |
| Midnight#9871 | Doesn't it? | 21:40:09 |
| Eievui#2759 | It doesn't, but that's why I said this ^ | 21:40:15 |
| Eievui#2759 | Do all of that work before HBlank, update scroll (load one value), then prepare the next line. | 21:40:41 |
| Eievui#2759 | ; Figure out what the scroll value is and store it in `b`
; Now wait for HBlank
: ld a, [rSTAT]
and a, STATF_BUSY
jr nz, :-
ld a, b
ld [rSCX], b
; Now prepare the next line
| 21:42:22 |
| Midnight#9871 | I can't think of how to do that. (for what it's worth, I didn't sleep well last night, so I'm pretty out of it) | 21:42:33 |
| Eievui#2759 | Just put the rSTAT loop before the part where it's actually needed. | 21:43:13 |
| Eievui#2759 | Like in that example | 21:43:18 |
| Midnight#9871 | ...Yeah, I think I'm probably going to have to take a break from this for today. I'm apparently tired enough that my ability to read code has hit pretty much zero. | 21:43:36 |
