tp Npc

match d_a_npc_cd2 Fix missing short conversion in d_a_movie_player (this will reduce slightly the % for gcn, but it will be functionally equivalent the actual code, contrarily to what the current code does)

Dec 14 '25 07:12 kipcode66

Report for ShieldD (4e6dfff - e619b70)

📈 1 improvement in unmatched functions

Unit	Function	Bytes	Before	After
`d_a_movie_player/d/actor/d_a_movie_player`	`daMP_MixAudio(short, short, unsigned long)`	+4	99.24%	99.91%

Report for RZDJ01 (4e6dfff - e619b70)

[!] Report not found. Did the build succeed?

Report for GZ2P01 (4e6dfff - e619b70)

📉 1 regression in unmatched functions

Unit	Function	Bytes	Before	After
`d_a_movie_player/d/actor/d_a_movie_player`	`daMP_MixAudio(short, short, unsigned long)`	0	99.32%	99.25%

Report for RZDE01_02 (4e6dfff - e619b70)

[!] Report not found. Did the build succeed?

Report for Shield (4e6dfff - e619b70)

[!] Report not found. Did the build succeed?

Report for RZDE01_00 (4e6dfff - e619b70)

[!] Report not found. Did the build succeed?

Report for GZ2E01 (4e6dfff - e619b70)

📉 1 regression in unmatched functions

Unit	Function	Bytes	Before	After
`d_a_movie_player/d/actor/d_a_movie_player`	`daMP_MixAudio(short, short, unsigned long)`	0	99.32%	99.25%

Report for GZ2J01 (4e6dfff - e619b70)

📉 1 regression in unmatched functions

Unit	Function	Bytes	Before	After
`d_a_movie_player/d/actor/d_a_movie_player`	`daMP_MixAudio(short, short, unsigned long)`	0	99.32%	99.25%

Dec 14 '25 08:12 decomp-dev[bot]

IIRC, the reason daNpcCd2_HIO_c's ctor was separated from the data that should go inside of it is because if you move the ctor up like you did it messes up function order, and if you move the data down it messes up data order. The bot doesn't report changes in function/data order, did you find a way to fix this, or did the order regress with this PR? You can run this command locally to check the function order (doesn't work well for data order so check that manually):

./tools/utilities/weak_order_diff.py d/actor/d_a_npc_cd2.o

Also, if you're moving cd2's data to be in-function statics anyway, there are two more of them: a_transScaleTbl$3998 and a_jntNumTbl$4692. I think those two can be fixed without messing up the order unlike the other four.

Dec 14 '25 17:12 LagoLunatic