NMM2CSV - Edit tables with Excel instead of Nightmare [UPDATED TO V1.0]


#21

Thanks man. Deleting the animation modules did the trick. I’ll just have to find ones that work.


#22

Hey there an aspiring hacker here :smiley: wanted to mention I keep getting an error when using the programs, I’m attempting to hack FE8 and everytime I go to use the program at the it goes

My apologies if I’m missing something obvious.


#23

are you sure you’re using Unicode 8


#24

Change your system from Japanese to English.


#25

:wave:

I made a custom alternative build of c2ea. The differences are a bit technical (see below), but basically if you’ve been using the “conventional” MAKE HACK_full.cmd/MAKE HACK_quick.cmd method it makes:

  • MAKE HACK_full.cmd slightly faster overall in theory (may or may not be insignificant).
  • MAKE HACK_quick.cmd significantly faster.

The reason for that is that this version, when repointing tables, instead of putting #inctext PFinder ... into the generated event files, it puts the repointing code directly. This basically means that Event Assembler won’t have to call PFinder itself for each repointed table. This does make c2ea itself take a bit more time, but the full build will be overall faster as it removes the overhead of calling the PFinder executable for each table.

Here are some numbers to prove my point (ran on some VBA-BT develop state from yesterday, using ColorzCore):

time ./Core A FE8 -output:"../VBA-BT.gba" -input:"../ROM Buildfile.event"

With old c2ea:
real    0m33.459s
user    0m0.000s
sys     0m0.031s

With new c2ea:
real    0m14.132s
user    0m0.015s
sys     0m0.000s

I don’t know either about those user/sys time values but I assume it’s just msys being msys, or maybe VM shenanigans. Just look at real time. So yeah, EA takes about half the time it took before.

All of this was done because PFinder is hella slow.


#26

:wave: again!

It seems that me messing with this has made @Crazycolorz5 also take an interest in optimizing this, and thanks to his efforts c2ea itself is now also lightning fast.

So here’s a link to the release download.

Basically what Colorz did is some caching shenanigans: when c2ea needs to do pointer finding, instead of starting to look into the ROM immediately it will check if it didn’t already do that search during a previous invocation, and if it did it will just load the cached results instead of doing the search again (because searching through 16MB worth of data is kind of slow).

n2c remains unchanged.


#27

Edited the new links into the OP.


#28

So can we just simply replace c2ea with this newer version and it’ll work better? We don’t have to do anything else like change our MAKE HACK files, do we?


#29

Yes, just a simple replace should work.