That's it. That's all it did. It did not even bother identifying symbols that I hadalready identified, other than GetWrestlerType
. I mean, since the gameuses overlays, I knew I was going to have to put in extra work, but the way people actabout this, you'd think the entire game is available just after running the program.At the very least, I thought it'd get ID2toID4and ID4toID2.
For full transparency, this is the vpw2.toml
I originally used:
[input]entrypoint = 0x80001000# Paths are relative to the location of this config file.elf_path = "vpw2_z.elf"output_func_path = "RecompiledFuncs"
The next time someone asks about "omg let's put the AKI games through recomp",link them this page.
(also, even if this worked, I would still have to do things that I'm alreadydoing in the context of the decomp, such as replacing hardcoded file IDs withsymbol names. getting the full game code output is just one piece of the largerpuzzle.)
2024/05/18 Update
I fixed the entry point to be 0x80000400, but that did not provide any newC files. So the problem is likely elsewhere. I am not insulting the tool or itsauthor(s), but I do want people to have realistic expectations. I have seenpeople assume that the recomp program can "port n64 games to PC in a matterof seconds", which is blatantly false (in some situations).
Another elephant in the room is the fact that the Mario and Zelda decompcommunities are much larger than the decomp community surrounding the N64 AKIgames. It's been this way for quite some time (decades, at last check); the onlyway to fix it is to get more knowledgable people involved in our scene.
2024/05/23 Update
Further research shows that the only functions that got decompiled properlywere ones that were already converted to C. There seems to be a reason for this...Quoting Mr-Wiseguy:
The recompiler can only use symbols that have valid sizes in the elf,so it may be ignoring them and then finding the function later on via the staticfunction detection process.
Most of the code in the VPW2 decomp is still assembly in GLOBAL_ASM
blocks, which might explain why I wasn't able to get anything reasonable.The VPW2 decomp was also started long before N64 decompilation became a bitbetter organized and/or standardized.
the situation might change in the future, so I'm keeping an update log just in casethis page gets any major changes