VC crash screen + load progress #2153
Conversation
fenhl
added
Type: Bug
fenhl
added
Status: Needs Review
|
@Rebbacus Could you post a screenshot of the load screen as well please? |
|
This will come up after the "You will need the classic controller" message when starting up the VC, and after bringing up the home menu. |
|
We've been discussing this internally and some concerns have been raised:
|
|
Main.py
Outdated
| if wad_app1_sha1 in wad_app1_sha1_usa: | ||
| if wad_app5_sha1 in wad_app5_sha1_usa: | ||
| wad_patch_name = "ootr_usa.gzi" | ||
| else: | ||
| raise Exception('Base WAD file is not a valid OoT USA or JPN wad.') | ||
| elif wad_app1_sha1 in wad_app1_sha1_jpn: | ||
| if wad_app5_sha1 in wad_app5_sha1_jpn: | ||
| wad_patch_name = "ootr_jpn.gzi" | ||
| else: | ||
| raise Exception('Base WAD file is not a valid OoT USA or JPN wad.') | ||
| else: | ||
| raise Exception('Base WAD file is not a valid OoT USA or JPN wad.') |
There was a problem hiding this comment.
I think that instead of having a bunch of branches here,you should pull out the values being checked into individual booleans e.g.
isUsaApp1 = wad_app1_sha1 in wad_app1_sha1_usa
isUsaApp5 = ...
isUsaWad = isUsaApp1 && isUsaApp5
[...]
if isUsaWad:
wad_patch_name = ...
elif isJapanWad:
[...]
else:
raise Exception(...)Also the exception should either be a user defined class derived from Exception, RuntimeException, or one of the other concrete exceptions. You should not use the raw Exception base class.
There was a problem hiding this comment.
That makes more sense, sure 1841b22
There was a problem hiding this comment.
Ah, you should probably still use the repo's standard for variable naming. I wrote them using camelcase in the suggestion out of habit from work.
A port of the VC functions I did for the Majora's Mask Randomizer VC - load progress of the ROM and custom crash screens.
The PPC assembly was converted to gzi commands and is placed in the unused spaces of app1. The hooks are in the exception handler, the start-up and rom loading routines, and in cpuFindFunction to save a register value when it finds an illegal instruction.
I've played for a little bit on dolphin and wii and it seems okay.
The wii crash addresses, N64 registers and last entered recomp node
An N64 stack trace is attempted. Due to how the VC doesn't really keep track of the program counter, a guess is made on where to start. This often fails however, and can result in incorrect info or the page showing unavailable.
Wii GP registers
Wii Back Chain
Some recomp code heap stuff
Couldn't get controller polling to work, so the crash pages advance on a timer, and this doesn't include any random crash patches yet - I'd like to see what crashes get reported first.