Menu
▾
▴
1 Attachments
#1598 5th Gen Pokémon games—scene changes are broken
As far as I can tell, scene changes when entering battle in the 5th Gen Pokémon games have been broken since at least revision 5284. I haven't been able to narrow it down any further than that, but the following behavior looks to me like a framebuffer bug:
Discussion
Anonymous
On second thought, ignore this—it seems to be a driver issue, not an emulator bug.
Last edit: Alec Warszycki 2016-10-07
View and moderate all "bugs Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Bugs"
if i select the game it will not load
Last edit: Anonymous 2016-12-24
try changing config > rom loading
View and moderate all "bugs Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Bugs"
I identified that the bug started to appear at build 5260.
Last edit: Anonymous 2016-12-24
View and moderate all "bugs Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Bugs"
r5259 my bad. If there is a way to debug the thing when the trainer goes into a pokemon battle, I'd like to know (since building on VS2015 is broken on r5259 and 5260, tested on r5260 buildbot)
You can just comment out the code that didnt compile that was fixed by r5260. What's the last build you have before r5260 (the one where the bug doesnt appear?)
View and moderate all "bugs Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Bugs"
The last build that didn't have the bug was r5258. The bug happens when JIT is enabled (interpreter works just fine).
Last edit: Anonymous 2016-12-24
Are you discussing the exact bug pictured in the screenshot on this thread? You say "trainer" and OP doesnt. It is hard to see how r5259 would have changed anything that is sensitive to the JIT
Last edit: zeromus 2016-11-08
Looks like he's talking about the same bug—I just tested, and turning off JIT does fix the bug when entering battle. The zoom-in effect that is supposed to happen when entering buildings is still broken, though.
This is the first I've heard of entering buildings. What does that have to do with this bug thread? That sounds like a different bug entirely.
Yeah, you're right. I'll make a new issue report for the building thing, once I've bisected it.
Last edit: Alec Warszycki 2016-11-08
So this bug is about hangs when entering battles, which can be solved by turning off the jit? See, that's the thing I can't possibly believe gets bisected to around this revision range. But if someone can confirm they have bisected the entering-battle issue to these revisions truly (and what is the difference between a battle and a trainer battle?) then I can look into it, simply because it's so unexpected--but anything fixed by turning off the jit will likely never, ever be fixed.
In my experience, the game isn't hanging—it's just graphical corruption that happens when entering any kind of battle, trainer or not.
I never said it was hanging, in fact...
The bug is about graphical corruption, correct? And there is no bug with JIT on build r5258. The bug begins to appear with JIT at build r5259 (the changelog says that there is heavy code cleanup in the GPU section, I tested it on r5260 because the buildbot didn't make a build for r5259 but the changes from the two builds look very minimal). And what do you mean by bisect?
I have attached a savefile to reproduce the bug with (just walk in the grass until a Pokemon appears). I'd debug it myself but I'm inexperienced (new student problems).
(oh yeah sorry for the multiple anonymous postings, the email for acount verification barely showed up an hour ago)
Edit: If you need any logs from the GPU stack, I can try to provide them (if you can tell me how to get them)
Last edit: Kevin 2016-11-08
View and moderate all "bugs Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Bugs"
Bug happens when JIT is enabled btw.
Last edit: Anonymous 2016-12-24
Fixed in git#36e03fe.
View and moderate all "bugs Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Bugs"
This bug is absolutely not fixed—entering battle still presents broken graphics during the transition.
looks pretty broken to me. like it's rendering trash vram. seems like the game captures the 3d to VRAM D right before doing framebuffer effects, but for some reason it is failing to capture in this case. It looks like timing may be involved, as I see it only trying (and failing) to capture part of a frame
He hasn't updated the github page to have that commit it seems, let's wait for that commit to hit the source.
Commit git#36e03fe is an older commit back from 2016/12/29.
It seems like this issue actually contains two completely unrelated bugs that I didn't notice the first time around.
The first bug involves the zoom-in effect not working when entering buildings. This bug has been addressed in git#36e03fe.
The second bug involves some stale graphics being drawn at the top of the screen upon entering battle. This bug has not been addressed in git#36e03fe.
So I'll reopen this issue since it is clearly a regression.
Fixed in git#5e06fbc.