Skull trap visual bugs about gemrb HOT 9 OPEN

2. is #128, so we can ignore it here.

from gemrb.

I tried casting it and can't see the issue in 1.. Do you only get it in that area?

from gemrb.

Was this testing done with your PR #2045 applied? Could be a regression from that, since the one problematic place is about triggered projectiles.

from gemrb.

Was this testing done with your PR #2045 applied? Could be a regression from that, since the one problematic place is about triggered projectiles.

I can replay this on master, and it's reproducible but under unknown conditions, my best guess it's that you need a small map to actually recognize this bug. But that's just an impression from testing through some IWD2 save games, so on larger maps I haven't spotted anything.

from gemrb.

Ah, I see now — for me they slowly drop from the top of the screen when positioned like in the screenshot. But fade eventually, it's odd that you managed to accumulate them.

What is it? I bet it's the exploded chunks and the positioning bug being in our gemrb/unhardcoded/iwd2/trapskul.pro. The rest is still the same old bug.
EDIT: it can't be the chunks, since the sprite appears sooner, before the skull even explodes. And there's some odd rotation going on.

from gemrb.

The bad rendering (at least the blending) is clearly just a missing application of ONE_MINUS_DST. This could either be the projectile missing PTF_TRANS or that it also has PTF_BLEND. You should be able to fix it by tweaking the flags on the pro.

from gemrb.

I doubt it should be there at all, I don't see anything interesting in the pro. SKLT is the head and a shadow bam is attached. SKLH has the parts for the explosion and that's what we don't draw properly. So perhaps there's also a desync with the explosion. We already look it up, so this must be it, I just don't see from where it comes — it will require more than code reading.

from gemrb.

something doesn't add up. SKLT and SKLH don't have a black background, but the rendering in the screenshot shows one. Are we palette munging them for some reason?

from gemrb.

Yes, this is the SP_TYPE_BITMAP type of sparkles, triggered in Projectile::DrawExplosion. So yeah, if it's not a duplicate of #128, it's very close to it.

from gemrb.