Conditional Chains don't chain

[ohrrpgce-bugbot]

[bz#851]

A newer nightly feature, Conditional Chains, don't seem to work when it checks for Tags. Regardless of whether the tag is on or off, the chained attack never happens.

From: Bobert-Rob <bobertinfiniti>
Reported version: 20110506 Zenzizenzic
Operating system: Windows

[ohrrpgce-bugbot]

Comment author: @bob-the-hamster

I just tested this, and the tag conditional chaining seemed to be working fine for me.

What is the version number of the nightly build you tested with?

Were you testing with regular chains, Else-Chains, or Instead-Chains?

Can you attach a test-case that demonstrates this problem?

[ohrrpgce-bugbot]

Comment author: Bobert-Rob <bobertinfiniti>

Well, when I tested it with a new file, it chained fine. Seems to only be happening in an old game I've been working on, one that's needed to be updated a few times. I have two attacks that won't work. One involves a tag set when the previous attack hits or misses, then chains depending on whether the 'miss' tag is on or off. The other depends on a tag that's set when you equip a certain item to give you an extra chain to a current multi-chained attack. The tags change like they're supposed to, but the attacks seem to ignore this data. Perhaps due to attack data not being updated to current standards properly? That's all I can guess at.

[ohrrpgce-bugbot]

Comment author: msw188

Created attachment 191
test case

Okay, here's what I've come up with so far. I cannot yet reproduce the bug where certain attacks are skipped. BUT I have managed to create a somewhat separate bug that I think is related. Here's the breakdown:

AtkA
if(mp<10),then instead of AtkA do AtkB
AtkB reg chain to AtkC

In this setup, if AtkA and AtkB target the same class, then everything's fine. HOWEVER if AtkA and AtkB have different targetting classes, then AtkB gets stuck in an endless loop chaining to itself. EVEN STRANGER in Tales 2 the issue was that AtkB would get skipped and AtkA would chain directly to AtkC. I cannot yet reproduce that behavior.

The test case is set up to show AtkB getting stuck. If you go into custom and change AtkA's target class to enemy (to match B and C), then everything works fine.

ALSO the instead chain seems to need an 'extra' turn to handle itself properly. Right now the hero has 11 MP, but AtkA costs 2MP. BUT I need to use AtkA twice to get the instead chain to roll. Odd? Bizarre? Yeah, I think so...

Attached file: TEST486.rpg (application/octet-stream, 274881 bytes)
Description: test case

[ohrrpgce-bugbot]

Comment author: msw188

Ok, ignore the last paragraph of my previous comment. It turned out my stupidity was in charge there; I had given my testing hero 50 MP~ (for fun...? I just gave him 50 of like everything for no real reason) which made AtkA actually only cost 1MP. I'm such a chud.

[ohrrpgce-bugbot]

Comment author: msw188

Really bizarre development! It seems to matter what order the attacks are in, in terms of their attack IDs. In the attached test case, if we simply make:
AtkA chain to AtkC instead,
then AtkC regular chain to AtkB

everything seems to work fine. With that in mind, I want to recap with numbers instead of letters:

Atk0 instead-chains to Atk1 which regular-chains to Atk2
-this gets stuck looping endlessly on Atk1 IF AND ONLY IF the targetting classes of Atk1 and Atk0 are not the same

Atk0 instead-chains to Atk2 which regular-chains to Atk1
-this seems to work fine

(In Tales 2)
Atk227 instead-chains to Atk229 which regular-chains to Atk39
-this skips Atk229 and goes right to Atk39 and terminates (no loops)

Atk227 instead-chains to Atk252 which reg-chains to Atk229 which reg-chains to Atk39
-this skips Atk252 and then 229 and 39 work correctly

Atk228 ins-chains to Atk230 which reg-chains to Atk39
-skips 230

Atk228 ins-chains to Atk253 which reg-chains to Atk 230 which reg-chains to Atk39
-skips 253, then 230 and 39 seem to work fine

[ohrrpgce-bugbot]

Comment author: @bob-the-hamster

I think I have this one beat.

When an instead-chain was delayed, or when an instead chain changed target class, the instead chain would get stuck in a loop. For some cases (the delay one, I think?) this would manifest as an endless attack loop. For the target change loop, it would spin out of control, hit a safety cutoff, and the next attack would behave crazy/wrong.

I am forcing a new nightly build with the fix, should be up in a matter of minutes. Let me know how it works.

[ohrrpgce-bugbot]

Comment author: msw188

Okay, I didn't test every imaginable case, but all of the issues I had in my test case and in Tales 2 seem to have been resolved, so I think it's safe for me to mark this fixed...? Thanks very much James.