Bug - Fixed crystalBallMonster incorrectly reset

[ckb]

In some cases, "crystalBallMonster" will be incorrectly reset to nothing (the property will be cleared) when the prediction is still true.

I believe this happens when the prediction gets set during a free turn (such as a free banish) and the next single turn is in the same location, "crystalBallLocation"

 

[gausie]

The whole thing needs implemented differently, as each location can have a concurrent prediction. We need to be able to track free noncombats to do it fully accurately

 

[ckb]

Two possible options that would allow a scripter to parse or script around the predictions :

A single preference, "crystalBallPrediction" or some such, that is a delimited list of location, monster, and turncount of when that prediction happened. It could just add to any existing list:
"LocationA:MonsterA:TurnA|LocationB:MonsterB:TurnB|LocationC:MonsterC:TurnC"

Option 2 would be to add a set of proxy records to each location, one for crystalBallMonster, one for crystalBallTurn

 

[fredg1]

A single preference, "crystalBallPrediction" or some such, that is a delimited list of location, monster, and turncount of when that prediction happened. It could just add to any existing list:
"LocationA:MonsterA:TurnA|LocationB:MonsterB:TurnB|LocationC:MonsterC:TurnC"

I did exactly that

 

[fronobulax]

I did exactly that... 10 months ago...

I, at least, stopped looking here for older patches to review so perhaps you could submit it as a PR?

 

[gausie]

This is now merged r25998 - please see the announcements for the new preference

 

[ckb]

Thank you! Looing forward to untangling the Orb mess even more in my script now.

 

[fewyn]

> Task :test

FightRequestTest > crystalBallPredictions() FAILED
    org.opentest4j.AssertionFailedError at FightRequestTest.java:173

ChatRequestTest > isShouldBuildAModernChatRequest() FAILED
    org.opentest4j.AssertionFailedError at ChatRequestTest.java:70

ChatRequestTest > isShouldBuildAnOlderChatRequest() FAILED
    org.opentest4j.AssertionFailedError at ChatRequestTest.java:80

744 tests completed, 3 failed, 7 skipped

> Task :test FAILED

 

[gausie]

> Task :test

FightRequestTest > crystalBallPredictions() FAILED
    org.opentest4j.AssertionFailedError at FightRequestTest.java:173

ChatRequestTest > isShouldBuildAModernChatRequest() FAILED
    org.opentest4j.AssertionFailedError at ChatRequestTest.java:70

ChatRequestTest > isShouldBuildAnOlderChatRequest() FAILED
    org.opentest4j.AssertionFailedError at ChatRequestTest.java:80

744 tests completed, 3 failed, 7 skipped

> Task :test FAILED

These aren't failing vocally or on GitHub. Can you post the test output?

 

[fewyn]

These aren't failing vocally or on GitHub. Can you post the test output?

Test results - Test Summary

 

[gausie]

H
unfurl="true"]https://ci.kolmafia.us/job/Kolmafia/26633/execution/node/3/ws/build/reports/tests/test/index.html[/URL]

Huh, this is leaking between tests I think

 

[fronobulax]

Given the quantity of work I did to fix leaks as well as the fact that this isn't happening locally I think I'd like to know more.

 

[fewyn]

I've temporarily disabled tests in the meantime on Jenkins.

 

[heeheehee]

I saw pwd mismatch issues in ChatRequestTest and turncount mismatch in FightRequest, I can look into it

 

[heeheehee]

Yep. Both of these are leaking from the charpane decorator tests (which parse `var turnsthisrun = 236` from the javascript, and set GenericRequest.passwordHash = abc123).

 

[fronobulax]

Yep. Both of these are leaking from the charpane decorator tests (which parse `var turnsthisrun = 236` from the javascript, and set GenericRequest.passwordHash = abc123).

OK. AFAIK the charpane test was added within the last 24 hours so that explains at least part of the reason why it pops up now. I'll put the clean up on my list although I am not working through my list as quickly as I'd like. If someone does it first, that's fine.