RWR bugfixes #7314
RWR bugfixes #7314
Conversation
| :req (req (let [all-cards (get-all-cards state)] | ||
| ;; TODO - assert this works when (e.g.) using the last click on lib-acc | ||
| ;; and having it trashed (all-cards wont find it) |
There was a problem hiding this comment.
Why won't all-cards find it?
There was a problem hiding this comment.
If the card gets trashed through use (ie last click on liberated accounts). I need to test this and check the behaviour, since I think things like that are where the issue actually is, and why some people aren't reproducing the bug and some are - and first-event is checking for cids using the current state.
|
I don't wanna get in the way of bug fixes, but could you add some test asserts for some/all of these? |
No worries, that was my plan anyway. Just gotta leave it for the weekend since I'm crunched on week days. |
|
That's all my fixes, I'll add more tests after work on Saturday and we can hopefully push this before the AMT on Sunday |
|
OK now it's DONE done. The only outstanding issue is Juli - she's still not CORRECT, but she's more correct that she was before. That's a bandaid, and I'll investigate making it correct correct next weekend. |
descent charges a credit when expended (closes #7312) - not testing
Juli moreira once per turn (closes #7310)
valentina only when runner removes tags (closes #7309) - unit tests
amaneunsis only works when runner removes tags (closes #7313) - unit tests
cloud eater trash program ability (closes #7321)
lightning lab demanding use every turn (closes #7324) - unit tests
kingmaking draw-up-to (and async on decline) (closes #7316, closes #7325) - unit tests
cupellation disables cards when hosting them (closes #7326) - unit tests
fixed async issues on descent and eminent domain (closes #7327 ) - no tests
fixed isaac lingering prompt (closes #7319) - unit tests
fixed Window of Op optional (closes #7315) - no tests
removed auto-break from pressure spike and boi-tata until we figure those out (closes #7322, closes #7320)
Updated force-ice-encounter to not offer jack-out when used during a run (but before access) (Closes #7308)
===============
Praccess actually works they way it's meant to now. The changes are mostly sane, needing to store an ephemeral state while resolving each install is so that Jarogniew Mercs taking a tag doesn't load a second copy of the same event (which is now resolving in a different window), which would then resolve first (before allowing you to resolve, say, your trigger from seb), before returning you to the first window with an empty button to press (because it's already resolved).
Reading that paragraph may induce a schizm. It is what it is. I'm not probably not writing a unit test for this case, because it's cumbersome to actually reproduce the bad behaviour.
If we get more complaints later then I'll work that out.
Closes #7317 and also Closes #7311 (I'm getting all three installs, before if you ordered wrong you missed one)