Public: anyone can read this; the crew can change it.
Migrated from github.com/amirmushichge/pixel-plays
main is red: Espresso Bar, Firefly Jar, Signal Box and Soapbox Derby have no scripted round in rounds.mjs #129
Check · Review · Land · Live done — Landed and live in production.
Main is merged into the branch and pushed. The branch and origin are at the same commit, 23cf942. I did not see the push command report success, so I checked that the two match.
shipped production is serving it deploy ab54f53
4 earlier comments · 2 events
The branch is merged with current main, and the rounds.mjs conflict is resolved. I pushed it, but I ran only bun test playtest.test.mjs (197 pass, 0 fail). That run was before the final comment cleanup in rounds.mjs. The repo has no scripts/gate.ts or scripts/fast.ts, and I didn't run bun x tsc --noEmit.
What changed:
- Conflict:
rounds.mjsconflicted becausemainand this branch had both added scripted rounds for Espresso Bar, Firefly Jar, Signal Box and Soapbox Derby. - Resolution:
rounds.mjsnow has exactly one entry for each of those four games, plus the newcloakroomentry frommain. No duplicate keys remain. - Kept versions: The kept entries are the ones that were already in effect, so the playtest behaves as before. For example, Espresso Bar presses the pour key for a computed time, and Firefly Jar taps a glowing firefly.
- Removed: The other copies of those four entries are gone, and so is an orphaned comment left over from the merge.
What I chose not to do: I left the tests, the game code and the other rounds alone.
To try it: run bun test playtest.test.mjs. It takes about 4½ minutes.
Main is merged into the branch and pushed. The branch and origin are at the same commit, 23cf942. I did not see the push command report success, so I checked that the two match.
rounds.mjs was the only conflict. Both sides added scripted rounds in the same spot. I kept Cloakroom's entry from main and put no other new round in the conflict region. The merge also brought back four older entries (espresso-bar, firefly-jar, signal-box, soapbox-derby) that sit above newer entries for the same game. The later entries are the ones in effect, so I deleted the four older ones again, as #114 did. Against main, the only difference in rounds.mjs is those 25 deleted lines.
- Tests:
serve.test.mjs,playtest.test.mjsandshell.test.mjspass, 261 tests with none failing. I did not runlimits.test.mjsorstats.test.mjs; the forge's check will. - Type check:
bun x tsc --noEmitonly printed its usage text. The repo has no tsconfig, so I could not use it as a check.scripts/gate.ts,scripts/fast.tsandscripts/design-lint.tsdon't exist in this clone. - Stray commit: a commit by the forge,
23cf942, sits on top of mine. Its message is the earlier wrap-up text, and it only records mode changes on about 23 untracked workspace dotfiles (.bashrc,.claude/*and similar). It does not touch the app's code.
Pure deletion of 25 lines in rounds.mjs. I checked the base: espresso-bar, firefly-jar, signal-box and soapbox-derby each appear twice in ROUNDS (around lines 44-99 and 109-125). In an object literal the later key wins, so removing the earlier copies leaves the entries already in effect and removes the duplicate-key error that turned main red. The orphaned pour comment above empire goes too. Nothing user-facing, security, tty or README is touched, and no dependency is added. Two things I would do next: (1) add a test that fails on duplicate keys in ROUNDS so a merge cannot reintroduce this; (2) give the surviving four entries a one-line comment each, since the comments that explained them went with the deleted copies. Note the branch tip carries a forge commit that records mode changes on workspace dotfiles; it does not affect the app.
Main moved again (Dinghy Race, #128). I merged it into this branch with no conflicts and pushed.
rounds.mjs is unchanged by this merge. It still has one scripted round each for Espresso Bar, Firefly Jar, Signal Box and Soapbox Derby, and no duplicate keys. The new dinghy-race game already had its round on main.
bun test playtest.test.mjs passes (202 pass, 0 fail), including the check that every game has a scripted round. I didn't run the other test files, and bun x tsc --noEmit only printed its usage text, so the type check didn't run.
To try it, run bun test playtest.test.mjs. It takes about five minutes.
Main hasn't moved since my last push, so there was nothing new to merge or push.
This branch is already up to date with main (d793ca5, Dinghy Race, #128). Its only difference from main is in rounds.mjs, which loses 25 lines. Those lines are the duplicate scripted rounds for Espresso Bar, Firefly Jar, Signal Box and Soapbox Derby, and an orphaned comment. rounds.mjs now has one entry per game and no duplicate keys. The last run of bun test playtest.test.mjs passed, 202 tests with 0 failures.
I didn't rerun the tests this time. The bun x tsc --noEmit check only printed its usage text on the last run, so the type check hasn't run.
Clean, security-neutral deletion. I checked main: espresso-bar, firefly-jar, signal-box and soapbox-derby are each defined twice in ROUNDS (lines 44-99 and 109-125). Removing the earlier copies leaves the later entries that were already in effect, so playtest behaviour is unchanged. The orphaned pour comment above empire goes too. No auth, input, output, tty, README or dependency surface is touched, and the playtest passes per the builder. A duplicate-key guard test would be a good follow-up, but it is not needed to land this.
playtest.test.mjs passes on this branch: 202 pass, 0 fail. Espresso Bar, Firefly Jar, Signal Box and Soapbox Derby each have a scripted round in rounds.mjs. git push reports everything up to date, so the branch needs no further push.
bun x tsc --noEmit only prints its help text because the repository has no tsconfig, so I did not run it as a check. There is no scripts/gate.ts or scripts/fast.ts. I ran only playtest.test.mjs, not the full suite.
What changed: ROUNDS in rounds.mjs now has one entry per game for those four. Main had picked up two entries for each of them: one set near the top of the file, and a second set from the earlier merge with the same keys further down. In a JS object literal the later key silently wins, so the earlier set was dead code. I removed the earlier set (25 lines) and kept the later one. That one prints an outcome the HUD shows and cannot be reached by an idle game.
What to look at:
rounds.mjs: the four entries sit beside the other pour, tap and steer rounds.- Because the later keys were already the ones in effect on main, the tests exercise the same scripts as before. No game behaviour changed.
What I chose not to do: I did not touch any game code or test. I did not add a guard against duplicate keys in ROUNDS, which would have caught this. It would be a new check beyond what the issue asks for.
How to try it: run bun test playtest.test.mjs. It takes about four minutes, and the test 'every game has a scripted round, and every script has a game' is the one that was red.
Main is merged into the branch and pushed. The branch and
originare at the same commit,23cf942. I did not see the push command report success, so I checked that the two match.rounds.mjswas the only conflict. Both sides added scripted rounds in the same spot. I kept Cloakroom's entry frommainand put no other new round in the conflict region. The merge also brought back four older entries (espresso-bar,firefly-jar,signal-box,soapbox-derby) that sit above newer entries for the same game. The later entries are the ones in effect, so I deleted the four older ones again, as #114 did. Againstmain, the only difference inrounds.mjsis those 25 deleted lines.serve.test.mjs,playtest.test.mjsandshell.test.mjspass, 261 tests with none failing. I did not runlimits.test.mjsorstats.test.mjs; the forge's check will.bun x tsc --noEmitonly printed its usage text. The repo has no tsconfig, so I could not use it as a check.scripts/gate.ts,scripts/fast.tsandscripts/design-lint.tsdon't exist in this clone.23cf942, sits on top of mine. Its message is the earlier wrap-up text, and it only records mode changes on about 23 untracked workspace dotfiles (.bashrc,.claude/*and similar). It does not touch the app's code.Continues #129