sganggs 5a66f48ad2 feedback4: the real-server suites' long "own battle over" waits poll in slices under the browser's CDP protocol timeout (test/e2e/client.mjs waitForFunctionLong, PROTOCOL_TIMEOUT_MS = the Client's 90 s) — puppeteer runs a page.waitForFunction's whole poll inside one Runtime.callFunctionOn, so watch-bonds.e2e subtest 3's 240 s wait (line 489) died with a bare "Waiting failed" (cause: Runtime.callFunctionOn timed out) whenever the own battle at 1× outlasted 90 s (the 0.1.4 candidate rerun: 152 s ≈ 62 s to COMBAT + 90 s; 65–67 s in isolated runs); reproduced at the same line with a 30 s protocol timeout, which passes with the slices; playtest2.real's 120 s wait had the same gap (a silent 'timeout' round); test/e2e/client-wait.test.js (slices until the deadline, other errors not retried, no Client suite with a plain wait as long as the protocol timeout); a test-harness timing flake, no product change; golden unchanged

2026-10-06 17:31:01 +08:00

..

client-wait.test.js

feedback4: the real-server suites' long "own battle over" waits poll in slices under the browser's CDP protocol timeout (test/e2e/client.mjs waitForFunctionLong, PROTOCOL_TIMEOUT_MS = the Client's 90 s) — puppeteer runs a page.waitForFunction's whole poll inside one Runtime.callFunctionOn, so watch-bonds.e2e subtest 3's 240 s wait (line 489) died with a bare "Waiting failed" (cause: Runtime.callFunctionOn timed out) whenever the own battle at 1× outlasted 90 s (the 0.1.4 candidate rerun: 152 s ≈ 62 s to COMBAT + 90 s; 65–67 s in isolated runs); reproduced at the same line with a 30 s protocol timeout, which passes with the slices; playtest2.real's 120 s wait had the same gap (a silent 'timeout' round); test/e2e/client-wait.test.js (slices until the deadline, other errors not retried, no Client suite with a plain wait as long as the protocol timeout); a test-harness timing flake, no product change; golden unchanged

2026-10-06 17:31:01 +08:00

client.mjs

feedback4: the real-server suites' long "own battle over" waits poll in slices under the browser's CDP protocol timeout (test/e2e/client.mjs waitForFunctionLong, PROTOCOL_TIMEOUT_MS = the Client's 90 s) — puppeteer runs a page.waitForFunction's whole poll inside one Runtime.callFunctionOn, so watch-bonds.e2e subtest 3's 240 s wait (line 489) died with a bare "Waiting failed" (cause: Runtime.callFunctionOn timed out) whenever the own battle at 1× outlasted 90 s (the 0.1.4 candidate rerun: 152 s ≈ 62 s to COMBAT + 90 s; 65–67 s in isolated runs); reproduced at the same line with a 30 s protocol timeout, which passes with the slices; playtest2.real's 120 s wait had the same gap (a silent 'timeout' round); test/e2e/client-wait.test.js (slices until the deadline, other errors not retried, no Client suite with a plain wait as long as the protocol timeout); a test-harness timing flake, no product change; golden unchanged

2026-10-06 17:31:01 +08:00

coop.e2e.mjs

feedback4: the real-browser e2e suites confirm the leftover-funds prompt on the Space path too (coop.e2e 4/4, real.e2e 2/2)

2026-10-05 14:28:47 +08:00

fastServer.mjs

feedback3 WJ: 高台 only for elite 歌蕾蒂娅 with HOK-Y

2026-10-04 20:31:46 +08:00