Public: anyone can read this; the crew can change it.
Full suite red: 3 files in check run 01m7g6jk92srkesh3cp1a4hzcw #13573
The real failure was the talk page's script byte cap. It is now 15,400 bytes, up from 15,200, and the other two files were timing flakes, so I left them alone. What changed test/talk-page.test.ts: the talkJs cap in "the page's own bytes are held" was 15,200.
stopped closed without shipping deploy ca1c170
2 earlier comments
The real failure was the talk page's script byte cap. It is now 15,400 bytes, up from 15,200, and the other two files were timing flakes, so I left them alone.
What changed
test/talk-page.test.ts: thetalkJscap in "the page's own bytes are held" was 15,200. The script is now 15,292 bytes, after the merged cost-display and starter-prompt features (#13456, #13339). That is the only assertion that failed. The cap rises by 200, enough to cover those features with about 100 bytes to spare. The CSS cap stays at 4,300 and the test still fails if the script bloats further.
What I chose not to do
test/llms-forks.test.tsandtest/tui-plan-row.test.tspassed when I ran them here. Their reports were over the 2 s test budget (2.22 s and 2.02 s) in a full run with four files at a time, not assertion failures. I did not add them toSLOW.- The
tui-plan-rowreport also lists the talk-page byte failure (shard 26), so the same cap change covers it. - I did not trim the script instead of raising the cap, because that would mean changing feature code for a cosmetic budget.
How to check
bun test test/talk-page.test.tspasses (62 tests), andbun scripts/gate.ts --fixandbun x tsc --noEmitran with nothing flagged.- The three files together gave 65 pass and 1 fail before the edit. I didn't run the other two again afterwards.
I merged main into the branch and pushed it. The only conflict was the talk page's script byte cap in test/talk-page.test.ts.
I took main's 15,900 over my 15,400. It is the higher of the two, so it still covers the script at 15,292 bytes. This pull now changes no code, since main already makes the same fix.
After the merge, test/talk-page.test.ts passes (65 tests) and bun x tsc --noEmit is clean. I did not run any other tests.
Because main has the fix, you may want to close #13573 rather than merge it.
Closed by the crew: superseded, its change is already on main in 575823e0.
The real failure was the talk page's script byte cap. It is now 15,400 bytes, up from 15,200, and the other two files were timing flakes, so I left them alone.
What changed
test/talk-page.test.ts: thetalkJscap in "the page's own bytes are held" was 15,200. The script is now 15,292 bytes, after the merged cost-display and starter-prompt features (#13456, #13339). That is the only assertion that failed. The cap rises by 200, enough to cover those features with about 100 bytes to spare. The CSS cap stays at 4,300 and the test still fails if the script bloats further.What I chose not to do
test/llms-forks.test.tsandtest/tui-plan-row.test.tspassed when I ran them here. Their reports were over the 2 s test budget (2.22 s and 2.02 s) in a full run with four files at a time, not assertion failures. I did not add them toSLOW.tui-plan-rowreport also lists the talk-page byte failure (shard 26), so the same cap change covers it.How to check
bun test test/talk-page.test.tspasses (62 tests), andbun scripts/gate.ts --fixandbun x tsc --noEmitran with nothing flagged.Closes #13571