You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
chore(console): lower the eager-closure line by the objectstack#22044 payback once objectui resolves an @objectstack/spec release carrying 93125aeeb8 #11762
Restart-when: an @objectstack/spec release whose tag commit has 93125aeeb8 as an ancestor is published on npm, and objectui resolves it (npm latest is 17.7.0 at this write; 17.7.0's tag 4e4e881427 predates it).
Filing gate: ③ a maintainer-directed task. The maintainer's ruling A on objectui#11717 (6027998226, selection verbatim 「A 抬预算+回收卡 (Recommended)」) raised the console eager-closure line for 17.7.0, filed the payback objectstack-ai/objectstack#22044, and set the rule this card executes: "When it lands, the ceiling and BASELINE come down by the amount it recovers." Filed by objectui's domain:devx seat 2 (seat post objectui#10917, session_01XgevVHGcN8dnegiiY3v75V). ⛔ Not a claim.
Readers: objectui triage first, for grading and the hold. Then the seat that lands the bump to the release above. This card should ride that bump, or be folded into its card, because the re-measurement needs the bumped lockfile.
Upstream measurement, with this repository's instrument (apps/console/dist/eager-closure.json, objectui c0862c1c with a spec dist built from objectstack main before and after): eager closure −226,238 gzip (3,591,379 → 3,365,141); vendor-objectstack −226,301 (1,604,719 → 1,378,418). ⛔ That is the upstream seat's measurement on an injected dist. The amount this card lowers by is re-measured on the real bump.
The change (when the restart condition holds)
In scripts/check-eager-closure-budget.mjs, in one commit:
re-pin BASELINE;
lower MAX_EAGER_CLOSURE_GZIP_BYTES;
lower the vendor-objectstack rows of PER_CHUNK_BASELINE and its per-chunk ceiling;
add a "⭐ LOWERED BY THE PAYBACK" header section, in the shape of the objectui#11101 and objectstack#20646 sections already in that file.
The section names this card, objectstack#22044 and the measured chunk table (control vs head, instrument named). The line moves by the recovered amount: ⛔ it is not re-derived as baseline plus half a regression, which is the rule those two sections state. ⛔ REGRESSION_THIS_GATE_MUST_CATCH_BYTES is untouched.
Dedupe
Corpus: objectui's 31 open issues, plus the 240 issues among the 500 most recently updated closed items, listed over REST and grepped locally.
Hits:
22044: 1 (objectui#10917, the seat post);
93125aeeb8: 0;
MAX_EAGER_CLOSURE: 0;
eager closure: 6, all seat posts or closed cards (objectui#11672, objectui#11666, objectui#11101 closed; objectui#10917, objectui#10243, objectui#5748 seat posts);
Restart-when: an
@objectstack/specrelease whose tag commit has93125aeeb8as an ancestor is published on npm, and objectui resolves it (npmlatestis 17.7.0 at this write; 17.7.0's tag4e4e881427predates it).Filing gate: ③ a maintainer-directed task. The maintainer's ruling A on objectui#11717 (
6027998226, selection verbatim 「A 抬预算+回收卡 (Recommended)」) raised the console eager-closure line for 17.7.0, filed the payback objectstack-ai/objectstack#22044, and set the rule this card executes: "When it lands, the ceiling andBASELINEcome down by the amount it recovers." Filed by objectui'sdomain:devxseat 2 (seat post objectui#10917,session_01XgevVHGcN8dnegiiY3v75V). ⛔ Not a claim.Readers: objectui triage first, for grading and the hold. Then the seat that lands the bump to the release above. This card should ride that bump, or be folded into its card, because the re-measurement needs the bumped lockfile.
Why now
sharedentry grew +152 KB gzip from 17.6.0 to 17.7.0 with the same exports, and objectui's console first screen carries it (the payback owed by objectui's 17.7.0 budget raise) objectstack#22044 closedcompletedthrough PR perf(spec): a bundle that never reads the ADR-0087 conversion table stops keeping it, 226 KB gzip off the console first screen (#22044) objectstack#22048 →93125aeeb8. Its landing record is6031203181, and the objectstackdomain:specseat 2's note on objectui#10917 is6031211318.apps/console/dist/eager-closure.json, objectuic0862c1cwith a spec dist built from objectstackmainbefore and after): eager closure −226,238 gzip (3,591,379 → 3,365,141);vendor-objectstack−226,301 (1,604,719 → 1,378,418). ⛔ That is the upstream seat's measurement on an injected dist. The amount this card lowers by is re-measured on the real bump.The change (when the restart condition holds)
In
scripts/check-eager-closure-budget.mjs, in one commit:BASELINE;MAX_EAGER_CLOSURE_GZIP_BYTES;vendor-objectstackrows ofPER_CHUNK_BASELINEand its per-chunk ceiling;The section names this card, objectstack#22044 and the measured chunk table (control vs head, instrument named). The line moves by the recovered amount: ⛔ it is not re-derived as baseline plus half a regression, which is the rule those two sections state. ⛔
REGRESSION_THIS_GATE_MUST_CATCH_BYTESis untouched.Dedupe
22044: 1 (objectui#10917, the seat post);93125aeeb8: 0;MAX_EAGER_CLOSURE: 0;eager closure: 6, all seat posts or closed cards (objectui#11672, objectui#11666, objectui#11101 closed; objectui#10917, objectui#10243, objectui#5748 seat posts);payback: 2 (objectui#11101 closed, objectui#10917).scripts/gives 22 hits in the same corpus.Generated by Claude Code