Intended, now documented and tested in 2f16713: only with [engine] model_role = "advises" (or without a model in classifier mode) and Jev's judgment of every file does a PR too large for the…
Fixed in 2f16713: an empty name or commit of the default branch is a configuration error that names its cause ("Gitea names no default branch" or "no commit of main"), covered by `test_the_reposito…
Answer to the review of a57d442 (ORANGE):
- ORANGE
cli.py:377-389(advisory downgrade) andcli.py:307-312(triage complete without units): intended, reasons in the threads. - YELLOW…
Not changed: Jev does get a deleted main-code file, with its removed lines, and asks security and weakens_checks (only a deleted judged path is unjudgeable). A deleted test file gives…
Not changed, intended: with Jev judging every changed file, the model only advises, also when it could not run because the prompt is too large; that is the same decision as `[engine] mode =…
Not changed: without a judged file, the changed files are docs, mechanical files, hard paths or DB schema files. Hard paths still give risky_paths, schema files are judged by [db_schema], and…
Not changed: the bullets name the two decisions a reader needs, that the guard keeps the exit code and which sums are bounded; the earlier review rounds asked for exactly this precision.
Answer to the review of f00192c (GREEN, merge held for 2 hints): both hints answered in their threads without a change and resolved. Restarting the review to release the merge.
Not changed: a broken stdout fails in the same guard as a broken file and is logged the same way; the test of the file path covers that guard. A partial write needs no distinction, since nothing…
Answer to the review of c875881 (GREEN, merge held for 2 hints):
- hint Scenario#37.02 vs. README: README completed in f00192c, resolved.
- hint Jev and LLM bounds vs. README wording: clarified…
Fixed in f00192c: the README says that a record that cannot be built is not reported either.
Clarified in f00192c: the README names the two bounds, the LLM cost and the Jev cost, each also as sum of its parts. Bounding their total would change the meaning of the bound; both stay far above…
Answer to the review of e3078d1 (GREEN, merge held for 2 hints):
- hint
werkrevisor/cli.py:715-718(stderr not tested): fixed in c875881, resolved. - hint Scenario#37.01 mixed two cases:…
Fixed in c875881: the record that cannot be built has its own Scenario#37.02; the cost bound is now Scenario#37.03.
Fixed in c875881: the failed-write test checks the stderr line. The stdout line before the failed file write is intended: Werkator keeps it in the build log, so the record is not lost.
Answer to the review of 33e46ad (GREEN, merge held for 2 hints):
- hint
UsageApiController.kt:196-199(excerpt_cost_usdunchecked withoutcost_usd): fixed in e3078d1, resolved. - hint…
Not changed: the same redactor and print to stderr run in every other log line of the run, so a failing one would have ended the review long before; guarding this one line would protect nothing.
Fixed in e3078d1: excerpt_cost_usd is read and checked also without cost_usd; the Scenario#37.02 test sends a negative, a string and a too large value with cost_usd null, red without the…