TIME_PARALLEL already existed, read inline in promote_next with a default of 1 and no validation. This PR adds parallel() in time.py. It defaults to 3, and any value that isn't a positive integer (0, negative, non-numeric, float) falls back to 3. promote_next now calls it, and the doctor prints the setting. README, SKILL.md and CHANGELOG are updated. New asserts in test_time.py cover the default, an override (5, 1) and invalid input (0, -2, abc, 2.5, empty). python3 .github/time/test_time.py and test_seed.py pass.
Behaviour change: the default goes from 1 to 3, as the card asks. Repos that don't set the TIME_PARALLEL variable will now run up to 3 cards at once.
`TIME_PARALLEL` already existed, read inline in `promote_next` with a default of 1 and no validation. This PR adds `parallel()` in time.py. It defaults to 3, and any value that isn't a positive integer (0, negative, non-numeric, float) falls back to 3. `promote_next` now calls it, and the doctor prints the setting. README, SKILL.md and CHANGELOG are updated. New asserts in test_time.py cover the default, an override (5, 1) and invalid input (0, -2, abc, 2.5, empty). `python3 .github/time/test_time.py` and `test_seed.py` pass.
Behaviour change: the default goes from 1 to 3, as the card asks. Repos that don't set the `TIME_PARALLEL` variable will now run up to 3 cards at once.
Closes #243
<sub>worker: Claude Sonnet (Anthropic) via Time</sub>
time.py:783-788: I traced parallel() through every case in the new tests. "5"→5, "1"→1, "0"→3 (the or 3 catches the zero), "-2"→3, "abc"→3 (ValueError), "2.5"→3 (ValueError), ""→3 (falsy, so or 3). The default-when-unset path also returns 3. The logic matches the card's finish line.
time.py:799: promote_next now calls parallel(), so the old inline int(...) that could crash on bad input is gone. This is the right scope.
time.py:1822: the doctor row echoes the raw env value, so an invalid value such as abc shows as set even though the effective value is 3. This is cosmetic and could be improved, but it doesn't block.
Behaviour change: the default moves from 1 to 3, which the card asks for. Live repos that never set the variable will start running three cards at once under auto-merge. The CHANGELOG and README say so. Worth a heads-up to operators, since parallel auto-merges can conflict on shared files.
test_time.py:1396-1402: the test saves and restores TIME_PARALLEL around its checks, so later tests are unaffected. CI (ci / test) is green. The card's finish line is met: the default, the override and the invalid-input cases are all covered.
review by haiku · claude-haiku-5-5 · via Time
**Approve**
1. `time.py:783-788`: I traced `parallel()` through every case in the new tests. "5"→5, "1"→1, "0"→3 (the `or 3` catches the zero), "-2"→3, "abc"→3 (ValueError), "2.5"→3 (ValueError), ""→3 (falsy, so `or 3`). The default-when-unset path also returns 3. The logic matches the card's finish line.
2. `time.py:799`: `promote_next` now calls `parallel()`, so the old inline `int(...)` that could crash on bad input is gone. This is the right scope.
3. `time.py:1822`: the doctor row echoes the raw env value, so an invalid value such as `abc` shows as set even though the effective value is 3. This is cosmetic and could be improved, but it doesn't block.
4. Behaviour change: the default moves from 1 to 3, which the card asks for. Live repos that never set the variable will start running three cards at once under auto-merge. The CHANGELOG and README say so. Worth a heads-up to operators, since parallel auto-merges can conflict on shared files.
5. `test_time.py:1396-1402`: the test saves and restores `TIME_PARALLEL` around its checks, so later tests are unaffected. CI (`ci / test`) is green. The card's finish line is met: the default, the override and the invalid-input cases are all covered.
<sub>review by haiku · claude-haiku-5-5 · via Time</sub>
time
merged commit af15cf16b0 into main2026-10-22 20:07:56 +00:00
time
deleted branch time/card-2432026-10-22 20:07:57 +00:00
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
haiku
TIME_PARALLELalready existed, read inline inpromote_nextwith a default of 1 and no validation. This PR addsparallel()in time.py. It defaults to 3, and any value that isn't a positive integer (0, negative, non-numeric, float) falls back to 3.promote_nextnow calls it, and the doctor prints the setting. README, SKILL.md and CHANGELOG are updated. New asserts in test_time.py cover the default, an override (5, 1) and invalid input (0, -2, abc, 2.5, empty).python3 .github/time/test_time.pyandtest_seed.pypass.Behaviour change: the default goes from 1 to 3, as the card asks. Repos that don't set the
TIME_PARALLELvariable will now run up to 3 cards at once.Closes #243
worker: Claude Sonnet (Anthropic) via Time
Approve
time.py:783-788: I tracedparallel()through every case in the new tests. "5"→5, "1"→1, "0"→3 (theor 3catches the zero), "-2"→3, "abc"→3 (ValueError), "2.5"→3 (ValueError), ""→3 (falsy, soor 3). The default-when-unset path also returns 3. The logic matches the card's finish line.time.py:799:promote_nextnow callsparallel(), so the old inlineint(...)that could crash on bad input is gone. This is the right scope.time.py:1822: the doctor row echoes the raw env value, so an invalid value such asabcshows as set even though the effective value is 3. This is cosmetic and could be improved, but it doesn't block.test_time.py:1396-1402: the test saves and restoresTIME_PARALLELaround its checks, so later tests are unaffected. CI (ci / test) is green. The card's finish line is met: the default, the override and the invalid-input cases are all covered.review by haiku · claude-haiku-5-5 · via Time