1. Time and turns. The long waits were the lower-std and build cycle, plus the clippy and it runs. In turns, I lost two to a scratchpad that did not exist yet. My first cp and git show > $TMPDIR/... failed quietly and I only saw it when git apply said the patch was missing. A third turn went on a blocked rm /x.diff, caused by the empty $TMPDIR.

2. Lookups the brief could have handed me.
   • The brief did not say that type_rules.nika and type_rules.rs were unchanged on develop, so I ran git diff --stat to find out. With that fact the patch could have been applied directly.
   • It did not say that the lowered code trips clippy::extend_with_drain. I found out when clippy failed, then looked up the allow list in lib.rs.
   • It did not say that the badge recipe needs the files staged first. checkout-index reads the index, so I had to stage and then regenerate.
   • The it module names were not given. I grepped for them, and my filter matched only 16 tests.
3. Checks run.
   • Useful: lower-std (twice), the build, cargo test --lib type_rules, the it filter run, the six tests/compiler programs with rule in the name, clippy, and pre-push.
   • Wasted or weak:
     • cargo test --lib type_rules matched 0 tests.
     • My first checkout-index ran before staging, so it did nothing useful.
     • pre-push still showed version red, which is the supervisor’s line, so I left it.
   • Not done: I did not confirm that the it filter covered the type_rules tests, and I did not re-run lower-std after the lib.rs change. I also never probed the fails-before case, since this is a move and not a fix.
4. One change. Add a “Preflight” line to the brief template: create <scratchpad>/<brief>/ first, and say whether the touched files changed on develop. For a move onto develop, also list the known clippy allows and the exact it module names.

This site is open source. Improve this page.