v0.1 implementation plan: review follow-ups
For the implementing model: do not work from this file. The owner gives you one task file at a time. This file is the index for the owner and the reviewer.
Goal: close the findings of the v0 review (docs/implementer-log.md, "v0 review"): the
proxy's status recorder must never panic on a writer without Flush, and the acceptance suite
must cover a configuration file that exists but cannot be read.
How this plan was made: acceptance tests first, from the review findings and PLAN.md; no
reference implementation. Both given tests were compiled and run against master at the merge of
v0: the recorder test fails there (it panics, which is the finding), the config tests pass
(finding 6 was a gap in the suite, not in the code).
Tasks
| # | File | Delivers | Tests that define it |
|---|---|---|---|
| 01 | 01-review-fixes.md |
Flush that degrades instead of panicking; the two config tests |
internal/proxy/recorder_test.go, internal/config/unreadable_test.go |
Branch v0.1. One task, one fresh OpenCode session, one commit.
For the reviewer
git log --oneline master..v0.1: one commit with the trailer.cmpboth copied tests against_files/;git diff master..v0.1 --stat -- PLAN.md AGENTS.md docs/plansempty.make gate,make smoke.grep -rn '\.(http\.' internal/— every type assertion on a writer is checked (v, ok :=).