mirror of
https://github.com/jmcorgan/fips.git
synced 2026-10-06 11:38:24 +00:00
Johnathan Corgan
fe9e51316f
Five conflicts, all from the pre-v0.5.0 batch landing in handshake and rekey code that next's peer/control-machine refactor has already rewritten. How each was resolved, since none of it is mechanical: src/node/handlers/rekey.rs keeps next's comment. Both sides describe the same removal and differ only in which sibling arms they name, and next's names next's own. src/node/handlers/handshake.rs takes master's comment. next already had the `let _slot` binding and the test-build assertion below it; what it lacked was the paragraph saying why the binding name is load-bearing, which is the thing a later tidy-up would delete. src/node/tests/acl.rs keeps next's file whole. next already covers the outbound ACL reject arm with a test written against its own structure; master's variant is written against master's API and its helper does not exist here. An earlier attempt spliced the two and produced a duplicate definition and four unresolved names. src/node/tests/unit.rs keeps next's side plus one test from master's, `msg1_handler_holds_its_pending_slot_while_the_handler_runs`. The rest of master's side already exists on next and would have duplicated it. The test is placed at module level: the conflict boundary fell inside another function's signature, so taking the hunk verbatim nested it and clippy caught it as an inner test item. src/node/tests/establish_chartests.rs is removed. next has no such module; master modified it in the same commit. Gated here rather than assumed: rustfmt clean, clippy clean at -D warnings, 2381 tests passed and 0 failed, all six repo guards exit 0, and the carried-over test confirmed present in the run output.