be get op-summary lists orphaned (non-DAG) post-rows
be get's summary prints wtlog post-rows for commits NOT reachable from the resulting branch tip, so it can show two "JS-025" posts while be log (the DAG) shows one — confusing. Surfaced in the JS-027 worktree from an orphaned sibling commit (an earlier botched + rolled-back commit). Method: Issues.
be get printed post ?9718389d#JS-025… AND
post ?371ad14b#JS-025…, while be log: shows only 371ad14b. Confirmed: 9718389d is NOT an ancestor of trunk tip 371ad14b — a sibling orphan (both children of b9c6b647); its diff is ONLY test/all/02-extension-dispatch/run.sh (a botched partial commit, see JS-025 Outcome).
be get's op-summary reads the wtlog (operations this wt performed); be logwalks the branch DAG. They legitimately differ, but surfacing a non-DAG post indistinguishably from a landed commit reads as "the branch has 2 commits".
be get's summary should mark or omit post-rows not reachable from the new tip(or otherwise distinguish op-log from DAG), so it can't be misread as branch history. Decide the intended op-log-vs-DAG semantics and make the views agree.
be get / sniff summary path; NOT a data bug (theorphan is a real wtlog row). Also decide whether orphaned local commits should be GC'd. Related to DIS-022 (be log empty in detached/submodule wt).
reachable from the new tip is not mentioned at all in the get banner.
miss (= behind ≠ rolled-back) and post (reads as forward
history); drop the base \ tgt loop in get_emit_commit_list. Forked.
be get + be log agree.
Done: test/get/56-orphan-banner-not-dag (ctest #258), red→green.
6bc22109: get_emit_commit_list (sniff/GET.c) drops the
disapplied (base \ tgt) loop — only tip-reachable commits print as post; banner now agrees with be log. Repro test/get/56 (#258).