Merge pull request #94 from fiddlerwoaroof/docs/ext-diff-and-ci

kb: a check that returns empty for the wrong reason passes

prolog-sessions/kb.pl

@@ -283,6 +283,10 @@ doctrine(a_guard_seeded_on_agreement_cannot_detect_drift,'lockstep.rs records co
 doctrine(a_substitution_is_a_defect_at_every_site_that_makes_it,'improvise-e5pf was filed as one command folded without a resolution. A sweep found the real fault is ONE SUBSTITUTION AT THREE DECISION POINTS: wherever the host asks layout.is_records_mode(), the client asks state.view.mode.is_records(). The wire command is only how the client REACHES the bad state; the substitution is why the bad state does damage, and it decides whether enriching the projection is even sufficient. Worse, the first site inverts its own failure mode: commit() is written to DECLINE a records commit it cannot compose, and with the mode wrong it does not decline -- it composes a typed_cell_command and a Basis claim against a records grid whose columns are [coords, +cat, measures, +], which is a WRONG WRITE rather than the intended black hole. When a ticket names a symptom, grep for the PROXY the symptom rests on; a proxy used once is used everywhere the same question is asked.').
 doctrine(comparing_a_function_against_itself_tests_nothing,'improvise-jmjz fused three grid passes into one and guarded the width half with \`one_row.col_widths == compute_col_widths(model, &app.layout, false, 0)\`. But compute_col_widths IS grid_body with an empty window -- the same function twice -- so windowing the widths degrades BOTH SIDES IDENTICALLY (both collapse to the MIN_COL_WIDTH clamp) and the assertion cannot fail. Injecting \`if drawn { widen(..) }\` left 1476 tests passing. The fixture hid it a second way: the column header \"Revenue\" is wider than the value \"300021\", so the width assertion passed on the header alone. A check whose two sides share an implementation is a tautology wearing the shape of a comparison -- and a consolidation is exactly when one gets written, because the second implementation has just been deleted. The discriminating fixture puts the widest DATA value outside the window and makes it exceed both the header and the total (values 1, 888888888, -888888888: the total cancels to 1.0 so it cannot supply the answer).').
 doctrine(a_negative_result_is_only_as_good_as_the_setup_that_produced_it,'PR #92 reported that a none_cats hazard was \"argued and unobserved at this shape\" after widening a fixture to probe it. Printing the actual axes showed Product on COLUMN in every view -- add_category defaults it there and nothing moved it -- so none_cats was {_Index, _Dim} identically everywhere and THE STATED INPUT TO THE HAZARD WAS ABSENT. The path was unprobed, reported as a negative result. This is an_injection_that_does_not_fail_names_an_untested_path applied to a SETUP rather than to an injection, and it is worse, because a recorded negative stops the next person looking. Before writing \"probed and not reproduced\", ASSERT THE PRECONDITION IN THE TEST -- print or assert the condition the hazard needs, so a fixture that drifted out of the shape fails loudly instead of reporting absence.').
+doctrine(the_elsewhere_path_is_only_as_correct_as_the_inputs_it_can_rebuild,'project::grid_in projects a view the host is NOT on. Twice now it has been wrong because it silently inherited an input computed for the host view. improvise-t6db was the first: it built with GridLayout::new, which skips prune_empty, so the other client grid never pruned -- one line to fix. improvise-ac2z is the second and is NOT local: App::rebuild_layout fills Model::formula_cache under the HOST view none_cats, and grid_in builds another view layout against that stale cache, so Total cells are keyed without the coordinate the other view puts on an axis, the lookup finds nothing, and prune_empty drops the formula columns entirely. ABSENT COLUMNS, not wrong numbers. The fix is not local because grid_in takes &App while recompute_formulas needs &mut Model. The general rule: when a function serves a context other than the one its inputs were computed for, enumerate those inputs and ask of EACH whether it was derived for the caller or for the subject -- a green suite says nothing, because every fixture where host and subject agree is vacuous for this class.').
+doctrine(a_conflicting_pr_runs_no_ci_because_it_has_no_merge_ref,'A pull_request workflow runs against refs/pull/N/merge, and GitHub cannot build that ref for a CONFLICTING pull request -- so NO RUN IS EVER CREATED and the PR simply shows zero checks. That is indistinguishable from a dropped webhook if you only look at the runs list. PR #92 sat with zero checks across two pushes and a close/reopen; the cause was that PR #91 had merged and both had appended to the same four context/ files. DIAGNOSIS ORDER: before theorising about event delivery, ask \`gh pr view N --json mergeable\` -- one call, and it answers first. The orchestrator ran a genuinely good control (push to another PR, watch a run appear in five seconds) which correctly proved delivery works repo-wide, and then drew the WRONG conclusion from a sound experiment, because the hypothesis space had two entries and the truth was a third. A control discriminates only between the hypotheses you listed.').
+doctrine(no_checks_reported_is_a_merge_state_symptom,'A pull_request-triggered workflow builds refs/pull/N/merge, and a CONFLICTING pull request HAS NO MERGE REF, so no run is ever created -- \"no checks reported\" then looks IDENTICAL to a dropped webhook. Measured on PR #92, 2026-09-21: 34 minutes of polling actions/runs?head_sha=, a close/reopen and an empty commit all produced zero runs; one git merge main produced a run in seconds. ASK gh pr view --json mergeable,mergeStateStatus FIRST -- one call answers what no amount of polling the runs API can distinguish. Both the agent and the orchestrator hypothesised a dropped event and a stalled delivery; NEITHER LISTED THE THIRD CASE, and the orchestrator control -- pushing to another PR and watching a run appear in five seconds -- is consistent with all three. A control discriminates only between the hypotheses you listed.').
+doctrine(grep_over_git_diff_is_silently_empty_under_an_external_differ,'This repo configures diff.external = difftastic, which prints a SIDE-BY-SIDE format carrying no unified-diff markers. So \`git diff | grep \"^-\"\` matches nothing REGARDLESS OF CONTENT, and an agent verifying \"nothing was dropped in the merge\" read three consecutive empty outputs as three clean bills. Verified by the orchestrator: on a file with one real deletion, git diff | grep -c \"^-\" gives 0 and git diff --no-ext-diff | grep -c \"^-\" gives 1. Use --no-ext-diff in any programmatic read of a diff, or --numstat / --name-only, which are unaffected. The general shape, and it is the sharpest form of check_the_checker: A CHECK THAT RETURNS EMPTY FOR THE WRONG REASON IS INDISTINGUISHABLE FROM ONE THAT PASSES -- and an empty result is the one outcome nobody investigates.').
 work_done('improvise-8tk',bug_fix,'ZZ/:wq quit even when the save failed, silently discarding the model. Save now aborts the effect batch on failure; save_to_current_path/1 returns Result<String,String>; three tests in ui/effect.rs.').
 work_done('improvise-907',bug_fix,'Import wizard step machine redesign: validation_error() gate at ReviewProposals (no-category and over-limit), history-stack previous() bound to Left on every step, and Done made a real confirmation step holding the built workbook. Superseded improvise-gfq, improvise-o1c, improvise-c6j.').
 work_done('improvise-6sq',wont_do,'Merging EditOrDrill and EnterEditAtCursorCmd into one EnterEdit{mode} is not safe: they differ in WHEN the cell value is read, not in a mode parameter. Added ui/app.rs::tests::i_and_o_diverge_on_an_empty_records_grid as executable evidence.').