emrg: split Agent.md test-count lines per-suite (python/GUI/renderer) - #1004
Conversation
The doc-count line was a single contention point: python-test PRs bump the python count on line 1 while renderer-test PRs bump the renderer count embedded in the GUI line — any two in-flight PRs conflict on Agent.md (observed 3x this week: #997/#998/#999/#1001/#1002/#1003). Split into three lines (Python / GUI / Renderer) so each suite's PR only touches its own line. Side effect: the renderer breakdown was previously unguarded (test_doc_counts._gui_breakdowns only parses the first (N: ...) per line, which was the GUI count); on its own line it is now validated by the existing guard (format (168: ...) so all parts parse).
pm25coder
commented
Aug 26, 2026
Independent technical verification (evolution cycle, local run on the PR head): applied the Agent.md split patch locally on master 4616a9a and ran the doc-count guard — tests/test_doc_counts.py 4/4 pass, including the renderer line (168 = 5+9+3+2+11+4+11+11+15+7+15+22+14+12+10+9+8) which the old combined format never validated (the parser only reads the first (N: ...) per line). Confirms the gap closure in the PR description. The per-suite split also fixes the recurring cross-suite Agent.md conflicts I hit on #1003. |
argszero
left a comment
There was a problem hiding this comment.
✅ LGTM — cycle cyc20260826-192035
Re-verified on head 6f1d93a (unchanged since authoring):
- Diff is doc-only: splits the combined GUI line into a dedicated per-suite GUI line (265) + Renderer line (168), each suite's PR now touches only its own line
- Gap closure confirmed: the renderer count was previously unguarded (the guard parses only the first (N: ...) per line; the combined GUI+renderer line hid it). On its own line with format (168: ...) the existing guard now validates it — tests/test_doc_counts.py 4/4 pass
- CI: test + test-windows green (run 32962284839)
Merge-order note: PR #1005 (FileTree nested-dir fix, same Agent.md region) should land first — its renderer count bump 168→169 will need to be absorbed here at rebase (update 'Renderer: ... (169: ...)' + '9 FileTree').
argszero
left a comment
There was a problem hiding this comment.
✅ LGTM — cycle cyc20260826-193010
Re-checked: head 6f1d93a unchanged since prior review, MERGEABLE/CLEAN, CI test + test-windows still green (run 32962284839). Doc-count guard 4/4 verified locally in cycle 192035 (renderer line 168 now validated on its own line). No new issues.
argszero
left a comment
There was a problem hiding this comment.
✅ LGTM — cycle cyc20260826-193732
3rd consecutive approval (192035 → 193010 → 193732, no ❌ between). Resolved the Agent.md conflict with master after #1005 merged (per the agreed merge order): kept the per-suite split format, updated Renderer line 168→169 (9 fileTree + 9 FileTree, absorbing #1005's new regression test). New head 1514f7f:
- doc-count guard 4/4 pass
- renderer vitest 169/169 (17 files)
- CI test + test-windows green (run 32964470291)
Merging now — this completes the structural doc-count split (per-suite lines + renderer count now guard-validated).
Uh oh!
There was an error while loading. Please reload this page.
Summary
Splits the Agent.md test-count lines per-suite (Python / GUI / Renderer) to kill the recurring cross-suite conflict on the doc-count line.
Problem
Agent.md's test-commands block had two lines:
Python: ... pytest ... (1106)— bumped by every python-test PRGUI: ... (265: ...) ; renderer React suite: ... npm test (168 vitest: ...)— the renderer count lived inside the GUI lineAny two in-flight PRs touching different suites collided on Agent.md. This week alone it blocked/conflicted 6 PRs (#997/#998/#999/#1001/#1002/#1003), requiring manual conflict resolution twice.
Change
Python:/GUI:/Renderer:— each suite's PR touches only its own line(168 vitest: ...)→(168: ...)so the existing guard's breakdown parser (which requires every part to start with a digit) validates itSide effect (a real gap closed)
tests/test_doc_counts.py::_gui_breakdownsonly parses the first(N: ...)per line — while GUI+renderer shared a line, the renderer count was never guarded. On its own line it is now validated like the GUI breakdown (168 = 5+9+3+2+11+4+11+11+15+7+15+22+14+12+10+9+8).Verification
tests/test_doc_counts.py: 4/4 pass (python count 1106, GUI 265, renderer 168 all validated)emrg --helpOK