Repository navigation
Commit ca0aac5
Integration: wave F on main a8aeeb2 -- cross-lane reconciliations, floors, anchors (INT16)
The 84 picks of lanes F12, F6, F3, F4, F5, F1, F14, F7, F9, F10 and F2 each
carry their own -x line. This commit is what only the merged tree can decide.
integration (i), F1 x F5, src/source/pine_scheduler_native.cpp: F5's
scheduler_update_session_state selects the kernel's session-day facts and
reads neither argument, so F1's retained-input lookahead that fed the old rule
(next_aggregated_script_open and next_script_open, the block computing
next_script_open_ms) is deleted, and the writer takes no parameters
(pine_strategy_host.hpp/.cpp). F1's COOF timing is kept verbatim: the close
callback asks only when no recalculation published the bar, a recalculation
asks at the bar's first publication. The design rows OT7 and AG3 no longer
cite the deleted helpers.
integration (ii), F7's orphans outside its files: the declarations of
submit_fx_margin_slice, carried_pooc_short_margin_before_script_scope,
carried_pooc_short_priced_exit_after_adverse_scope and
next_coof_waypoint_price (pine_adapter.hpp), and of the deleted 4-argument
entry_stop_first_touch overload (src/engine_internal.hpp), are removed.
PineScheduler::next_input_waypoint no longer restates the path-order rule: it
asks PineExecutionAdapter::source_path_uses_high_first for the input bar, the
kernel's path_high_first over the same spec path_order and the same
|H - O| < |O - L| comparison, so every answer is the copy's.
integration (iii), doc rows the lanes reported, re-read on the merged tree:
- F12's three examples: README's example sentence is count-free and names the
new hosts; ADR (twice) and docs/ci.md say eighteen / sixteen C++.
- F4's 39 C symbols: README (the count, five new table rows, three callback
layouts), index.md (104 PF_API, 39), ADR, native-engine.md; the hash hook,
the anchored-level hook, declare_auxiliary_feed and native_sized_units have
C routes now (native-engine.md, pine-to-native.md: four members excluded,
not eight).
- F3's contracts in the headers (comment-only, strip_comments identical):
pineforge.h's per-bar hash equals the scalar for every host and the
"[F3 will change: ...]" marker is gone; the close-cause derivation,
execution.hpp's CloseCause and engine.hpp's exit_from_bracket / close_cause
say an armed WaitForApplied close records Bracket through the fill.
- integration: F2's d62ff389 rewrote strategy_configure_native_ext_v1's doc
to "leaves the host Failed"; F3's dab5e8ea validates first and keeps the
handle usable, so the doc states that. F2's d9e972fc said a native bracket
leg reads SCRIPT (F3 item 1 made it BRACKET) and listed eight C-less
members (F4 left four): both pine-to-native passages restated.
- integration: F5's COVERAGE note reasoned from the struct's end; after F4's
tail the four bytes sit in the base layout's padding, as it now says, and
the decision struct note orders F5's bytes before F4's tail.
- F14's ruling: a bucket a session break clips is Confirmed, named beside the
session-close clip in native-engine.md's lookahead_off rule.
- Stale citations: native_execution_consumer.cpp cites native_run_spec.hpp:560
(F2); scripts/bar_identity_lane.py and src/source/pine_security_eval.cpp
cite timeframe.cpp's current lines (F14).
- F9: pine_adapter.hpp no longer calls placement rows immutable evidence.
- F6's finding 2: the ADR residual section describes the widened vocabulary
and the installed-header surface, P2c's paragraph says why its four names
are ruled in a residual row, the front doors say the gate reads both.
- integration, F6 x F5: check_kernel_residuals called session_ismarket_,
session_isfirstbar_ and session_islastbar_ stale (F5 moved them off
BacktestEngine); F6's residual row now rules barstate_islast_ alone.
integration, F4 x F5 tests: both lanes' sections of tests/test_native_c_api.c
defined session_state / session_bars / session_fill / session_on_bar and
SESSION_BARS; F5's section is renamed dayfact_* / DAYFACT_*, no check changed.
integration, F4 x F6: test_native_c_api_c99 reads P2c's historical spellings
on purpose, so it builds with -Wno-deprecated-declarations as test_c_abi does.
integration (iv), floors from the rows that register on the merged tree
(ctest -N, not summed): release 572 -> 586, kernel 202 -> 213 run (214
registered, the WebSocket row skips); the six CONTRIBUTING.md /
contributing-llm.md citations of the two constants carried by hand.
integration (v), F10's harness row is NOT wired: python3 -m unittest discover
-s benchmarks/tests and benchmarks/check_provenance.py pass on this host (25
tests OK; all 163 headline numbers traced), but in a depth-1 clone -- the
checkout .github/workflows/ci.yml uses -- the unit tests fail once and the
checker reports 9 errors ("pinned 933fe58:... is not in this clone"), so the
row would turn every CI job red. Wiring it needs fetch-depth 0 there first.
Doc anchors: check_doc_anchors.py --fix, then reanchor_doc_citations.py over
HEAD, fd78592, origin/main and the eleven lane tips (152 carried), then
twelve by hand; check_doc_anchors 1625/1625, check_doc_lint TOTAL 0.
No epoch moves. No behaviour change beyond the lanes': (i) and (ii) are
byte-identical by construction, (iii) is documentation and comment-only
headers.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>1 parent 3668af9 commit ca0aac5
26 files changed
Lines changed: 508 additions & 489 deletions
File tree
- docs
- adr
- design
- pages
- include/pineforge
- source
- scripts
- src
- source
- tests
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
137 | 137 | | |
138 | 138 | | |
139 | 139 | | |
140 | | - | |
| 140 | + | |
141 | 141 | | |
142 | 142 | | |
143 | 143 | | |
| |||
149 | 149 | | |
150 | 150 | | |
151 | 151 | | |
152 | | - | |
153 | | - | |
| 152 | + | |
| 153 | + | |
154 | 154 | | |
155 | 155 | | |
156 | 156 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
159 | 159 | | |
160 | 160 | | |
161 | 161 | | |
162 | | - | |
163 | | - | |
164 | | - | |
165 | | - | |
166 | | - | |
| 162 | + | |
| 163 | + | |
| 164 | + | |
| 165 | + | |
| 166 | + | |
| 167 | + | |
167 | 168 | | |
168 | 169 | | |
169 | 170 | | |
170 | 171 | | |
171 | 172 | | |
172 | 173 | | |
173 | | - | |
| 174 | + | |
174 | 175 | | |
175 | 176 | | |
176 | 177 | | |
| |||
390 | 391 | | |
391 | 392 | | |
392 | 393 | | |
393 | | - | |
| 394 | + | |
394 | 395 | | |
395 | 396 | | |
396 | 397 | | |
| |||
462 | 463 | | |
463 | 464 | | |
464 | 465 | | |
465 | | - | |
| 466 | + | |
466 | 467 | | |
467 | 468 | | |
468 | 469 | | |
| |||
480 | 481 | | |
481 | 482 | | |
482 | 483 | | |
| 484 | + | |
| 485 | + | |
483 | 486 | | |
484 | | - | |
| 487 | + | |
| 488 | + | |
485 | 489 | | |
486 | | - | |
| 490 | + | |
487 | 491 | | |
488 | 492 | | |
489 | 493 | | |
| |||
501 | 505 | | |
502 | 506 | | |
503 | 507 | | |
504 | | - | |
505 | | - | |
| 508 | + | |
| 509 | + | |
| 510 | + | |
506 | 511 | | |
507 | 512 | | |
508 | | - | |
509 | | - | |
| 513 | + | |
| 514 | + | |
510 | 515 | | |
511 | 516 | | |
512 | 517 | | |
| |||
0 commit comments