Document FSP lifecycle, SHM publication, backfill ordering,
causal signal-processing semantics, and deterministic tests.
Record research directions separately from implementation and
expose the package through the relative Claude skill symlink.
(this commit msg was generated in some part by `codex` using
`gpt-6` (`openai`))
Preserve the renderer's `reset_cache` result when publishing
rebuilt paths so zoom changes invalidate Qt's cached graphics.
Retain the earlier live-debug provenance for cadence, negative
volume, repaint suppression, and downsample cache repair.
Prompt-IO: ai/prompt-io/opencode/20260903T232525Z_fadab3d2_prompt_io.md
(this commit msg was generated in some part by `codex` using
`gpt-6` (`openai`))
Cache destination bounds and timestamp/value bytes on `Viz`.
Skip identical quote-driven refreshes, including stable NaNs,
while changed samples and forced history repairs still redraw.
Add a regression covering each invalidation path.
(this commit msg was generated in some part by `codex` using
`gpt-6` (`openai`))
Clamp negative trade sizes and cumulative bar volume in sampling,
and sanitize history and realtime volume in `tina_vwap()` and
`dolla_vlm()`. Provider resets must not create negative outliers
that hide the derived curve. Cover the history and tick paths.
(this commit msg was generated in some part by `codex` using
`gpt-6` (`openai`))
Bootstrap FSP output and timestamps from one source snapshot,
validate the history schema, and compare absolute SHM bounds.
Subscribe before bootstrap, replay history revisions, filter
foreign sampler notices, and append each destination step once.
Close replaced generators before signaling completion.
Use declared `Flume` sample periods so market-closure gaps cannot
be mistaken for the realtime cadence. Cover bounds, snapshot
isolation, backfill routing, sample steps, and targeted redraws.
(this commit msg was generated in some part by `codex` using
`gpt-6` (`openai`))
Export `FeedQuote` and `Tick`, describe historical and realtime
FSP yields, and preserve callable parameters through `@fsp`.
Align builtin history arrays and emit named realtime updates.
Keep the staged Tuicr review fixes. Let Numba specialize `ema()`
lazily so omitted defaults work alongside explicit arguments,
retaining nopython/nogil compilation and single-sample seeding.
Cover builtin yields and EMA argument forms with regressions.
Prompt-IO: ai/prompt-io/opencode/20260908T020619Z_fadab3d2_prompt_io.md
Prompt-IO: ai/prompt-io/codex/20260915T155229Z_fadab3d2_prompt_io.md
(this patch was generated in some part by `codex` using
`gpt-6` (`openai`))
Replace whole-linked-chart refreshes after FSP history rebuilds
with updates selected by normalized destination `NDToken` identity.
Clear each matching `Viz` formatter and max/min cache, redraw
its path, then rescale only the affected view. This keeps primary
and foreign market curves stable while near-term history is
backfilled.
Record the concrete piker `PlotItem`, `ChartView`, `Axis`, and
`YAxisLabel` relationships used by the redraw path.
Prompt-IO: ai/prompt-io/opencode/20260828T195032Z_9115d24d_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Distill the gap-overlay E2E findings into a staged audit plan for
the full PyQtGraph iface. Keep Qt and graphics real while replacing
feeds, services and provider search with deterministic seams.
Deats,
- define the headless PyQt6 env and pytest-qt fixture ownership;
- layer primitive, input, SHM, full-window and actor checks;
- guard window SIGINT, provider search and overlapping symbol boot;
- require signal/postcondition waits and lifecycle leak assertions;
- split reviewable PRs before distilling a repo-local UI skill.
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Schema checks do not exercise ctx discovery, directional decoding,
stream setup, request correlation or endpoint-finally cleanup.
Boot an isolated child actor around `remote_gap_overlays()` and
drive it through `AnnotClient`. Cancel one published request, leave
its reply queued and prove the next request discards that stale
msg.
Also,
- verify generated request IDs and remote-owner cleanup;
- document automated coverage and broker-backed manual boundaries.
Prompt-IO: ai/prompt-io/opencode/20260817T222036Z_f65df6a1_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
The keyboard loop treated every display not owning the focused
history chart as a realtime target. One `Ctrl+G` could therefore
toggle unrelated cached symbols.
Match the focused chart explicitly before selecting its timeframe.
Factor startup registration and toggle routing into testable
helpers, and use scoped PyQt6 enums throughout the input path.
Deats,
- preserve independent 60s startup layers per display;
- toggle only the focused 1s or 60s chart;
- drive the production event relay with a real `QKeyEvent`.
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
`prepareGeometryChange()` ran after the gap rects and arrow specs
were mutated. Qt could retain the item's old scene index even
though `boundingRect()` returned its new prepended-history
position.
Invalidate cached bounds and notify Qt before changing any
geometry. Cover both scene lookup positions with a real offscreen
item.
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
`PlotItem.addItem()` records overlays outside the Qt scene. Direct
scene removal leaves stale plot refs and range bookkeeping behind.
Track each annotation's owning plot and remove it through the same
API used for insertion. Exclude pixel-sized arrows from data
bounds.
Deats,
- centralize replacement, hiding and owner teardown;
- retain scene removal as a defensive fallback;
- cover real Qt registries and duplicate-FQME chart identity.
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Detect positive timestamp gaps from chart OHLCV SHM with one
vectorized pass. Route local UI and remote callers through one
`GapOverlayMngr` renderer, and key each layer by owner, chart,
FQME and timeframe so remote ctx teardown cannot remove local
graphics.
Deats,
- add tagged `SetGapOverlay` requests and typed
`GapOverlay` responses;
- keep `remote_annotate()` intact and serve its typed dialog from
a dedicated `remote_gap_overlays()` Tractor ctx;
- admit `SetGapOverlay`, `GapOverlay`, `list[str]` and `None` in
the Tractor payload spec, then strictly narrow requests to
`SetGapOverlay` and replies to `GapOverlay`;
- serialize shared-stream dialogs and correlate request IDs so
stale replies cannot skew later exchanges;
- rename `AnnotCtl` / `open_annot_ctl()` to `AnnotClient` /
`open_annot_client()` across storage and annotation callers;
- enable 60s history overlays at startup and toggle the focused
1s or 60s chart directly with `Ctrl+G`;
- guard redraws by chart, FQME and timeframe; tolerate empty
arrays and out-of-range lookups; update rect and arrow geometry;
- cover detector geometry, msgpack roundtrips, schema rejection,
typed stale-FQME errors and renamed storage seams;
- document ownership, refresh, stack landing and live
qualification.
Prompt-IO: ai/prompt-io/opencode/20260814T032716Z_2aa66c2c_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Track the untracked Claude session note with its benchmark, client
bottleneck, and follow-up optimization sketch intact.
(this commit msg was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Introduce `AnnotCtl.add_batch()` and `serve_rc_annots()` batch
handler to submit 1000s of gaps in single IPC msg instead of
per-annot round-trips. Server builds `GapAnnotations` from specs
and handles vectorized timestamp-to-index lookups.
Deats,
- add `'cmd': 'batch'` handler in `serve_rc_annots()`
- vectorized timestamp lookup via `np.searchsorted()` + masking
- build `gap_specs: list[dict]` from rect+arrow specs client-side
- create single `GapAnnotations` item for all gaps server-side
- handle `GapAnnotations.reposition()` in redraw handler
- add profiling to batch path for perf measurement
- support optional individual arrows for A/B comparison
Also,
- refactor `markup_gaps()` to collect specs + single batch call
- add `no_qt_updates()` context mgr for batch render ops
- add profiling to annotation teardown path
- add `GapAnnotations` case to `rm_annot()` match block
(this patch was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
For a ~1000x perf gain says ol' claudy, our boi who wrote this entire
patch! Bo
Introduce `GapAnnotations` in `.ui._annotate` for batch-rendering gap
rects/arrows instead of individual `QGraphicsItem` instances. Uses
upstream's `pyqtgraph.Qt.internals.PrimitiveArray` for rects and
a `QPainterPath` for arrows. This API-replicates our prior annotator's
in view shape-graphics but now using (what we're dubbing)
"single-array-multiple-graphics" tech much like our `.ui._curve`
extensions to `pg` B)
Impl deats,
- batch draw ~1000 gaps in single paint call vs 1000 items
- arrows render in scene coords to maintain pixel size on zoom
- add vectorized timestamp-to-index lookup for repositioning
- cache bounding rect, rebuild on `reposition()` calls
- match `SelectRect` + `ArrowItem` visual style/colors
- skip reposition when timeframe doesn't match gap's period
Other,
- fix typo in `LevelMarker` docstring: "graphich" -> "graphic"
- reflow docstring in `qgo_draw_markers()` to 67 char limit
(this patch was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
NOTE, this reversion was discovered as needed by @goodboy after
extensively manually testing the new zoom-by-font-size feats introduced
alongside macOS support.
Use class-body `if _friggin_macos:` branching to
conditionally define `size_to_values()` for both
`PriceAxis` and `DynamicDateAxis` — macOS gets the
new `_updateWidth()`/`_updateHeight()` + geometry
recalc path, other platforms fall back to the
original `setWidth()`/`setHeight()` calls.
Deats,
- add `platform` import and module-level
`_friggin_macos: bool` flag.
- `PriceAxis.size_to_values()`: macOS branch calls
`_updateWidth()` + `updateGeometry()`; else branch
uses `self.setWidth(self.typical_br.width())`.
- `DynamicDateAxis.size_to_values()`: macOS branch
calls `_updateHeight()` + `updateGeometry()`; else
uses `self.setHeight(self.typical_br.height() + 1)`.
- reorder imports: `platform` before `typing`.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Refine zoom methods in `MainWindow` and font helpers
in `_style` to return `px_size` up the call chain and
log detailed zoom state on each change.
Deats,
- make `_set_qfont_px_size()` return `self.px_size`.
- make `configure_to_dpi()` and `_config_fonts_to_screen()`
return the new `px_size` up through the call chain.
- add `font_size` to `log.info()` in `zoom_in()`,
`zoom_out()`, and `reset_zoom()` alongside
`zoom_step` and `zoom_level(%)`.
- reformat `has_ctrl`/`_has_shift` bitwise checks and
key-match tuples to multiline style.
- comment out `Shift` modifier requirement for zoom
hotkeys (now `Ctrl`-only).
- comment out unused `mn_dpi` and `dpi` locals.
Also,
- convert all single-line docstrings to `'''` multiline
style across zoom and font methods.
- rewrap `configure_to_dpi()` docstring to 67 chars.
- move `from . import _style` to module-level import
in `_window.py`.
- drop unused `screen` binding in `boundingRect()`.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Again a patch (vibed) from our very own @dnks
(just a commit msg reworking using his new `/commit-msg` skill added by
@goodboy B)
Deats,
- add `Axis.update_fonts()` to recalculate tick font, text offset,
bounding rect and `pyqtgraph`'s internal text-width/height tracking
after a zoom change; store `_typical_max_str` at init for later reuse.
- rework `PriceAxis.size_to_values()` and
`DynamicDateAxis.size_to_values()` to use pyqtgraph's
`_updateWidth()`/`_updateHeight()` with `updateGeometry()` instead of
raw `setWidth()`/ `setHeight()` so auto-expand constraints are
respected.
- fix `GlobalZoomEventFilter` to mask out `KeypadModifier` and
explicitly require both Ctrl+Shift, letting plain Ctrl+Plus/Minus pass
through to chart zoom.
- add `_update_chart_axes()` to walk all plot-item axes during
`_apply_zoom()` and call `splits.resize_sidepanes()` to sync subplot
widths.
(this commit msg, and likely patch, was generated in some part by
[`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Add `Ctrl+Shift+Plus/Minus/0` shortcuts for zooming all
UI widget font sizes via a `GlobalZoomEventFilter`
installed at the `QApplication` level.
Deats,
- `.ui._window`: add `GlobalZoomEventFilter` event
filter class and `MainWindow.zoom_in/out/reset_zoom()`
methods that reconfigure `DpiAwareFont` with a
`zoom_level` multiplier then propagate to all child
widgets.
- `.ui._style`: extend `DpiAwareFont.configure_to_dpi()`
and `_config_fonts_to_screen()` to accept a
`zoom_level` float multiplier; cast `px_size` to `int`.
- `.ui._forms`: add `update_fonts()` to `Edit`,
`Selection`, `FieldsForm` and `FillStatusBar` for
stylesheet regen.
- `.ui._label`: add `FormatLabel.update_font()` method.
- `.ui._position`: add `SettingsPane.update_fonts()`.
- `.ui._search`: add `update_fonts()` to `CompleterView`
and `SearchWidget`.
- `.ui._exec`: install the zoom filter on window show.
- `.ui.qt`: import `QObject` from `PyQt6`.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Note since it's actually `xonsh` code run with either,
- most pedantically: `xonsh ./snippets/calc_ppi.xsh`
- or relying on how shebang: `./snippets/calc_ppi.xsh`
* an sheboom.
For wtv reason on nixos importing `pyqtgraph` first is causing `numpy`
to fail import?? No idea, but likely something to do with recent
`flake.nix`'s ld-lib-linking with `<nixpkgs>` marlarky?
- set `QT_USE_PHYSICAL_DPI='1'` env var for Qt6 high-DPI
* we likely want to do this in `piker.ui` as well!
- move `pxr` calc from widget to per-screen in loop.
- add `unscaled_size` calc using `pxr * size`.
- switch from `.availableGeometry()` to `.geometry()` for full
bounds.
- shorten output labels, add `!r` repr formatting
- add Qt6 DPI rounding policy TODO with doc links
(this patch was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
A one-day MNQ schedule reply can start in the next session,
leaving the gap's earlier trading hours uncovered. Widen
`Client.history_schedule()` lookbacks up to 31 days, keeping the
requested end fixed and sharing one timeout across attempts.
Preserve unknown coverage when widening cannot cover the range.
Share `SECONDS_PER_DAY` between query and reservation arithmetic,
with one explanation of elapsed versus calendar/session days.
Replay the captured MNQ replies through the allocator: reserve
51,824 candidate slots while excluding the maintenance closure.
Cover bounded retries and timeout cancellation, and document the
query contract. This does not repair existing SHM or parquet.
Prompt-IO: ai/prompt-io/codex/20260924T232405Z_441358e2_schedule_coverage_prompt_io.md
(this patch was generated in some part by `codex` using
`gpt-6` (`openai`))
Interrupted backfill can leave stored observations separated by
missing trading time. Query dated IB sessions before prepending
stored history and reserve synthetic samples only during covered
open sessions. Keep venue closures compressed and preserve the
indexes of already published rows.
Deats,
- Add `Client.history_schedule()` with UTC coverage, extended hours,
bounded duration, timeout, and rejection of ambiguous DST times.
- Keep schedule normalization in `venues` and allocation in
`.tsp._reserve`, with the project AGPL headers.
- Expose schedule and synthetic-range queries through `datad.ib`.
- Bound reservations by available SHM capacity and inspect only the
newest 32 gaps by default via `newest_gaps`. Zero disables
queries; leave unknown and oversized ranges unresolved.
- Record synthetic ranges separately from OHLCV. Automatic storage
writes retain provider observations; placeholders copy close.
- Group the new tests under `tests/providers`.
- Cover request formatting, DST, failure and capacity cases, plus
real SHM/NativeDB startup placement with a fake provider.
- Document a common CLI/chart segment-repair interface and defer
coordinated insertion, repair execution, and reader integration.
Validation: 55 targeted tests passed in the working tree; live IB
schedule coverage remains unqualified. After the module moves,
33 affected tests passed again; the local review follow-up passed
18 affected cases, including four new request-budget cases.
The commit runner repeats
lint and isolated regression checks against this exact boundary.
Prompt-IO: ai/prompt-io/codex/20260924T185456Z_1da9b760_schedule_prompt_io.md
(this patch was generated in some part by `codex` using
`gpt-6` (`openai`))
Extract `_copy_close_to_null_rows()` with explicit mutation
bounds, searchable close-copy comments, and preserved activity
fields. Type and document `maybe_fill_null_segments()` including
its distinction from packed timestamp-gap repair.
Disable relative epoch tolerance when validating synthetic
stamps: an overfull reservation must not duplicate the right
boundary. Cover modern epochs, typed empty replies, packed gaps,
and unchanged rejected inputs.
Prompt-IO: ai/prompt-io/codex/20260924T165757Z_1da9b760_prompt_io.md
(this patch was generated in some part by `codex` using
`gpt-6` (`openai`))
Complete HMDS no-data replies without silently retrying a day
older. Let `open_history_client()` classify the original request
and leave frame stepping to `start_backfill()`.
Distinguish `request_end_dt` from `oldest_shm_dt` so the first
frame after an empty interval retains its unwritten end bar.
Inline the storage time cutoff and drop the unused two-direction
`diff_history()` helper. Explain both dates and filtering rules.
Add request-boundary tests and real SHM/NativeDB regressions:
all 2,004 supplied bars survive, while the closure stays empty.
Prompt-IO: ai/prompt-io/codex/20260923T164656Z_efa72980_prompt_io.md
(this patch was generated in some part by `codex` using
`gpt-6` (`openai`))
- preserve dirty main-tree work while transferring branch ownership
- hold provider, FQME, timeframe, and symptom fixed while debugging
- capture read-only baselines before disposable live runs
- turn each reproduced failure into deterministic coverage
- qualify interruption and restart behavior after the normal lifecycle
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Lay out a correctness-first path from deterministic regressions
through normalized gap detection, structured backfill completion,
and crash-safe `NativeDB` persistence.
Deats,
- separate checker, storage, provider, and annotation gates
- record confirmed current defects and data invariants
- define commit boundaries and timeout-wrapped verification
Prompt-IO: ai/prompt-io/opencode/20260721T215055Z_c6ec3d41_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Override `NonShittyWrapper._setTimer()` to compute packet silence
in seconds from integer epoch nanoseconds, avoiding upstream's
`datetime` subtraction error after `tcpDataArrived()`. Delegate
pre-traffic datetime and sentinel handling upstream.
Preserve the packet timestamp on expiry and clear the timer before
`timeoutEvent` emission so listeners can safely rearm it. Add three
offline regressions under `tests/providers/` for rescheduling, single
expiry, datetime and sentinel states, disabling, and reentrant
listener rearming. Document the wrapper overrides on the class.
This fixes an explicitly armed packet-idle timer; heartbeat, default
arming, and reconnect behavior are outside this change.
Prompt-IO: ai/prompt-io/opencode/20260924_ib_idle_timer.md
(this patch was generated in some part by `opencode` using
`gpt-6-astra` (`openai`))
Propagate the authoritative `disconnectedEvent` as a
`ConnectionError` so proxy callers and trade consumers fail loudly
instead of waiting on a dead socket. Preserve the last `apiError`
text as the reason without treating numeric farm/order errors as
socket failures.
Deats,
- Disconnect the one-shot callback before sending the exception.
- Raise when `recv_trade_updates()` loses its socket, and relay
disconnect exceptions through both Trio-side consumers.
- Exercise real `eventkit` error/disconnect callbacks and preserve
the supplied commission fill while testing bridge termination.
- Record the aborted integration stage as historical provenance;
its test passes do not validate this post-abort boundary.
Prompt-IO: ai/prompt-io/opencode/20260916_ib_connectivity_integration.md
(this patch was generated in some part by `opencode` using
`gpt-6-astra` (`openai`))
Keep each commission callback's exact `Fill` paired with its
report, so overlapping executions on one contract cannot borrow
each other's fees. Cache positions by account and `conId`, queue
commissioned fills by execution ID until a position arrives, and
accept zero commissions. Reject mismatched execution/report IDs.
Keep `Execution.execId` authoritative when flattening ledger
entries; an empty commission ID must not erase a valid fill. Add
regressions for empty commission IDs, reversed fee arrival, and
positions arriving before or after overlapping executions.
The human live-tested fills and reports the accounting fix works.
Exact-boundary automated checks remain pending; the earlier nine
passes covered the mixed working tree. Included Prompt-IO logs
retain whole-session context, including separate roaming work.
Prompt-IO: ai/prompt-io/opencode/20260916T194052Z_749ca1f0_prompt_io.md
(this patch was generated in some part by `opencode` using
`gpt-6-astra` (`openai`))
Normalize structured and raw IB errors without indexing strings as
message dictionaries.
Deats,
- match each error form before request-id handling;
- parse `[code N]` and `[Errno N]` prefixes without crashing;
- cover raw errors through `deliver_trade_events()`.
Prompt-IO: ai/prompt-io/opencode/20260828T063126Z_1606122b_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
IB returns trading samples, not one row per wall-clock period.
Converting elapsed time to a physical prepend offset left 42,720
zero rows in NVDA's 60s buffer even though both valid margins
shared the same timestamp.
Use the actual earliest provider sample as the NativeDB boundary.
After reverse retrieval, prepend older stored rows directly beside
it.
Deats,
- start reverse insertion at the current first index;
- update `ShmArray._first` after every provider prepend;
- trim provider/NativeDB overlap by its timestamp;
- publish the final storage prepend to chart and FSP consumers;
- derive the source `time` field from `ohlc_key_map`;
- cover the full lifecycle with a real multi-frame `ShmArray`.
Prompt-IO: ai/prompt-io/opencode/20260805T195001Z_f0f3c806_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Make `store ls` show every NativeDB FQME and its available sample
periods by default. Keep `store series` as the detailed view of
each exact sub-series, including its period, size and path. Share
the same optional FQME filter and numerically ordered payload.
Use `store shm` as the canonical volatile-buffer ep and make its
default operation a read-only report. Enter the prior persistence
workflow only through `--write-parquet`, require that flag before
`--reload-parquet-to-shm`. Reject read-only formatting options in
repair mode instead of silently ignoring them.
Also,
- update the qualification docs, debug helper and harness refs;
- cover no-flag listings, command help and SHM mode validation;
- retain exact generation and object-name selection for stale SHM.
Prompt-IO: ai/prompt-io/opencode/20260801T015321Z_74ea2152_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Make bare storage groups and eps render help, so the CLI is
discoverable without memorizing `Typer` option conventions. Add
read-only `series` and `shm` eps for durable `NativeDB` series and
live history buffers, with `Rich`-table or `JSON` output.
Deats,
- preserve exact `(fqme, period)` identities for `NativeDB`
series;
- parse exact `(generation, fqme, kind)` names for SHM buffers;
- reconstruct provider dtypes when attaching to live buffers;
- tolerate buffers disappearing between discovery and attach.
Also,
- report expected and observed cadence separately, plus invalid
rows, ordering defects and ranked gaps;
- add exact `ldshm --shm-name` selection and keep disabled write
and reload paths from reading undefined state or entering `pdb`;
- document the triage layers and cover help, identity parsing,
read-only diagnostics, dtype selection and `ldshm` flow.
Prompt-IO: ai/prompt-io/opencode/20260731T165215Z_0846cbd4_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Keep reverse history and stored SHM layout aligned so sparse IB
frames do not expose permanent zero-time reservations.
Deats,
- omit inclusive query endpoints from SHM but retain full storage
- serialize and bound farm resets, blanks, and cancellations
- run the final null sweep after reverse retrieval completes
- require valid timestamp boundaries for synthetic gap rows
- cover the live QQQ 653-row failure and reset interleavings
Prompt-IO: ai/prompt-io/opencode/20260729T204046Z_ad6c560f_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Do not pass unpublished or corrupt timestamp rows into dedupe,
`NativeDB` replacement writes, or live SHM reload.
Deats,
- infer cadence from finite positive observed steps
- choose the smaller observed step when frequency ties
- skip invalid or degenerate snapshots as a whole
- cover the QQQ zero-slot failure with mutation-failing spies
Prompt-IO: ai/prompt-io/opencode/20260729T185406Z_d359b1cb_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Filter known zero-initialized epoch slots before persisted history
hydrates SHM or merges with a fresh provider frame.
Deats,
- reindex the repaired read view for same-session chart publication
- treat sentinel-only files as empty history for fresh backfill
- reject new first-slot sentinels but preserve modern zero-price bars
- classify raw sentinel evidence separately in `store audit`
Prompt-IO: ai/prompt-io/opencode/20260729T041723Z_ce33deb6_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Serialize aware reverse-fill boundaries as `YYYYMMDD-HH:MM:SS`
before handing them to `ib_async`.
This avoids Gateway error 10314 from its trailing-`UTC` format and
drops the unfinished EST comparison breakpoint from `Client.bars()`.
Prompt-IO: ai/prompt-io/opencode/20260729T034106Z_922a5df8_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Drive known-bad, fresh, restart and append checks from a private
disposable NativeDB root.
Deats,
- bind child config and imports to the marked worktree root
- preserve checksum-matched reports, snapshots and case metadata
- compare timestamp coverage, seam values and persisted structure
- document manual chart runs, gap triage and the expansion matrix
Prompt-IO: ai/prompt-io/opencode/20260728T003247Z_ed85721c_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))