Time dependent dur rate f and mtime by mattfidler · Pull Request #991 · nlmixr2/rxode2

mattfidler · 2026-03-02T01:36:25Z

No description provided.

Remove parse-time restriction that blocked state variable references in f(cmt), rate(cmt), and dur(cmt) expressions. These expressions now fall through to the default branch which emits correct state variable references. Absorption lag-time (alag) restriction is kept since state-dependent lag is not supported (sort-phase ordering cannot handle it). Addresses RxODE#216 and RxODE#222. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Add double *__zzStateVar__ parameter to generated RxRate and RxDur function signatures, and populate state variables from __zzStateVar__ instead of NA_REAL for ode_rate and ode_dur modes. This allows state-dependent rate(cmt) and dur(cmt) expressions to access actual compartment values at runtime. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Update t_RATE and t_DUR function pointer typedefs to include a double *y parameter for passing ODE state to rate/duration functions. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Update getRate and getDur signatures to accept double *y and forward it to the RATE/DUR function pointers. Update updateDur and updateRate to pass yp through to getDur/getRate. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Replace rx->ypNA with op->inits in sort-phase calls to updateDur and updateRate (handleTurnOnModeledDuration, handleTurnOnModeledRate). Using NA state caused state-dependent rate/dur expressions to return NA, triggering badSolve=1. Using initial conditions provides valid non-NA estimates for event ordering. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

In handle_evid for EVIDF_MODEL_RATE_ON and EVIDF_MODEL_DUR_ON events, call updateRate/updateDur with the actual solver state (yp) before reading getDoseIndexPlus1. This ensures InfusionRate and infusion end-time use the correct state-dependent values at dose time. Add needSortDefines.h include and forward declarations for updateRate/updateDur to resolve the forward-reference dependency with rxode2parseGetTime.h. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…ignature and direct-index bug - Allow alag(cmt) expressions referencing d/dt(state) syntactically by removing the ALAG parse-time restriction in handleDdtRhs - Add double *__zzStateVar__ parameter to generated Lag function signature (matching Rate/Dur/F functions); state vars remain NA_REAL in Lag so state-dep alag expressions produce a clean runtime 'NA lag' error - Update t_LAG typedef from 3-arg to 4-arg (adds double *y) - Update getLag() to accept and forward double *y state pointer - All sort-phase getLag calls pass rx->ypNA so state-dep lag always produces a runtime error (not a crash) - Fix direct-index bug in handle_evid: ind->idx is an ix-array position but updateRate/updateDur expect a direct all_times index; compute _directIdx = (ind->idx >= 0) ? ind->ix[ind->idx] : ind->idx before calling updateRate/updateDur (fixes the li=0.5 regression) - Update getRate in handleSS and getLag in handleSS to pass yp Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

State-dependent f(cmt) is now allowed when combined with modeled rate (rate=-1) or modeled duration (rate=-2) events; update the two expect_error() calls that tested these cases to expect_no_error(). Also update the test description to reflect the current behavior: - rate(cmt)/dur(cmt) depending on state still error (unsupported) - f(cmt) depending on state with fixed rate still errors - f(cmt) depending on state with modeled rate/dur now succeeds Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Document the new capabilities added in this feature branch: - State-dep f(cmt) with modeled-rate/duration events now allowed - Runtime recomputation of modeled rate/duration with actual ODE state - Sort-phase uses initial conditions instead of NA for better ordering - alag(cmt) expressions using d/dt(state) now syntactically accepted Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

- Remove TMTIME restriction in parseDdt.h that blocked state variable references in mtime(var) <- expr definitions - Add double *__zzStateVar__ parameter to generated mtime C function so state values are available during model-time computation - Update t_calc_mtime typedef to accept state pointer (double *y) - Move calc_mtime() call in iniSubject to after u_inis/memcpy so initial conditions are populated in ind->solve before use - Pre-copy op->inits to ind->solve when no u_inis but nMtime > 0, ensuring state-dep mtime expressions get valid non-NA initial values - Add test: mtime(t1) <- intestine * 0.01 parses and solves correctly Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

In handleTurnOnModeledDuration and handleTurnOnModeledRate, use ind->solve (per-subject initial conditions from u_inis) instead of op->inits (global all-subject estimates) when op->neq > 0. When op->neq == 0 there is no ODE state so fall back to op->inits. iniSubject populates ind->solve via memcpy(op->inits) + u_inis before sortInd is called, so ind->solve has correct per-subject values at sort time. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…Sorted=0 - Add reSortMainTimeline(ind, startI): re-sorts ind->ix[startI..n-1] using ind->timeThread as sort key (wraps timsort SORT macro). - Add recomputeMtimeIfNeeded(rx, ind, yp, nextI): recomputes ind->mtime with current state yp and updates timeThread for changed mtime events; returns 1 if any mtime value changed. - Refactor sortInd to call reSortMainTimeline instead of inline SORT. - In handle_evid at EVIDF_MODEL_RATE_ON / DUR_ON events, set ind->mainSorted = 0 (guarded by ind->idx >= 0 to skip extra doses). - In all 6 solve loops (ind_indLin0, ind_liblsoda0, ind_lsoda0, ind_linCmt0H, ind_linCmt0, ind_dop0): add re-sort trigger at top of event loop and mtime recomputation after handleEvid1/handleSS. The re-sort refreshes timeThread for remaining stop events (using getTime_ to include lag correction) and mtime events before calling reSortMainTimeline(ind, i). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Tests verify that: 1. State-dependent modeled rate uses per-subject initial state at sort phase, producing different trajectories for subjects with different initial conditions. 2. Runtime re-sort after rate recompute places stop event correctly (no NA output, no solver errors). NEWS entries document the three bug fixes: per-subject initial conditions at sort time, main timeline re-sort after runtime rate/dur update, and mtime recomputation at dose events. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…branch) The `ode_lag` branch in codegen.c previously set all state variables to NA_REAL, causing runtime NA errors whenever a state-dependent alag() expression was evaluated. Merge it with the general else-branch so that Rate/Dur/F/Lag functions all populate state vars from `__zzStateVar__`. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Five related fixes for state-dependent rate/duration/alag interaction: 1. updateDur/updateRate: store absolute lagged stop time (lagged_start + duration) in all_times[stop_raw] instead of unlagged (t + duration). This ensures that when alag() and dur()/rate() are both state-dependent, the stop event lands at the correct absolute time. 2. getTimeCalculateInfusionTimes: RATE_OFF/DUR_OFF cases now return getAllTimes() directly (already the absolute stop time) instead of re-applying lag via getLag — which would have double-lagged stop events. 3. getTime__: add RATE_OFF/DUR_OFF guard in update=0 path to likewise return getAllTimes() directly. 4. getTimeCalculateInfusionTimes catch-all: replace rx->ypNA with ind->solve so sort-time lag uses per-subject initial state. 5. handleInfusionItem: replace rx->ypNA with ind->solve for amt>0 and tB computations so constant-alag infusion start/stop times are computed with the per-subject initial state rather than a sentinel NA array. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…igger Three changes across all 6 solve loops and one new helper: 1. refreshLagTimesIfNeeded(): new helper that, when needSortAlag is set, recomputes timeThread entries for remaining bolus/ON dose events using the current state yp via getLag(). Skips RATE_OFF/DUR_OFF (store absolute times) and all EVIDF_INF_RATE events (stop key = lagged_start + f*dur, not raw_stop + lag — handled at sort time via handleInfusionItem). Returns 1 if any time changed; callers set mainSorted=0 to trigger re-sort. 2. xout computation: changed from getTime_(ind->ix[i], ind) (which called getLag with rx->ypNA) to ind->timeThread[ind->ix[i]] (precomputed correct lag-adjusted value from sortInd / runtime refresh). 3. Re-sort trigger: stop event refresh now uses getAllTimes(ind, _raw) instead of getTime_(_raw, ind), since stop events already store absolute lagged times after commit fix(getTime). 4. All 6 loops call refreshLagTimesIfNeeded after recomputeMtimeIfNeeded, setting mainSorted=0 when lag-adjusted times change. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Three tests covering the new state-dependent lag functionality: - Sort phase uses actual initial state (not NA) for alag(cmt) <- state expr - dur(cmt) + constant alag: stop event placed at absolute lagged time - Multiple doses: runtime lag refresh produces no NAs across dose events Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Document four fixes: - State-dep alag() now uses actual state values (not NA) at sort phase - dur(cmt)/rate(cmt) stop events store absolute lagged times correctly when combined with alag() - xout uses precomputed timeThread values (correct lag at each event) - Runtime lag refresh via refreshLagTimesIfNeeded re-sorts timeline when state-dep lag values change between events Note cache invalidation: cached compiled models must be rebuilt after installing this version. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…ation 1. Skip evid==3 (reset events): getWh(evid=3) gives cmt=-1, which causes _alag[-1] array underrun (UB) inside RxLag. 2. Zero ind->curShift before the scan loop, matching the sortInd convention. After a reset event curShift = -maxShift; without zeroing, getLag returns rawTime + lag + maxShift instead of rawTime + lag, producing a spurious 'changed' flag and a corrupted event timeline. 3. Set ind->idx = j (the ix[] position) inside the loop before calling getLag/_update_par_ptr. getValue() in approx.cpp reads y[ind->ix[ind->idx]], so the covariate (e.g. lagt) is looked up at the sorted position of the dose event being processed — not at the position of the preceding observation event (which has lagt=0). All three changes save and restore the affected fields after the loop so no side-effects leak into subsequent event handling. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

mattfidler · 2026-03-09T11:50:28Z

@john-harrold the last commit made me laugh -- it said I fixed everything that I broke, but it had just forgot the context. It was "justifying" itself for only fixing some of the bugs it was introducing.

Both liblsoda (C) and dlsoda (Fortran) refuse to integrate when |tout - t| < 2*UROUND*max(|tout|,|t|) with the error "tout too close to t to start integration". The prior isSameTime threshold was only 1*DBL_EPSILON, creating a gap where the check would pass the guard but the solver would reject the step. For id=509 in nmtest (SS infusion with lag and modeled rate), the EVIDF_INF_RATE stop-event time computed via handleInfusionItem differed from the solver's current xp by ~1.5 ULPs at t≈12 and t≈31.28, which fell in the 1x–2x epsilon gap. Widening isSameTime to 2*DBL_EPSILON eliminates these spurious "too close" failures. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

State-dependent lag/F/dur/rate/mtime expressions (including those using d/dt(state) terms) now parse successfully and compile, so the corresponding tests should expect success rather than failure. Changes: - First group (rxode2): mtime(z)=d/dt(x)+3 now compiles fine - Second group (rxode2parse): {f,F,alag,lag,rate,dur}(depot)=d/dt(central)+3 and mtime(z)=d/dt(x)+3 now parse successfully Jacobian-dependent expressions (df(x)/dt(ETA[1])) remain invalid and their badParse tests are unchanged. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Add mtime0[90] to rx_solving_options_ind to hold the initial sort-time mtime values. recomputeMtimeIfNeeded() now only re-evaluates mtime[k] when the solver is at ind->mtime0[k] (guarded by isSameTime), then marks mtime0[k] = R_NegInf to prevent double-fire. Previously mtime was re-evaluated after every solver step, converging to a fixed-point T = expr(state(T)). Now it fires at T0 (the initial sort time computed from initial state) with one optional update when the solver reaches T0. Also update test-mtime.R: use sparse sampling grids that exclude mtime fire times (so the mtime event adds its own time point), and rewrite test 4 to verify fires at T0 not the old fixed-point Tfp. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…te-f

Copilot

Pull request overview

This PR enables and validates state-dependent expressions for dosing modifiers (alag, rate, dur, f) and model times (mtime) by passing the current ODE state into generated helper functions, recomputing timing at runtime where needed, and re-sorting event timelines to preserve correct temporal ordering.

Changes:

Allow parsing and execution of state-dependent mtime(), alag(), rate(), dur(), and f() expressions (including dependence on d/dt(...)).
Update solver internals to compute event times from timeThread, recompute lag/mtime at runtime, and re-sort timelines when timing changes.
Add extensive test coverage for state-dependent behavior across alag, rate, dur, f, sorting, and mtime.

Reviewed changes

Copilot reviewed 22 out of 22 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
tests/testthat/test-state-dep-sort.R	Adds regression tests ensuring sort-phase uses per-subject initial state for state-dependent rate.
tests/testthat/test-state-dep-rate.R	Adds analytical tests verifying state-dependent modeled rate behavior and infusion stop-times.
tests/testthat/test-state-dep-lag.R	Adds tests for state-dependent lag and NA regressions, including multi-dose refresh.
tests/testthat/test-state-dep-f.R	Adds analytical tests for state-dependent bioavailability across single/multi-dose scenarios.
tests/testthat/test-state-dep-dur.R	Adds analytical tests verifying state-dependent modeled duration and correct stop-times.
tests/testthat/test-parsing.R	Updates parsing expectations to allow state-dependent `mtime` and dosing modifiers.
tests/testthat/test-mtime.R	Adds tests for state-dependent `mtime` scheduling and runtime behavior.
tests/testthat/test-model-rate.R	Updates expectation: `f(state)` now allowed with modeled rate/duration infusions in specific cases.
tests/testthat/test-alag.R	Removes a test that previously enforced runtime errors for state-dependent `alag`.
src/parseDdt.h	Removes parse-time restrictions preventing state-dependent `mtime`/`alag`/`rate`/`dur`/`f`.
src/par_solve.h	Moves `calc_mtime` timing and adds `mtime0` capture + conditional final `sortInd`.
src/par_solve.cpp	Adds runtime recomputation for `mtime` and lag, timeline re-sort utilities, and fixes an incorrect `std::copy`.
src/linCmt.cpp	Widens `isSameTime` tolerance slightly.
src/handle_evid.cpp	Includes get-time header to enable runtime update helpers.
src/codegen.c	Extends generated Lag/Rate/Dur/mtime signatures to accept state array and populates state vars from it.
inst/include/rxode2parseStruct.h	Extends function pointer typedefs and adds `mtime0` + `mainSorted` fields.
inst/include/rxode2parseHandleEvid.h	Adds runtime update hooks for modeled rate/duration and flags timeline re-sort.
inst/include/rxode2parseGetTime.h	Threads state (`y`) through lag/rate/dur helpers and fixes stop-time to use lagged start.
inst/include/rxode2.h	Widens `isSameTime` tolerance slightly.
R/summary.R	Adjusts `summary.rxC()` to return the C text invisibly.
NEWS.md	Documents new/changed behavior for state-dependent timing and parsing.
CLAUDE.md	Updates tester workflow commands and adds note on viewing generated C code.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

src/par_solve.h

src/par_solve.cpp

NEWS.md

tests/testthat/test-parsing.R

tests/testthat/test-mtime.R

CLAUDE.md

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

mattfidler · 2026-03-17T04:19:38Z

The f(state) <- d/dt(state) needs to bring back a proper parsing error.

mattfidler · 2026-03-17T04:20:34Z

@copilot modify claude.md for copilot instructions md file

Copilot · 2026-03-17T04:20:40Z

@mattfidler I've opened a new pull request, #996, to work on those changes. Once the pull request is ready, I'll request review from you.

This reverts commit b3ed59d.

Copilot

Pull request overview

This PR extends rxode2’s event-time machinery to correctly support state-dependent dosing properties (alag, modeled rate/dur, f) and model-time events (mtime) by evaluating them against the appropriate ODE state and re-sorting timelines when event times change.

Changes:

Update generated model helper callables (lag/rate/dur/mtime) and the solver/event pipeline to pass real state vectors, recompute state-dependent times at runtime, and re-sort the main timeline when needed.
Allow additional previously-rejected combinations (notably f(cmt) with modeled rate=-1 / rate=-2) and add .rxLastCompileSuccess() to distinguish parser failures from compiler failures in tests.
Add comprehensive test coverage for state-dependent alag, rate, dur, f, mtime, and sorting/resorting behavior; update NEWS and contributor docs.

Reviewed changes

Copilot reviewed 23 out of 23 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
tests/testthat/test-ui-rename.R	Switch a couple of tests from `%>%` to base `
tests/testthat/test-state-dep-sort.R	New tests exercising per-subject init state in sort and runtime re-sort behavior.
tests/testthat/test-state-dep-rate.R	New analytical tests for modeled `rate(cmt)` depending on state.
tests/testthat/test-state-dep-lag.R	New tests for state-dependent `alag()` (sort-time + runtime refresh) and dur+alag interaction.
tests/testthat/test-state-dep-f.R	New analytical tests for state-dependent `f(cmt)` including multi-dose behavior.
tests/testthat/test-state-dep-dur.R	New analytical tests for modeled `dur(cmt)` depending on state.
tests/testthat/test-parsing.R	Track last compile success across expected parser errors; adjust some test descriptions.
tests/testthat/test-mtime.R	New tests for state-dependent `mtime()` scheduling semantics.
tests/testthat/test-model-rate.R	Update expectations: allow `f(state)` with modeled rate/duration; keep other error cases.
tests/testthat/test-alag.R	Remove a test expecting runtime errors for state-dependent `alag()` (now supported).
src/par_solve.h	Initialize mtime using per-subject initial state; adjust sorting behavior between solve/LHS phases.
src/par_solve.cpp	Add mtime requeue + lag refresh + main timeline re-sort support across all solve loops.
src/linCmt.cpp	Relax `isSameTime()` tolerance.
src/handle_evid.cpp	Include getTime header to access updateRate/updateDur helpers.
src/codegen.c	Change generated Lag/Rate/Dur/mtime signatures and populate state from provided state vector.
inst/include/rxode2parseStruct.h	Add `mtime0` and `mainSorted`; update callable typedef signatures.
inst/include/rxode2parseHandleEvid.h	Add sorting-related includes and runtime recompute hooks for modeled rate/dur.
inst/include/rxode2parseGetTime.h	Pass state into Lag/Rate/Dur; store absolute lagged stop-times; adjust time retrieval.
inst/include/rxode2.h	Relax `isSameTime()` tolerance.
R/summary.R	Make `summary.rxC()` return the generated code text (invisibly) while printing it.
R/rxode2.R	Add `.rxLastCompileSuccess()` and track compilation success state for tests/debugging.
NEWS.md	Document the new/changed state-dependent timing semantics and related fixes.
CLAUDE.md	Update local test/check commands to remove stale object files first; add tip for printing generated C.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

src/codegen.c

src/par_solve.cpp

tests/testthat/test-parsing.R

NEWS.md

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

mattfidler and others added 9 commits March 1, 2026 10:07

feat: add state parameter to t_RATE and t_DUR typedefs

f51ba6a

Update t_RATE and t_DUR function pointer typedefs to include a double *y parameter for passing ODE state to rate/duration functions. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

This comment was marked as outdated.

Sign in to view

mattfidler changed the title ~~Time dependent dur rate f~~ Time dependent dur rate f and mtime Mar 2, 2026

mattfidler and others added 5 commits March 2, 2026 13:40

Add NEWS entry for state-dependent mtime support

2b1228e

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

mattfidler mentioned this pull request Mar 3, 2026

Make rxode2 more thread safe #992

Open

mattfidler and others added 6 commits March 3, 2026 08:39

This comment was marked as outdated.

Sign in to view

mattfidler closed this Mar 10, 2026

This comment was marked as outdated.

Sign in to view

mattfidler reopened this Mar 12, 2026

mattfidler and others added 3 commits March 16, 2026 22:16

Merge remote-tracking branch 'origin/main' into time-dependent-dur-ra…

56dbc8d

…te-f

Merge remote-tracking branch 'origin/main' into time-dependent-dur-ra…

225942c

…te-f

mattfidler requested a review from Copilot March 17, 2026 03:52

Copilot AI reviewed Mar 17, 2026

View reviewed changes

Copilot started reviewing on behalf of mattfidler March 17, 2026 04:06 View session

mattfidler and others added 5 commits March 16, 2026 23:12

Potential fix for pull request finding

b3ed59d

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Potential fix for pull request finding

466fe35

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Potential fix for pull request finding

2097ffa

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Potential fix for pull request finding

48c9a26

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Potential fix for pull request finding

8b7b419

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Copilot AI mentioned this pull request Mar 17, 2026

Add .github/copilot-instructions.md for GitHub Copilot repository context #996

Draft

mattfidler added 6 commits March 17, 2026 14:13

Revert "Potential fix for pull request finding"

6d4779a

This reverts commit b3ed59d.

Don't use magrittr

416ccef

No magrittr

a5d432d

Fix mtime

b057073

Add test for rate() etc depending on d/dt()

287800b

Change mtime/f/alag/dur so they cannot depend on d/dt(state)

abcf155

mattfidler requested a review from Copilot March 18, 2026 02:04

Copilot started reviewing on behalf of mattfidler March 18, 2026 02:04 View session

Copilot AI reviewed Mar 18, 2026

View reviewed changes

mattfidler and others added 4 commits March 17, 2026 21:21

Potential fix for pull request finding

9de875f

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Potential fix for pull request finding

6b229bb

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Update news take out AI's struggles

b4a61a9

Fix parsing test

76c4674

mattfidler merged commit 1d0d883 into main Mar 18, 2026
5 of 8 checks passed

mattfidler deleted the time-dependent-dur-rate-f branch March 18, 2026 14:41

Conversation

mattfidler commented Mar 2, 2026

Uh oh!

This comment was marked as outdated.

mattfidler commented Mar 9, 2026

Uh oh!

This comment was marked as outdated.

This comment was marked as outdated.

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

mattfidler commented Mar 17, 2026

Uh oh!

mattfidler commented Mar 17, 2026

Uh oh!

Copilot AI commented Mar 17, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants