Repository navigation
Conversation
A protected record set by modifiers (MSL R134a getPhase_ph) declared its fields unset; field defaults keep their scope prefix (b = 2 * a reads w.a). Assisted-by: Claude Opus 5.5
…eters A parameter bound to not Qini (MSL RSFlipFlop) was "not a constant". Assisted-by: Claude Opus 5.5
At zero flow they hold for any intermediate slope; MSL Media's test models (ShortPipe to FixedAmbient) went Unstable at the first step. Assisted-by: Claude Opus 5.5
…cord literal arguments by field MSL Spice3 Graetz: (D1.cc, D1.m_dCap) = diodeNoBypassCode(...) left D1_cc unresolved and the tuple residual unlowered; a folded record parameter went to the function as a NamedTuple. Assisted-by: Claude Opus 5.5
…as elimination -y = semiLinear(x, a, b) is y = semiLinear(-x, b, a): a massless FluidHeatFlow pump between ambients forms such a chain. Assisted-by: Claude Opus 5.5
Fields as arrays from the binding or zero-filled; element reads, writes and field access through the body's record map (MSL ReferenceAir Ice09 Gibbs). Assisted-by: Claude Opus 5.5
MSL 4 ContinuousSignalExtrema read an undefined terminal. Assisted-by: Claude Opus 5.5
MSL 4 SoftStarter: FieldError (INLINE_IF_EQUATION has no field lhs). Assisted-by: Claude Opus 5.5
…roblem MSL PumpingSystem (TriggeredTrapezoid) read initial; MSL 4 DCPowerSupply read vLim, set by its initial algorithm. Assisted-by: Claude Opus 5.5
…m parameter expressions MSL 4 lightning examples (start reading another parameter); RobotR3's BearingFriction table from tau_pos = [0, Rv0 / unitTorque; ...]. Assisted-by: Claude Opus 5.5
IF97 rho_ph := rho_props_ph(p, h, waterBaseProp_ph(...)): MeasuringTemperature's index reduction had no derivative of rho_ph (omc inlines it first). Assisted-by: Claude Opus 5.5
MSL 4 RobotR3: BoundsError in first_tstop at the event that ends its path. Assisted-by: Claude Opus 5.5
The bound value comes boxed; unbox it and splat its field tuple (MSL 4 Inverse_sh_T). Assisted-by: Claude Opus 5.5
MSL Digital MUX2x1Table[4, 4, 4] stayed a Vector{Matrix}: BoundsError at [1, 3, 1].
Assisted-by: Claude Opus 5.5
With symbolic arguments an array whose size follows the arguments was one opaque term (MSL PlanarLoops_analytic: normalizeWithAssert(n)[2], a BoundsError). Assisted-by: Claude Opus 5.5
… branch The arithmetic lowering turned the other branch's NaN into the result (MSL ContinuousMean: the Noise examples Unstable at the first step). A Boolean held as a Real is true above 1/2. Assisted-by: Claude Opus 5.5
… own operands One scale for all made it relative to the largest operand: MSL Spice3 V_pulse (PW = 1e60) kept time < TD true after TD, the inverters' input stayed at 0 V. Assisted-by: Claude Opus 5.5
…xplicit fold Folded as x := 1/y, x's other equation determined y nonlinearly and the initialization's Newton diverged (MSL 4 FluxTubes QuadraticCoreAirgap, both variants; now as omc). Assisted-by: Claude Opus 5.5
…dual initial equations give no variable a value MSL 4 SMPM_ResistiveBraking: two residuals for idq_sr were read as two values of a variable named 0.0 (refused), and one equation at a time the free parameters kept their start values. Assisted-by: Claude Opus 5.5
…, as before The parameter binding's call is then evaluated once for all its elements (a regression of f29fe9a: Borefields' impure g-function was called per element). Assisted-by: Claude Opus 5.5
…cided at the next check MSL TanksWithEmptyingPipe1/TankWithEmptyingPipe2: level_start = height, the level from the mass 2 + 4e-16 failed level <= height at t = 0, or not, by run. Assisted-by: Claude Opus 5.5
…inations A validation's reference signal folded to a parameter (MSL ShowTransferFunction's const.y) could not be read from the solution. Assisted-by: Claude Opus 5.5
From the integrator they arrived as Float64 (MSL Digital RAM's Logic tables: invalid index 1.0; now as the reference and omc). Assisted-by: Claude Opus 5.5
…one-sided partials at a domain boundary As OpenModelica's solvers: MSL MeasuringTemperature's IF97 region error at a relation crossing's re-evaluation of the step's start no longer ends the run (now as the reference). A persistent error is thrown when the solve fails. MTK path; the array path is unchanged. Assisted-by: Claude Opus 5.5
DirectRHSGeneration.jl is included before asserts.jl: the package did not load (5b467a2). Assisted-by: Claude Opus 5.5
Positional options are not Modelica (omc: no matching function); OMFrontend 1.7.x refuses them. Assisted-by: Claude Opus 5.5
A RuntimeGeneratedFunction's id is the SHA1 of its printed body and its cache is global: two builds differing only in the object got the first build's function and table. Assisted-by: Claude Opus 5.5
Assisted-by: Claude Opus 5.5
Assisted-by: Claude Opus 5.5
… at the initialization MSL Blocks.Continuous.Filter's poles were 0 and its states did not move (SMEE_Rectifier, ThreePhaseTwoLevel_PWM). Assisted-by: Claude Opus 5.5
Assisted-by: Claude Opus 5.5
Each build's tables stayed allocated for the session. Objects are told apart by a serial number, not their address, which a freed table passes on. Assisted-by: Claude Opus 5.5
A VSS model with a table in a mode did not simulate: the modes' SimCode had no functions (unqualified calls), and the module assigned only the top level's data structures and parameters. The modes now get the model's functions, the module assigns their data structures (after the parameters they read), owned by the model's build. simulateIMTK evaluates a structural module again after a translate (it kept the first one); a module evaluated again frees the replaced module's external objects (forgetBuild, or the objects in _buildAndCache). Assisted-by: Claude Opus 5.5
A model recompiled during its simulation got no functions (the frontend's cache was dropped) and was evaluated in the Runtime module, without the parameters and data structures its code reads at module level: an UndefVarError at the first Modelica function or table. The recompiled SimCode now gets the model's functions (record call sites flattened), and its code is a module Recompiled in the model's module with its functions, parameters and tables, owned by the model's build. Assisted-by: Claude Opus 5.5
_buildAndCache built the active mode's problem (a full build) and failed to put it into BUILT (a caught MethodError); the module's simulate built it again. A module without simulateFromBuild is now only evaluated. Assisted-by: Claude Opus 5.5
…anslate's passes The recompiled program was the modified class alone (no MSL), and the recompiled SimCode went through four of the SimCode passes (a MultiBody model's redundant equation stayed). Now the modified class with the classes the model was instantiated with (OMFrontend.recompilationProgram, compat OMFrontend 1.7.3), and mtkSimCode: the MTK path's SimCode pipeline, out of translate. The recompiled problem's init and reinit! run in the latest world (its module's closures: an MSL table's time-event affect). Assisted-by: Claude Opus 5.5
updateInitialConditions! matched every variable by its name without the first `_` segment (body_frame_a_f took revolute_frame_a_f's value), fixed it and replaced its attributes: a MultiBody model's recompiled system was unbalanced. Now by full name, start only; by the name without the component only when no full name matches and one old variable has it (a dynamic transition: the ball's y from the pendulum's), fixed as before. Assisted-by: Claude Opus 5.5
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes from an MSL coverage stretch (OMLibraryTesting, the failing models rerun), external objects freed with their build, VSS and recompilation, and a release workflow.
x = 1/yoriented, coupled free parameters, enumeration inputs.clearCaches!,forgetModel).overwriteCachewithout a wasted build. Recompiled models with functions and tables (a module of their own in the model's), with libraries (the frontend's program, translate's SimCode passes:mtkSimCode), start values by full name (a MultiBody model's recompiled system was unbalanced).REGISTRY_DISPATCH_TOKENsecret; else the daily run).Checked locally: tests 454/454; OM.jl 1539 pass, 6 broken; MSL validate 3.2.3 347/425, 4.1.0 366/534 (CI 314/343).
Version 0.5.5; compat OMFrontend 1.7.3 (OpenModelica/OMFrontend.jl#104, same branch).
🤖 Generated with Claude Code