Enum SimulateError 

pub enum SimulateError {
    Compile(CompileError),
    DiffsolError {
        details: String,
    },
    ToleranceNotMet,
    NotDynamic {
        details: String,
    },
    CompilerUnavailable {
        compiler: &'static str,
        details: String,
    },
    RangeRestriction {
        code: &'static str,
        message: String,
    },
    DataSourceUnbound {
        parameter: String,
        data_source: String,
    },
    Cancelled {
        details: String,
    },
    UnsubstitutableBinding {
        name: String,
        class: String,
    },
    InvalidTimeSpan {
        start: f64,
        end: f64,
    },
    InvalidParameter {
        name: String,
    },
    AmbiguousParameter {
        name: String,
        candidates: Vec<String>,
    },
    CollidingParameterKeys {
        name: String,
        keys: Vec<String>,
    },
    InvalidInitialCondition {
        name: String,
    },
    AmbiguousInitialCondition {
        name: String,
        candidates: Vec<String>,
    },
    CollidingInitialConditionKeys {
        name: String,
        keys: Vec<String>,
    },
    InvalidFieldInitialCondition {
        name: String,
        details: String,
    },
    ProviderError {
        name: String,
        details: String,
    },
}

Errors raised building an crate::problem::EsmProblem or running crate::problem::solve.

Variants

Compile(CompileError)

Wraps a CompileError raised by crate::problem::esm_problem before solving even starts.

DiffsolError

diffsol returned a solver-internal error (build failure, step failure, etc.).

Fields

details: String

The underlying diffsol error message.

ToleranceNotMet

The integrator could not satisfy the requested tolerances.

NotDynamic

crate::problem::solve was called on a EsmProblem whose document declares no differential equations, so there is nothing to integrate. Its build-time products are still readable with crate::problem::observed_field.

Fields

details: String

Why the EsmProblem carries no integrable right-hand side.

CompilerUnavailable

crate::problem::ProblemOptions::compiler named a value that is in API_SPEC §5.8’s closed vocabulary but that this binding, this build or this process cannot provide (esm-spec §9.6.6 compiler_unavailable).

Never answered by building with a different compiler: the whole point of naming one is that the caller knows which ran. details names what would have to be loaded or built.

Fields

compiler: &'static str

The vocabulary spelling the caller asked for.

details: String

What would have to be loaded or built to provide it.

RangeRestriction

A range-restricted or operator-restricted right-hand side (API_SPEC §5.8: select, halo, tile, lanes) was refused. code is one of range_halo_insufficient, range_global_locality, range_selection_writes_incomplete or range_box_invalid, and message names the rule, variable and axis involved.

Fields

code: &'static str

The diagnostic code.

message: String

What was refused, naming the rule, variable and axis.

DataSourceUnbound

A DATA-FED parameter — one whose update is kind: "data" (esm-spec §5.4, §8.5) — reached a build with NOTHING bound to it: no provider, no loaded array, and no caller-supplied p value (esm-spec §9.6.6 data_source_unbound, CONFORMANCE_SPEC §5.46).

Raised at CONSTRUCTION, before any right-hand side is built, so the answer does not depend on the compiler: this asks whether the DOCUMENT’s inputs are bound, not what a compiler can lower. A compiler that additionally cannot lower a read of the forcing channel reports that separately as CompileError::CompilerRefusedRule.

Refused rather than evaluated at the parameter’s default, which is not a fallback value but a placeholder: a forcing at its default integrates to a complete, smooth trajectory reported under the label of a quantity the document says is read from a file.

Fields

parameter: String

The flattened parameter nothing bound.

data_source: String

The data_sources key its update names. Spelled data_source rather than source, which thiserror reserves for an error’s CAUSE — a String cannot be one.

Cancelled

A build progress observer asked crate::problem::esm_problem to stop (returned Flow::Cancel).

Distinct from every other variant in that nothing went wrong: it is the caller’s own decision. It stays an ERROR — unlike a cancelled solve, which is ReturnCode::Terminated with a partial trajectory — because a half-built EsmProblem is not a usable result.

Fields

details: String

The phase and item the build stopped at.

UnsubstitutableBinding

crate::problem::remake was asked to substitute a binding the EsmProblem cannot honour without redoing part of construction.

Raised rather than silently rebuilding or silently ignoring the substitution (esm-libraries-spec.md §2.5.5): the name and the class that makes it un-substitutable are both reported so the caller knows to build a fresh EsmProblem instead.

Fields

name: String

The binding the caller tried to substitute.

class: String

The class that makes it un-substitutable.

InvalidTimeSpan

The caller supplied a time span whose start or end is not a finite number — NaN, +∞ or −∞.

A document cannot spell one: the schema’s TimeSpan bounds are JSON numbers, and esm-spec §6.6.4 records that “JSON has no infinite literal”. So this is a caller-supplied argument of a solve, validated alongside the parameter and initial-condition maps of the same call.

Refused rather than interpreted. Every ordering test against NaN is false, so a NaN endpoint reads as a run that cannot advance and would hand back the initial state as though it were a trajectory; an infinite end gives the solver loop a stop time it can never reach.

Fields

start: f64

The requested start of the interval.

end: f64

The requested end of the interval.

InvalidParameter

The user supplied a parameter name that does not appear in the flattened system.

Fields

name: String

The unknown parameter name.

AmbiguousParameter

The user supplied a key that two or more of the flattened system’s parameters carry as a DOTTED SUFFIX — a shared local name (gain for Left.gain and Right.gain) or a shared partial qualification (sub.g for Left.sub.g and Right.sub.g), esm-spec §6.6.2 rule 4. Distinct from SimulateError::InvalidParameter: the name exists, it just does not identify ONE parameter, and binding it to either candidate would be a wrong answer rather than a missing one.

Fields

name: String

The ambiguous local or partially qualified name.

candidates: Vec<String>

The qualified parameters that carry it.

CollidingParameterKeys

Two or more NON-EXACT parameter_overrides keys designate the one flattened parameter (esm-spec §6.6.2): a bare local spelling and a more-qualified one (solo and Doc.Left.solo both landing on Left.solo), or two more-qualified ones (A.M.g and B.M.g both landing on M.g). The caller wrote two overrides and only one can take effect, so picking a winner would be a wrong answer rather than a missing one. An EXACT key is never part of a collision — it identifies its parameter outright and wins over any suffix or bare claim on it.

Fields

name: String

The flattened parameter every colliding key designates.

keys: Vec<String>

The colliding keys, sorted.

InvalidInitialCondition

The user supplied an initial condition for a name that is not a state variable, or a state variable has no initial value (no entry in initial_conditions and no default on the ModelVariable).

Fields

name: String

The variable name.

AmbiguousInitialCondition

The user supplied a key that two or more of the flattened system’s states carry as a DOTTED SUFFIX (esm-spec §6.6.2 rule 4). The state-side counterpart of SimulateError::AmbiguousParameter.

Fields

name: String

The ambiguous local or partially qualified name.

candidates: Vec<String>

The qualified states that carry it.

CollidingInitialConditionKeys

Two or more NON-EXACT initial_conditions keys designate the one state element (esm-spec §6.6.2). The state-side counterpart of SimulateError::CollidingParameterKeys.

Fields

name: String

The state element every colliding key designates.

keys: Vec<String>

The colliding keys, sorted.

InvalidFieldInitialCondition

An ic(target) field initial condition could not be resolved to a per-cell value. Carries the ic(...) target state name plus a diagnostic saying why (wrong field rank, unresolvable RHS, …), so the name field stays a plain identifier.

Fields

name: String

The ic(...) target state name.

details: String

Why the initial condition could not be resolved.

ProviderError

A data provider bound in crate::problem::ProblemOptions failed to materialize its loader field, or produced the wrong number of fields for its target forcing variable.

Fields

name: String

The forcing variable the provider was bound to.

details: String

What went wrong.

Implementations

impl SimulateError

pub fn is_cancelled(&self) -> bool

Whether this is a build progress observer’s own Flow::Cancel rather than something going wrong — the counterpart of reading ReturnCode::Terminated off a Solution.

Trait Implementations

impl Debug for SimulateError

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Display for SimulateError

fn fmt(&self, __formatter: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Error for SimulateError

fn source(&self) -> Option<&(dyn Error + 'static)>

Returns the lower-level source of this error, if any.

fn description(&self) -> &str

👎Deprecated since 1.42.0:

use the Display impl or to_string()

fn cause(&self) -> Option<&dyn Error>

👎Deprecated since 1.33.0:

replaced by Error::source, which can support downcasting

fn provide<'a>(&'a self, request: &mut Request<'a>)

🔬This is a nightly-only experimental API. (error_generic_member_access)

Provides type-based access to context intended for error reports.

impl From<CompileError> for SimulateError

fn from(source: CompileError) -> Self

Converts to this type from the input type.