- Handle empty arrays [] in std::vector<cvf::Vec3d> parser
- Only treat ReadCorruptData as error in generic template (ReadPastEnd
is expected for types like caf::FilePath that consume all input)
- Use strict mode (allowExtraCharacters=false) for Python GRPC parsing
Implements comprehensive type and range validation to prevent invalid data
from being set via Python GRPC interface. Invalid values are rejected with
clear error messages. Uses two-phase commit pattern to ensure atomic updates
where either all fields update or none do.
Validation Features:
- Type validation: Ensures field value types match (parsing errors caught)
- Range validation: Checks values against min/max constraints defined with
setRange(), setMinValue(), setMaxValue()
- Object validation: Validates cross-field constraints via validate() override
- Atomic updates: Two-phase commit ensures all-or-nothing semantics
- Rollback on error: Failed validation leaves ALL fields at previous values
- Clear error messages: Aggregates and returns all validation failures
Implementation:
- Use std::expected<void, QString> for type-safe error handling (C++23)
- assignFieldValue(): Validates after parsing, rolls back individual field on error
- copyPdmObjectFromRipsToCaf(): Two-phase commit pattern
- Phase 1: Validate all fields and collect changes (defer UI notifications)
- Phase 2: If any validation fails, rollback all fields; otherwise commit all
- UpdateExistingPdmObject(): Returns grpc::INVALID_ARGUMENT with validation details
Testing:
- Comprehensive Python tests demonstrating validation behavior
- Test atomic rollback: multiple fields rolled back when one fails
- No regression in existing functionality
Documentation:
- Added validation documentation to CLAUDE.md with examples
- Explains how to set field ranges and handle validation errors in Python
Built on validation infrastructure from commit f342ca77 (PdmFieldHandle and
PdmObjectHandle validation support).
* #12044 Python: Add Non-Darcy properties available in Python
* #12049 Python: Add checking for multiple consecutive upper case letters
The scripting keyword is transformed to snake_case for use in Python. When data in Python is sent back to ResInsight, the opposite operation happens. This concept works well for most variants of keywords, but there are some corner cases that is not working.
Add compile time checking to make sure that the scripting keywords are formatted correctly.
Fix keyword that has been formatted the wrong way and has never worked.
* Avoid running test if ResInsight executable path is not defined in env
* Make sure dash is allowed in enumeration text string
* Refactor interface to PdmObjectHandle and PdmFieldHandle
Return objects instead of passing in structures as parameters
* Add nodiscard to several functions
* Remove redundant this->
* Rename to ptrReferencedObjectsByType
* #7885 Update opm-common with optimized coordinate import
* #7885 Allow null as default result from a script method
* #7885 Propagate default parameter values to generated Python code
* #7885 Add CommandRouter as hub for worker methods
* #7885 Add support for use of CommadRouter from Python
* #7797 Well Targets: Add scripting capability
* #7794 Python : Do not update childField or childFieldArray
* #7797: Python - Add scripting to well path collection
- Extend the pdmobject.py with method add_object()
- allow objects to be created from Python in well path collections
- add well targets to modelled well path
* #7795 Python : Make sure referenced generated classes are defined
* #7810 StimPlanModel: clean-up python generation
* Python : Always use empty string as default value for ptrFieldValue
It can happen that a ptrField is assigned to a pointer on object construction. (FaciesProperties) Make sure that constructor always assigns an empty string.
Co-authored-by: magnesj <magnesj@users.noreply.github.com>
Co-authored-by: Kristian Bendiksen <kristian.bendiksen@gmail.com>
Kristian Bendiksen
Magne Sjaastad
Gaute Lindkvist