Region-to-region addresses were reconstructed as "R1 - R2" (with spaces) when
looking up keywords for ensemble realizations without a pre-built address map,
but opm-common's on-disk keyword format has no spaces ("R1-R2"). This caused
silent data loss for region-to-region vectors (RGFT*, ROFT*, RWFT*, etc.) in
all but one ensemble realization.
Fix formatUiTextRegionToRegion() to use "R1-R2", and add a regression test.
Fixes#14853.
Migrate all assert macros in ApplicationLibCode to CAF_ASSERT and remove every
use of cvfAssert.h.
CVF_ASSERT is replaced one to one. CVF_TIGHT_ASSERT is also replaced by
CAF_ASSERT, which is semantically exact: CVF_ENABLE_TIGHT_ASSERTS is 1 only
under _DEBUG, and that is what CAF_ASSERT now does. The two CVF_FAIL_MSG sites
become CAF_ASSERT( false && "message" ), preserving the message with the idiom
already used elsewhere in the code base.
Counts before and after: CVF_ASSERT 1044 to 0, CVF_TIGHT_ASSERT 66 to 0,
CVF_FAIL_MSG 2 to 0, cvfAssert.h references 154 to 0.
Include handling: files that included cvfAssert.h directly now include
cafAssert.h instead, includes left dead by the migration are removed, and files
that were relying on cvfAssert.h transitively get an explicit cafAssert.h. Files
that reach cafAssert.h through another caf header are left unchanged; a missing
include here is a compile error, not a silently disabled assert.
ResultStatisticsCache links only LibCore and therefore had no path to
cafAssert.h. Add the cafPdmCore directory as a private include path rather than
linking the library, since cafAssert.h is header only.
Note that this stops these asserts from firing in Release and RelWithDebInfo,
where CVF_ASSERT was previously active.
Refactored RifEclipseSummaryAddress to use setter methods for string-to-index assignments (vector, name, LGR), centralizing logic and improving maintainability. Added debug-only string members for easier inspection in debug builds, wrapped in a struct to preserve default comparison behavior. Updated all relevant constructors and factory methods to use the new setters. No impact on release builds.
Introduced custom percentile handling for ensemble statistics:
- Added "CUSTOM" statistics type and percentile field to RifEclipseSummaryAddress
- Refactored RimEnsembleStatisticsCase to store and compute arbitrary percentiles
- RimEnsembleStatistics now parses custom percentiles from UI input
- RimEnsembleCurveSet and related interfaces updated to handle custom percentile curves and labels
- Hashing and statistics calculations now include custom percentiles for cache validity
Replaced direct string members with indices into RiaStringPool for vector, well, group, and LGR names. Updated all constructors, accessors, and logic to use the string pool for storage and retrieval. This change reduces memory usage and improves performance for repeated string values.
Fixes by clang-format
Improves the parsing logic for well names in Eclipse
summary addresses, allowing well names to contain colons.
Previously, the parsing would incorrectly split well names
at colons.
Adds a unit test to verify correct parsing of well names with
colons.
Removes the dedicated `SUMMARY_ENSEMBLE_STATISTICS` category and instead uses a `statisticsType` field within `RifEclipseSummaryAddress` to represent ensemble statistics.
This change simplifies the code and avoids potential confusion by representing statistics as attributes of existing summary types, rather than a separate category. It also improves code robustness.
* Add category WELL_CONNECTION
* Throw exception if category string is invalid
* Make it possible to step connections
* Add definition of completion vectors
* Use WELL_CONNECTION for CPRL
* Convert "WGPRL__2:MY-WELL1-A5" to "WGPRL:MY-WELL1-A5:2"
Use the syntax for well completions in opm-common
* More tests for different variants of summary texts
* Qt6: Avoid insertWidget, use addWidget
In Qt6, the insertWidget function checks if the index parameter is valid based on current widgets present in the layout. This is error prone, and use addWidget to avoid manual counting of index.
* Disable use of Qt keyword foreach
* Replace use of QRegExp with QRegularExpression
Replace use of QRegExp with QRegularExpression
Remove dependency on qt5compat module
Simplify an expression based on review
* Remove Qt5 ifdefs
* Guard access out of bounds seen in debug build
* Avoid reuse of string variable
* Disconnect all signals from the QOpenGLContext
The call stack when this assert happens indicates that there are more signals to be disconnected from the object. Crash is fixed by disconnecting all signals.
Assert seen in debug build:
ASSERT failure in caf::Viewer: "Called object is not of the correct type (class destructor may have already run)", file C:\Qt\6.6.3\msvc2019_64\include\QtCore/qobjectdefs_impl.h, line 130
* Fix issue related to delete of a linked view
Guard null pointer use in view linker. Remove complicated cleanup in destructor in Rim3dVew.
Required changes to use Qt6 and disable support for Qt5. There are still some adjustments related to Qt6 to be done, but these changes are not required to make Qt6 compile on relevant systems.
* Build system changes Qt6
* Override enterEvent
* Update QKeySequence
* QtChart changes
* Use QScreen to instepct dotsPerInch
* Add app->quit()
* Required updates for code related to QString
* Use RiaQDateTimeTools
* Required changes related to regular expressions
* Support compile on Qt < 6.5
When version < 6.5 is found, qt_generate_deploy_app_script() is disabled. Compilation of ResInsight will work, but the install target will be incomplete.
* Octave: add missing header.
* Qt Advanced Docking: force Qt6 where both Qt5 and Qt6 is available.
---------
Co-authored-by: magnesj <1793152+magnesj@users.noreply.github.com>
Co-authored-by: Kristian Bendiksen <kristian.bendiksen@gmail.com>
* Make sure clang-tidy action use .clang-tidy config file
Use add-paths to instruct create-pull-request to a sub folder to avoid diff from Qt and vcpkg
* Use empty() in macro to avoid clang-tidy warning
* Add NOLINT to CAF_ASSERT
* Add NOLINT to cvfAssert
* Use one common variable for object name, use three ints
* Move enums to separate file
* Refactor use of enums
* Move implementation to cpp
* Refactor includes
* #9966 Summary calculations: add support for imported vectors.
* #9966 Summary Calculation: add support for imported and observed data.
* Summary Address: avoid extra level in tree for imported data.
* #9966 Update observed data when calculation is updated.
* Summary Calculation: use shorter name in title
* Sort summary addresses to group calculated results last.
* Summary Calculation: add option to distribute calculation to other items.
* Add support for summary calculation for RimGridSummaryCase.
* Use short name for calculations in plot titles.
* Update ensembles in Data Sources when calculation is added.
* Summary Calculation: allow drag-and-drop of ensembles
* Summary Plot: fix axis range aggregation for calculated ensemble addresses.
---------
Co-authored-by: Kristian Bendiksen <kristian.bendiksen@gmail.com>
Fixes incorrect classification for vectors with 8 characters starting with
'N' (e.g NLINSMAX). They were classified as "network", but lookup table have
them correctly classified as "misc".
Magne Sjaastad
magnesj
magnesj
jonjenssen
Gaute Lindkvist