Hence staying closer to the original implementation that didn't distinguish
Sphinx and Docutils specific run configuration internally. It may be beneficial
to return the exact run configuration types once they start to actually differ
in some way or become applicable in some overlapping contexts.
GitOrigin-RevId: 8104c4949dea558ca4199ad06172d1ce9fb1ab8c
The main problem with them is that due to an odd design decision they both share
the same configuration type -- RestRunConfigurationType, and it doesn't quite fit in
this newer typed API, e.g. isConfigurationFromContext() can receive a configuration
instance of a wrong type (DocutilsRunConfiguration for SphinxConfigurationProducer and
vice-versa) due to it, producing a CCE at runtime. I employed a workaround already
used in Python testing run configurations sharing the same problem. Namely,
overloaded getConfigurationSettingsList() to additionally check that instances of
RunnerAndConfigurationSettings with RestRunConfigurationType type actually
correspond to run configurations of the proper concrete class.
Note that setupConfigurationFromContext can also receive a run configuration of a wrong
type but in reality it doesn't happen since Sphinx and Docutils run configurations are
applied to different contexts -- directories and individual .rst files respectively.
GitOrigin-RevId: 15fbb90917edd68e6f29b696c527ff04d4eed433
The default implementation delegates to 'getName' method and it may cause problems if it returns localized value (IDEA-232878).
GitOrigin-RevId: 28495a477af06896f0c3f98141b424c33920f1e0
It was forgotten during implementation of IDEA-106449. Without this method
we tried to parse the whole injected reStructuredText content as a valid
Python file not surprisingly containing tons of syntax errors and then
run all the normal Python inspection over it.
GitOrigin-RevId: ed12c9c6ead90ebb0b5a083126410b2c8cbff233
in production InspectionEP#displayName should be used; highlighting tests just do not use default names
GitOrigin-RevId: f0c172e568219499e7b96982fe352c7d1d3acc69
Python run configurations don't use any indices, so it's safe to update/invoke them while indexing. This change enables
Run / Debug / Profile / Run with Coverage / Concurrency Diagram.
GitOrigin-RevId: 4993c859983c336b88984033ae209fc221290e2a
Build scripts set proper values for bundled plugins and plugins for publishing.
Still, it's possible to break the logic if explicit since-build value is not match \d+.\d+ pattern or if until-build is omitted. So it's better to get rid of it at all.
They don't make any sense since
- they are not required for bundled plugins (com.intellij.ide.plugins.IdeaPluginDescriptorImpl#myVersion will be filled with IDE version)
- they will be overwritten during the build for pluginsToPublish (org.jetbrains.intellij.build.impl.DistributionJARsBuilder#setPluginVersionAndSince)
Still they may affect plugins loading mechanism, e.g. see https://youtrack.jetbrains.com/issue/GO-6741 or https://youtrack.jetbrains.com/issue/IDEA-188964
