Fix bad-import-yang-version handling #237
Conversation
It is allowed to import YANG 1.1 modules from YANG 1.0 when importing without revision-date. bad-import-yang-version was flagged even when a YANG 1.1 module was imported by YANG 1.0 module without specifying revision-date and is now fixed. This fix addresses TypeFox#228 Signed-off-by: Siddharth Sharma <[email protected]>
Added build settings for vscode Signed-off-by: Siddharth Sharma <[email protected]>
esmasth
force-pushed
the
esmasth/issue228
branch
from
e6ebf1c
to
8760d7e
Compare
|
@esmasth |
| if(baseModuleVersion != importedModuleVersion) { | ||
| val message = '''Cannot import a version «importedModuleVersion» module in a version «baseModuleVersion» module.'''; | ||
| val revisionDate = importStatement.substatementsOfType(RevisionDate) | ||
| if (!revisionDate.nullOrEmpty && baseModuleVersion != importedModuleVersion) { |
There was a problem hiding this comment.
[nit] substatementsOfType can not return null, you can just use .empty.
[nit] revisionDate is a lazy evaluated iterable, if you move it after baseModuleVersion != importedModuleVersion it will only be calculated in the erroring case and save some computation time.
Proposed change:
val revisionDate = importStatement.substatementsOfType(RevisionDate)
if (baseModuleVersion != importedModuleVersion && !revisionDate.empty) {There was a problem hiding this comment.
Fixing this in following update
|
@esmasth |
|
@dhuebner Thanks, is it decided then to make these issues configurable? I rather agree with your statement on yang-vscode that it's a potential to break the engine if we allow unambiguously wrong input. |
No, but we can allow to configure it by the user. I think the user can then decide what to do.
Do you what to apply the suggestions I made, or should I merge your PR as it is? |
|
I will do the late evaluation and It might be good to have some strictness for MUST/SHALL statement violations in YANG RFC with some consistent policy for the project. Maybe good to discuss that on TypeFox/yang-vscode#40 to conclude? |
|
@esmasth |
substatementsOfType does not return null and hence only empty needs to be checked. Also changed the order of checks to improve performance. Signed-off-by: Siddharth Sharma <[email protected]>
|
Ok with merging this now, if you are. |
|
@esmasth |
It is allowed to import YANG 1.1 modules from YANG 1.0 when importing without revision-date. bad-import-yang-version was flagged even when a YANG 1.1 module was imported by YANG 1.0 module without specifying revision-date and is now fixed.
This fix addresses #228