Fix Doxygen Ref issues - #536
Conversation
|
I'm getting folowing error: |
Is this when running the I believe this is due to a bug in Doxygen. My doxygen version is 1.17.0 and I do not have these errors. When I downgrade to 1.9.8, I get similar errors. |
|
Error on RTD when I tried to build docs If you fix those I can run it again until it builds on RTD |
Yes those are exactly what I get with the outdated version of doxygen. This is a doxygen bug - you should try updating. |
|
Well that's the unfortunate thing, we can't do that on RTD unless we just use a more recent version of Ubuntu In build:
os: ubuntu-26.04 |
Increasing Ubuntu version will increase Doxygen version for fixing a bug.
|
https://gridkit.readthedocs.io/en/alex/ Take a look, it builds now! |
Description
The PR #524 was created to fix some error messages Doxygen was giving about refs not being found. It was fixed by removing the refs, but those refs are useful so it would be nice to fix the root issue instead. This PR does that.
Proposed changes
The root of the issue is a quirk with Doxygen - since the docs of
CircuitComponentare inherited byPowerElectronicsModel, some of the refs inCircuitComponentdocs appear in thePowerElectronicsModelcontext, but some of these items are re-imported into thePowerElectronicsModelnamespace. The re-imports aren't documented and Doxygen can only tell that they refer to the same things if in the code they clearly come fromCircuitComponent.DOXYGEN_EXTRACT_ALLhas been set in the normal docs to help with normal docs as well - many items aren't documented, but they should still appear in the documentation. Refs will also be able to resolve for them even if they aren't documented.EXTRACT_PRIVATEhas been set in the Sphinx documentationDoxyfile. Sphinx will strip privates from the doxygen xml, and this will help errors from being generated, since the Doxygen documentation is developer-oriented, it might mention private members in the documentation.Checklist
-Wall -Wpedantic -Wconversion -Wextra.This is part of a previous fix
Further comments
@lukelowry can you verify that this fixes the issue you were having in the previous PR?