refactor: migrate Boost.Python syntax to nanobind#86
Conversation
There was a problem hiding this comment.
This is a very early port with only syntax checks performed. The pickle suite naming is left but underneath it is handled differently, the final ways of pickling depends on migration for rest of the files.
There was a problem hiding this comment.
There's an explicit behavior change (or enhancement?) in atomadapter_indexing that non-unit and negative-step slices are handled. We may need to test this on all of the packages to see if it breaks anything.
…nd BondCalculator
…nd libdiffpy_version
…ty and PDFBaseline
|
@sbillinge Now the syntax have been migrated and build and tests are passed. I'll keep marking the PR as draft since I found some of parts are not done elegantly, but we can start reviewing this. |
|
@SYMSCAE pleae can you, on a separate PR, do the updates of the pre-commit versions so we can get that passing also? I will start to look at this. @pavoljuhas if you have time, there is more c++ reviewing here. @ycexiao @danielsirakov, if you guys have time, please could you also take a look. This is the work to migrate from boost to nanobind |
I'm working on that, but the change of docformatter breaks some things. I'll open a new PR and bring the discussion over there. |
|
I merged that PR so you can update main and merge it into this branch |
2 participants
This PR is the initial port that migrates the original Boost.Python bindings to nanobind as mentioned in diffpy/pyobjcryst#88, noticed future improvements will be mentioned in TODOs but will be addressed in future PRs.