Closed pramsey closed 1 month ago
I think they were copied from a previous conda-forge project (maybe @benbovy set that up?)...I have no problem removing them.
I'm sure you've figured this out by now, but I had lofty ambitions to port the R tests ( https://github.com/r-spatial/s2/tree/main/tests/testthat ) to C++ after I split this bit of code off, but I think what is there now is non-existent.
Yes, I see there's very few tests, and I have another branch with a few extra WKT tests... I kind of banged up against the limits of my knowledge WRT how precisely the gtests are supposed to be done. It seems really duplicative to have a separate gtest executable for every .cc
file, but I am not sure what the expected gtest conventions are. Do you have any guidance?
Writing tests is hard; reorganizing them is easy! Any tests you're willing to contribute are welcome.
I think they were copied from a previous conda-forge project (maybe @benbovy set that up?)...I have no problem removing them.
Sorry for the late reply (better late than never!).
I agree this is annoying.
If you haven't figured out how to remove the 3rd-party warnings, s2geometry v0.11 has now an option S2_USE_SYSTEM_INCLUDES
to silence the warnings in s2 headers. Not sure that applies to the s2 dependencies (absl, openssl), though. I think we could also use here some cmake workarounds and tweak some properties of the imported targets to avoid that noise.
When build tests is turned on, the following warnings are enabled, which mostly (as far as I can tell) warn about things in the S2, OpenSSL and absl headers.
https://github.com/paleolimbot/s2geography/blob/master/CMakeLists.txt#L176-L178
This makes for a noisy build which probably hides any issues that might be in s2geography. Is there a rationale to keep them?