Closed nacho closed 11 months ago
I went at the end for a more conservative approach given the CI failures. We can revisit it later on. Can you please approve again the run? Thanks
Hi,
Thank you for your patch. While we're generally not opposed to an idea like this, maintaining support for BoringSSL which has no guarantees of API stability is a fairly heavy burden. It might make more sense for us to refactor libfido2 to use a thin crypto abstraction layer that is statically linked. That would make it easier for a third party to write their own wrapper to use something else if they prefer. This has been discussed previously in the context of other crypto libraries.
Sounds fair, thanks for the comment
To build with BoringSSL CRYPTO_INCLUDE_DIRS and CRYPTO_LIBRARY_DIRS must be defined
Also fix all the compile issues when building with BoringSSL