This patch series fixes the issues caused by --enable-pubkey-ciphers=ecc (and hopefully also fixes the use of ECC without a proper #if or runtime check).
Revisions and Commits
|rC9d6203532d90 tests/basic: Skip non-FIPS tests in FIPS mode|
|rCc5480b4934bb fips: Skip selftests of unsupported PK algos|
|rC572b0bf9668d tests/keygen.c: Skip unavailable PK algorithms|
|rC56000fb5c42f build: Skip PK-specific tests if algo is disabled|
|rC78c0d76f8092 tests/benchmark: Skip unavailable PK algorithms|
|rCe78cf3df23a2 tests/basic: Skip tests if PK algo is unavailable|
|rC4f4da6cbf065 tests/pubkey: Skip tests if PK algo is unavailable|
|rC96fafffeeba5 kdf: Skip tests if hash algo is not available|
Here's another one related to this: https://lists.gnupg.org/pipermail/gcrypt-devel/2022-July/005344.html
Turns out I should also have tested in FIPS mode. This particular one seems to be a regression from e96980022e5ec079c9d4e3492eb6a1131c68e0f2, which moved the continue into the else branch.