Upgrade to clubcard-crlite 0.5

[djc]

No description provided.

[djc]

Going to see if we can get

• Downgrade MSRV to 1.85 mozilla/clubcard-crlite#31

[djc]

It's a little weird the benchmark is failing here...

I'm also seeing that on one of my macOS machines, the performance effect from

• Swap aws-lc-rs for a portable, private SHA256 #181

on revocation-input-from-certs seems to be not good:

revocation-input-from-certs
                        time:   [761.48 ns 763.69 ns 766.11 ns]
                        change: [+32.633% +33.217% +33.805%] (p = 0.00 < 0.05)
                        Performance has regressed.

Do we still think this is the right trade-off?

When the benchmark passes (seems intermittent locally -- also bad), "use internal SHA-256 implementation for hashing" seems to worsen performance on the same benchmark by another 15%.

[codspeed-hq]

Merging this PR will not alter performance

✅ 4 untouched benchmarks

---

_{Comparing clubcard-crlite-0.5 (0e4e447) with main (9c812ec)}

[djc]

Okay, part of the problem was that I got the hashing operation for the key wrong. With that fixed, my results from benchmarking 1c9dd7a (last commit before #181) against the tip of this branch are:

• +59% on revocation-input-from-certs (570ns -> 910ns)
• -4.7% on revocation-check (1.0ms -> 0.99ms)

I guess that's fine; still a little weird/worrying that these changes are not visible on Codspeed.

[djc]

Going to hold off on releasing until we have

• upki: check all filters #185

merged at least. Let me know if you know of other things that should block this.