Behn is sensitive to the order in which ova are emitted, which was not
correct. Previously, ova emitted first within a single vane activation
were processed as effects last.
+sloy is practically how we perform most scrying. It parses wires
using the full +slay when several of the elements. Replaces all
usages we can with +slaw.
The long term fix is to move to not parsing wires, but that won't
happen before current release-candidate is released.
* release-candidate:
[CI] updates pinned vere commit
Use a trap to lazily evaluate the arguments to %dude in Eyre.
Don't do two passes in the unjetted +sort
%hold removed
losing the `b` reference
remove !:
use assertion
Update pin-vere-commit.txt
type variance bug fix
Trim ethereum snapshot in Jael (#911)
Moved to +jack
Add ~| to arvo to try to debug intermittent crash.
Add special cases for %da and %ta to +slaw
The majority of time in handling a scry is spent in the compiler since it
doesn't use the compiler cache since scrying can't have side-effects and modify
the cache. So prime the worm cache at vane install time so the cache is
prepopulated before we call a scry.
This shaves off >10ms per scry.
* research-constitution: (120 commits)
removes obsolete doccords-style comments
Revert "fixy uno~by, other vor-based code duplication"
Implement argon2.
Use face over lark when available.
Implement blake2b.
Use ++rep over ++can when we can
Use ++rev over ++swp, for safety.
Clean up base58 en/decoding
Rename secp point serialization arms
Add HMAC-SHA1
Put +=byts to use
Restructure and complete BIP32 core
Fix bug in RIPEMD-160 implementation.
BIP32 key derivation logic. Messy, in dire need of restructuring.
Fix bug in HMAC implementation.
SHA-1 re-implementation, can account for leading zeroes.
Converting points on curve to un/compressed numbers.
RIPEMD-160 implementation.
Start work on BIP32 support. Library file includes wrappers for the SHA family to make it take and produce sane byte order data, and a new HMAC implementation that depends on it. Also includes @belisarius222's secp256k1 implementation, plus experimental (and broken) support for other secp variants.
Add ++rev to bit arithmetic, for reversing block order while accounting for leading zeroes.
...
