Skip to content

The cross-harness, self-improving memory layer for AI agents. Join our community: https://discord.com/invite/zdNChS2fBu

Hash the event identity by streaming rather than into a sized buffer

Claudee325d9341883
diff --git a/pkg/asymptoteobserve/identity.go b/pkg/asymptoteobserve/identity.go
index f67e246b..22018b7e 100644
--- a/pkg/asymptoteobserve/identity.go
+++ b/pkg/asymptoteobserve/identity.go
@@ -5,6 +5,7 @@ import (
"encoding/hex"
"encoding/json"
"fmt"
+ "io"
"strings"
)
@@ -99,14 +100,19 @@ func EventIDForLine(line []byte) string {
//
// SHA-1 is what version 5 specifies. It is used here as a namespaced identifier
// function over data Beacon just wrote, never as a security boundary.
+//
+// The parts are streamed into the hash rather than concatenated into a sized
+// buffer first. name is a whole serialized event, so sizing an allocation from
+// its length is a computation on a value the event's own content decides -- and
+// there is no reason to hold a second copy of the event in memory to hash it.
func eventUUID(kind, name string) string {
- buf := make([]byte, 0, len(eventIDNamespace)+len(kind)+len(name))
- buf = append(buf, eventIDNamespace[:]...)
- buf = append(buf, kind...)
- buf = append(buf, name...)
- sum := sha1.Sum(buf)
+ digest := sha1.New()
+ // hash.Hash.Write is documented never to return an error.
+ digest.Write(eventIDNamespace[:])
+ io.WriteString(digest, kind)
+ io.WriteString(digest, name)
var id [16]byte
- copy(id[:], sum[:16])
+ copy(id[:], digest.Sum(nil))
id[6] = (id[6] & 0x0f) | 0x50 // version 5
id[8] = (id[8] & 0x3f) | 0x80 // RFC 4122 variant
out := make([]byte, 32)