Avoid repeated prefix scans when projecting item policies

This commit is contained in:
2026-10-24 00:19:14 +00:00
parent 25cf7ce05c
commit 8d40983271
3 changed files with 81 additions and 9 deletions

View File

@@ -1,7 +1,7 @@
# Sequence Implementation Review Remediation
Status: remediation in progress. Completed findings are marked below; this
record introduces no new release gate.
Status: all eleven findings complete. Each finding has a separate remediation
commit and completion note below. This record introduces no new release gate.
Reviewed commit: `0be7cdb00eff576fda6b30c0600b2134d4e2e086`.
Implementation range: `bc6b977..0be7cdb`.
@@ -25,6 +25,10 @@ Validation on the reviewed commit:
- Finding 10 follows directly from the maintained text; finding 11 follows from
the prefix-scan loop, not a timing comparison or performance threshold.
Final validation after remediation: `./scripts/check.sh` and
`go test -race ./...` passed with Go 1.26.10 on Linux. The original public-API
probes for findings 1–9 now demonstrate the required corrected behavior.
P1 denotes the highest-priority correctness issue here; P2 denotes normal-priority
work needed to complete the selected behavior. Preserve the agreed scope while
fixing these issues. In particular, do not resolve failures by restricting valid
@@ -376,7 +380,16 @@ Keep YAML-only enforcement distinct from registered final validation.
## 11. Remove The Quadratic Prefix Scan In Item Policy Projection — P2
Status: [ ] Open.
Status: [x] Complete.
Completion: assigned sequence ranges now supply item offsets directly. Handoffs
without a range lazily accumulate prefix lengths, visiting each preceding
contribution once rather than rescanning every prefix. Regression coverage keeps
inline and policy-less contributions in the correct positions with complete,
partial, or absent ranges. Existing ownership/placement tests remain intact.
Verified focused item-policy tests, `./scripts/check.sh`, `go test -race ./...`,
and inspection of the remaining error-only prefix-scan callers. No new benchmark
fixture or performance threshold was needed for this bounded change.
Owner: [`projectNestedCollectionPolicies`](../../internal/engine/imports.go),
especially line 968; the scanned helper is `priorSequenceLength` in the same file.

View File

@@ -961,17 +961,28 @@ func projectNestedCollectionPolicies(composed *composedDocument, scope acquisiti
return
}
baseOffsets := make(map[string]int)
var prefixLengths map[string]int
for sourceIndex, contribution := range scope.contributions {
if prefixLengths != nil && sourceIndex > 0 {
addSequencePrefixLengths(scope.contributions[sourceIndex-1].document, Path{}, prefixLengths)
}
if contribution.nested == nil || len(contribution.nested.collectionPolicies.collections) == 0 {
continue
}
itemIndex := priorSequenceLength([]acquisitionScope{{contributions: scope.contributions[:sourceIndex]}}, contribution.destination)
if assigned, found := ranges[sequenceRangeKey{sourceIndex: sourceIndex, path: contribution.destination.String()}]; found {
if projected, ok := assigned.outputForInput(0); ok {
itemIndex = projected
}
}
key := contribution.destination.String()
assigned := ranges[sequenceRangeKey{sourceIndex: sourceIndex, path: key}]
itemIndex, found := assigned.outputForInput(0)
if !found {
// Normal composition supplies the assignment directly. For a
// handoff without ranges, scan each preceding tree only once.
if prefixLengths == nil {
prefixLengths = make(map[string]int)
for _, previous := range scope.contributions[:sourceIndex] {
addSequencePrefixLengths(previous.document, Path{}, prefixLengths)
}
}
itemIndex = prefixLengths[key]
}
base, initialized := baseOffsets[key]
policy, _ := composed.collectionPolicies.at(contribution.destination)
if !initialized {
@@ -986,6 +997,20 @@ func projectNestedCollectionPolicies(composed *composedDocument, scope acquisiti
}
}
func addSequencePrefixLengths(node *document.Node, path Path, lengths map[string]int) {
if node == nil {
return
}
if node.Kind == document.Sequence {
lengths[path.String()] += len(node.Sequence)
return
}
// Contribution destinations select mapping paths, not sequence members.
for _, entry := range node.Mapping {
addSequencePrefixLengths(entry.Value, path.Key(entry.Key), lengths)
}
}
func projectNestedOrigins(target map[string]Origin, root *document.Node, origins map[string]Origin, prefix Path) {
var visit func(*document.Node, Path)
visit = func(node *document.Node, path Path) {

View File

@@ -195,6 +195,40 @@ func TestNestedItemPolicyProjectionUsesStableOutputIndexes(t *testing.T) {
}
}
func TestItemPolicyProjectionCountsContributionsWithoutLocalPolicy(t *testing.T) {
t.Parallel()
items := Keys("items")
for _, assignments := range []map[int]int{nil, {1: 2, 3: 4}, {1: 2}, {3: 4}} {
t.Run(fmt.Sprint(assignments), func(t *testing.T) {
local := collectionPolicyScope{}
local.set(Keys("steps"), collectionPolicy{identityFields: []string{"name"}, identityPresent: true})
scope := acquisitionScope{contributions: []acquiredContribution{
{document: parseItemPolicyDocument(t, "inline", "items: [{id: inline-a}, {id: inline-b}]\n")},
{document: parseItemPolicyDocument(t, "first", "items: [{id: first}]\n"), destination: items, nested: &composedDocument{collectionPolicies: local}},
{document: parseItemPolicyDocument(t, "gap", "items: [{id: gap}]\n"), destination: items},
{document: parseItemPolicyDocument(t, "second", "items: [{id: second}]\n"), destination: items, nested: &composedDocument{collectionPolicies: local}},
}}
ranges := make(map[sequenceRangeKey]sequenceRange)
for sourceIndex, output := range assignments {
ranges[sequenceRangeKey{sourceIndex: sourceIndex, path: items.String()}] = sequenceRange{start: output, end: output + 1}
}
composed := &composedDocument{}
projectNestedCollectionPolicies(composed, scope, ranges)
policy, _ := composed.collectionPolicies.at(items)
if len(policy.recordOrder) != 5 {
t.Fatalf("record policies = %d, want 5 positions", len(policy.recordOrder))
}
for index := range 5 {
record, _ := policy.recordPolicyAt(index)
_, found := record.scope.at(Keys("steps"))
if found != (index == 2 || index == 4) {
t.Errorf("local policy at item %d = %t", index, found)
}
}
})
}
}
func TestOverlaySequenceTransfersIncomingItemPolicyOwnership(t *testing.T) {
path := Keys("records")
registered := mustCompileSequenceIdentityPolicy(t, []sequenceIdentityRegistration{{selector: sequenceIdentitySelector{literalSequenceSelector("records")}, identityFields: []string{"id"}}})