fogfish / arcnet-cli

80%

main: 81%

LAST BUILD ON BRANCH 016-arc-revert

branch: 016-arc-revert

CHANGE BRANCH

x

Reset

• 016-arc-revert
• 001-cli-infrastructure
• 002-arc-init
• 003-apply-patch
• 004-arc-lint
• 005-graph-schema-first-class
• 006-arc-grep-content-search
• 007-arc-subgraph
• 008-arc-serve-mcp
• 009-node-timestamp-attrs
• 010-predicate-node-model
• 011-machine-readable-schema
• 012-predicate-merge-policies
• 013-predicate-role-rendering
• 014-lint-spec-conformance
• 015-predicate-node-shape-cli
• 017-subclass-of-predicate
• 018-apply-schema-patch
• 019-camelcase-node-types
• fix-apply-and-lint-sourceCitekey
• fix-github-actions
• fix-predicate-apply
• main

Build # 29356384995

Build Type

Pull #19

github

Committed by fogfish

Commit Message

[Spec Kit] Fix BUG-002: union/append outcome must reflect actual result

mergeTexts/mergeAttrs's isListMerge branch (internal/core/merge.go)
always reported a union/append predicate as "appended" (or "created")
based on which merge behavior dispatched, never checking whether the
dispatch actually changed the value. A contribution that mergeText's own
near-duplicate detection correctly drops (or unionPredicates correctly
deduplicates) was still reported "appended", misleading arc apply
--verbose users into thinking content was added when nothing changed.

Now compares the merged value against the existing one and reports
OutcomeUnchanged for a genuine no-op, matching every other dispatch
class's own accuracy standard (spec 012 FR-019).

Adds regression coverage in internal/core/merge_test.go (exact and
near-duplicate text, duplicate Attrs list entry) and an E2E test in
cmd/arc/graph/apply_test.go reproducing the reported scenario; fixes
TestApplyReportsStepPerNode's fixture, which had encoded the buggy
expectation.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

Pull Request Pull Request #19: (fea) `arc revert` part of the graph

Coverage Stats

551 of 731 new or added lines in 8 files covered. (75.38%)

3878 of 4830 relevant lines covered (80.29%)

0.93 hits per line