feat(pipeline): report when a re-ingest disagrees with the page on disk (#104) #129

Merged
mathias merged 2 commits from fix/104-surface-merge-conflicts-v2 into main 2026-09-23 05:22:28 +00:00
2 changed files with 45 additions and 1 deletions
Showing only changes of commit b05e9c48f4 - Show all commits
+16 -1
View File
@@ -84,7 +84,7 @@ func MergeReport(a, b Page) (Page, []string) {
secsA[i].content = appendNew(sA.content, sB.content)
default:
// a wins — but say that b disagreed, rather than dropping it mute.
if strings.TrimSpace(sA.content) != strings.TrimSpace(sB.content) {
if canonicalBlock(sA.content) != canonicalBlock(sB.content) {
conflicts = append(conflicts, sB.heading)
}
}
@@ -329,6 +329,21 @@ func canonicalLine(line string) string {
return strings.Join(strings.Fields(line), " ")
}
// canonicalBlock applies canonicalLine per line of a multi-line section body
// — the same whitespace-insensitivity appendNew/mergeBullets already give a
// single line, extended to a whole section for MergeReport's disagreement
// check (found in review, PR #129: comparing raw strings.TrimSpace only
// tolerated leading/trailing whitespace, not a re-extraction's mid-line
// spacing, which every other comparison in this file already tolerates).
func canonicalBlock(block string) string {
lines := strings.Split(block, "\n")
out := make([]string, len(lines))
for i, l := range lines {
out[i] = canonicalLine(l)
}
return strings.Join(out, "\n")
}
// appendNew concatenates the lines of b that a does not already carry.
//
// Append sections used to concatenate wholesale, which is correct exactly once.
+29
View File
@@ -385,3 +385,32 @@ Appetite is fixed, scope is variable.
_, conflicts := MergeReport(page, page)
assert.Empty(t, conflicts, "identical content is not a disagreement")
}
func TestMerge_ReportsNothingForInternalWhitespaceVariantsOnly(t *testing.T) {
// Found in review of PR #129: the conflict comparison used raw
// strings.TrimSpace, which only tolerates LEADING/TRAILING whitespace —
// unlike appendNew/mergeBullets, which already canonicalize internal
// whitespace runs too (#103: a re-extraction's spacing isn't guaranteed
// byte-stable). A mid-line double-space from re-extraction would
// false-positive as a "disagreement" here even though every other
// whitespace-sensitive comparison in this file already tolerates it.
existing := Page{Path: "wiki/sources/x.md", Content: `---
title: 'X'
---
## Summary
Appetite is fixed, scope is variable.
`}
fresh := Page{Path: "wiki/sources/x.md", Content: `---
title: 'X'
---
## Summary
Appetite is fixed, scope is variable.
`}
_, conflicts := MergeReport(existing, fresh)
assert.Empty(t, conflicts, "internal whitespace differences alone are not a disagreement")
}