You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 53edf3f
Browse filesBrowse the repository at this point in the historyBrowse files
Two small cross-port corrections, identical in TypeScript, Python, C# and Java.
M7 — rule (d) rejected @sourceRefField with "but is not a M:N relationship.", which
reads as "this attribute is never allowed here" now that @Cardinality: "one" is a
legal home for it (the whole point of this branch). The message now names both:
"...but is neither a M:N relationship (requires @through with @Cardinality: "many")
nor a @Cardinality: "one" relationship." Wording is byte-identical across the four
ports.
M6 — rule (e)'s header comment still justified itself against a rule (d) that was
"own-scoped" ("so own-scoping there is correct"). Rule (d) stopped being own-scoped
in this same branch: it walks the effective relationship set and dedupes by node
identity, reporting against the declaring entity. The comment now says what actually
differs — the SUBJECT of each pass, not which relationships it can see.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Lfd8nat1WcSpXffd8YetiC
Copy file name to clipboardExpand all lines: server/java/metadata/src/test/java/com/metaobjects/relationship/Issue368RelationshipReferenceValidationTest.java
+5-1Lines changed: 5 additions & 1 deletion
Original file line number
Diff line number
Diff line change
@@ -224,7 +224,11 @@ public void sourceRefFieldOnAbsentCardinalityStillErrors() {
0 commit comments