Support ObjectId-based document IDs.

Fixes: peerlibrary#12
kerf007 · Aug 1, 2016 · 7000871 · 7000871
1 parent 09120cc
commit 7000871
Show file tree

Hide file tree

Showing 2 changed files with 866 additions and 849 deletions.
diff --git a/server.coffee b/server.coffee
@@ -111,7 +111,7 @@ Meteor.publish = (name, publishFunction) ->
             # iteration) but this is OK because in afterRun of other computations this will be corrected and documents
             # with those IDs removed.
             for id in _.difference currentlyPublishedDocumentIds, currentComputationAddedDocumentIds, otherComputationsAddedDocumentsIds, otherComputationsPreviouslyAddedDocumentsIds
-              @removed collectionName, id
+              @removed collectionName, @_idFilter.idParse id
 
           computation.beforeRun =>
             oldDocuments[computation._id] = documents[computation._id] or {}
@@ -142,7 +142,7 @@ Meteor.publish = (name, publishFunction) ->
         # with value set to "undefined". So we look into current session's state and see which fields are currently
         # known and create an object of same fields, just all values set to "undefined". We then override some fields
         # with new values. Only top-level fields matter.
-        for field of @_session.getCollectionView(collectionName)?.documents?[id]?.dataByKey or {}
+        for field of @_session.getCollectionView(collectionName)?.documents?[stringId]?.dataByKey or {}
           oldFields[field] = undefined
         @changed collectionName, id, _.extend oldFields, fields
       else