340995 - complexType without complexContent fails validation

Status: RESOLVED FIXED

(Core Graveyard :: XForms, defect)

Description

[aaronr]

If the user has a schema containing a complexType that does not contain complexContent or sequence, etc., then we will fail document validation inside nsSchemaValidator::ValidateComplextype.  The mContentType that we'll encounter will be nsISchemaComplexType::CONTENT_MODEL_EMPTY which is not implemented currently.

Comment 1

[aaronr]

Attachment: testcase

testcase that shows the problem.  You won't see any problem visually, but if you go to the error console, you'll see that the error, "ERR: XForms Warning (3): Instance document did not validate"

Comment 2

[Steve Speicher]

I tested the associated test case with patch from bug 310962 (attachment 253149 [details] [diff] [review]) and it works.  Perhaps this bug depend on 310962?

Comment 3

[aaronr]

(In reply to comment #2)
> I tested the associated test case with patch from bug 310962 (attachment
> 253149 [details]) and it works.  Perhaps this bug depend on 310962?
> 

Looks like 310962 depends on trunk only features.  But perhaps we can use the piece of it that fixes this problem?  That might live only in the extension and thus be able to be checked into 1.8.

Comment 4

[Steve Speicher]

Attachment: allows empty content for complex types

Comment 5

[Doron Rosenberg (IBM)]

Comment on attachment 256069 [details] [diff] [review]
allows empty content for complex types

looks good

Comment 6

[aaronr]

Comment on attachment 256069 [details] [diff] [review]
allows empty content for complex types

>Index: schema-validation/src/nsSchemaValidator.cpp
>===================================================================

> nsresult
>+nsSchemaValidator::ValidateComplexModelEmpty(nsIDOMNode* aNode,
>+                                    nsISchemaComplexType *aSchemaComplexType,
>+                                    PRBool *aResult)
>+{
>+  PRBool isValid = PR_TRUE;
>+  nsresult rv = NS_OK;
>+
>+  nsCOMPtr<nsIDOMNode> currentNode;
>+  aNode->GetFirstChild(getter_AddRefs(currentNode));
>+  NS_ENSURE_SUCCESS(rv, rv);

nit: why do you need rv?  You never use it.  Just return NS_OK at the end.

>Index: schema-validation/src/nsSchemaValidator.h
>===================================================================
>RCS file: /cvsroot/mozilla/extensions/schema-validation/src/nsSchemaValidator.h,v
>retrieving revision 1.10
>diff -u -8 -p -r1.10 nsSchemaValidator.h
>--- schema-validation/src/nsSchemaValidator.h	15 May 2006 21:34:38 -0000	1.10
>+++ schema-validation/src/nsSchemaValidator.h	22 Feb 2007 20:58:17 -0000
>@@ -298,16 +298,20 @@ private:
>   nsresult ValidateComplextype(nsIDOMNode *aNode,
>                                nsISchemaComplexType *aSchemaComplexType,
>                                PRBool *aResult);
> 
>   nsresult ValidateComplexModelElement(nsIDOMNode *aNode,
>                                        nsISchemaComplexType *aSchemaComplexType,
>                                        PRBool *aResult);
> 
>+  nsresult ValidateComplexModelEmpty(nsIDOMNode *aNode,
>+                              nsISchemaComplexType *aSchemaComplexType,
>+                              PRBool *aResult);
>+

nit: line up param types...looks like that is what the surrounding ones do.

with those, r=me

Comment 7

[Steve Speicher]

Attachment: patch with nits fixed, ready for checkin

patch for checkin

Comment 8

[Doron Rosenberg (IBM)]

Fixed in trunk.  Does this need to go into any branches?

Comment 9

[Steve Speicher]

I'll just tag it xf-to-branch and can go with rest of XForms 0.8 items.  Thanks for the ci

Comment 10

[aaronr]

checked into 1.8 branch on 2007-04-12
checked into 1.8.0 branch on 2007-04-16