User-Agent: Mozilla/5.0 (X11; U; Linux i686; fr; rv:1.9.0.3) Gecko/2008092816 Iceweasel/3.0.3 (Debian-3.0.3-3) Build Identifier: Mozilla/5.0 (X11; U; Linux i686; fr; rv:1.9.0.4) Gecko/2008102920 Firefox/3.0.4 The RDF parser generates anonymous resources when it sees "rdf:nodeID". This works properly when a <rdf:Description rdf:nodeID="foo"> appears in the file before the id "foo" is referenced. Yet, when the rdf:nodeID appears as the object of a predicate before being described in a <rdf:Description>, the parser generates two different resources for the same identifier. Such an input is valid according to the RDF spec, see the section about nodeID: http://www.w3.org/TR/rdf-syntax-grammar/#section-Syntax-blank-nodes I think this should block bug 11650. Reproducible: Always Steps to Reproduce: 1. Load the following data (example from the RDF spec): <?xml version="1.0"?> <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:ex="http://example.org/stuff/1.0/"> <rdf:Description rdf:about="http://www.w3.org/TR/rdf-syntax-grammar" dc:title="RDF/XML Syntax Specification (Revised)"> <ex:editor rdf:nodeID="abc"/> </rdf:Description> <rdf:Description rdf:nodeID="abc" ex:fullName="Dave Beckett"> <ex:homePage rdf:resource="http://purl.org/net/dajobe/"/> </rdf:Description> </rdf:RDF> 2. Serialize it with "ds.FlushTo(...)". Actual Results: For the moment, FlushTo writes the following file. Note the different resources "rdf:#$Y+Ujz3" and "rdf:#$X+Ujz3". <?xml version="1.0"?> <RDF:RDF xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:ex="http://example.org/stuff/1.0/" xmlns:NC="http://home.netscape.com/NC-rdf#" xmlns:RDF="http://www.w3.org/1999/02/22-rdf-syntax-ns#"> <RDF:Description RDF:about="rdf:#$Y+Ujz3" ex:fullName="Dave Beckett"> <ex:homePage RDF:resource="http://purl.org/net/dajobe/"/> </RDF:Description> <RDF:Description RDF:about="http://www.w3.org/TR/rdf-syntax-grammar" dc:title="RDF/XML Syntax Specification (Revised)"> <ex:editor RDF:resource="rdf:#$X+Ujz3"/> </RDF:Description> </RDF:RDF> Expected Results: We should get instead something like that: <RDF:Description RDF:about="rdf:#$X+Ujz3" ex:fullName="Dave Beckett"> <ex:homePage RDF:resource="http://purl.org/net/dajobe/"/> </RDF:Description> <RDF:Description RDF:about="http://www.w3.org/TR/rdf-syntax-grammar" dc:title="RDF/XML Syntax Specification (Revised)"> <ex:editor RDF:resource="rdf:#$X+Ujz3"/> </RDF:Description> Note that the previous example works fine if you swap the two <RDF:Description>.

I found bug 232623 which introduced the rdf:nodeID feature. I'm not familiar at all with Mozilla's code, yet the following lines of its patch look suspect to my eyes: + if (!*aResource) { + mNodeIDMap.Put(nodeID,*aResource); + return gRDFService->GetAnonymousResource(aResource); It seems that we're associating "nodeID" with a null resource in "mNodeIDMap", then we create the anonymous resource. Which, I think, will make the parser create another resource the next time it sees the same "nodeID".

Comment on attachment 349648 [details] [diff] [review] Proposed fix Nice catch. r-, though, we shouldn't put *aResource into mNodeIDMap on failure. Put in a if (NS_FAILED(rv)) return rv; and then no explicit return in the if clause?

Comment on attachment 349649 [details] [diff] [review] Same patch, but check rv before adding the resource to the map r=me, this needs sr still. Trying shaver. Mike, I think this is small and low-risk enough to take in RDF-land.

Comment on attachment 349649 [details] [diff] [review] Same patch, but check rv before adding the resource to the map sr=shaver, just for nostalgia's sake

Comment on attachment 349649 [details] [diff] [review] Same patch, but check rv before adding the resource to the map a191=beltzner

Landed on mozilla-central for baking, http://hg.mozilla.org/mozilla-central/rev/f4ca2f598902. Leaving open to remind me on 1.9.1

Landed on 1.9.1, too, http://hg.mozilla.org/releases/mozilla-1.9.1/rev/0fcb3bd9bd58. Thanks for the patch, Florian. FIXED and fixed1.9.1.