Meaning of "ordered by" in Node Map Generation step 6.12

Recent versions of the "Latest editor's draft" have changed many of the references of "lexicographic [order]" to "Unicode code point order" which is great.

I'm having trouble getting my implementation to pass toRDF test c019 because, I think, I'm not sure what Node Map Generation step 6.12 means when it states: "Finally, for each key-value pair property-value in element ordered by property perform the following steps:" ... what does "ordered by property" mean? Should I sort the properties by "Unicode code point order" before executing steps 6.12.x or should the properties be processed in the order they were originally inserted into the 'element' being processed?

If I sort the properties by "Unicode code point order" then the test fails because my generated blank node names created in step 6.2 are out of order. I just wanted to make sure I understood the meaning of the phrasing "ordered by property" before I simply stop sorting the properties--which does make the test pass, but leaves me worried I'm missing something.

Read the latest editor's draft around Node Map Generation step 6.12 and compare its wording with the toRDF test c019. Determine whether “ordered by property” requires Unicode code point order or preserves insertion order, then clarify the specification so the intended behavior and test result are unambiguous.