Extraneous xmlns:e attribute

[Paul J. Lucas]

If you run the full-text query FTSelection-FTTimes-q1a.xq, the W3C expected result starts with:

<paragraphs>
  <p>
    <termdef id="dt-cardinality-selection" term="cardinality selection">

The actual result starts with:

<paragraphs>
  <p xmlns:e="http://www.w3.org/1999/XSL/Spec/ElementSyntax">
    <termdef id="dt-cardinality-selection" term="cardinality selection">

with the difference being the addition of the xmlns:e attribute of the <p> element.
The presence of the extra attribute causes the test to fail.

[Markos Zaharioudakis (markos-za)]

I believe it's the serializer who adds the xmlns:e attribute. So, I am transferring this Nicolae.

[Nicolae Brinza (nbrinza)]

There is no bug here, Zorba's result is correct, and W3C's expected result should be corrected.

[Paul J. Lucas (paul-lucas)]

Then open a bug with the W3C (and try to convince them it's an error), make this bug dependent on that bug, add that bug to the W3C Wiki bug page, and add this bug as an expected W3C failure in Zorba.

[Paul J. Lucas (paul-lucas)]

Sorry, for the last step I meant: ... and add the W3C test mentioned in this bug to the list of expected failing tests in Zorba.

[Nicolae Brinza (nbrinza)]

Well, why don't you do that? You definitely know more than I do about the FullText.

[Paul J. Lucas (paul-lucas)]

You claim the W3C result is wrong. You must have some evidence to support your assertion. You gave no such evidence here. Hence, only you know why you believe it's wrong. This is a serialization bug in a W3C test which you know more about. This has nothing to do with full-text other than it just so happens to be a full-text test's result that is wrong. It's about comparing the expected result, not the test itself. Since you know far more about serialization, you should do it. I wouldn't know what to say in trying to open a W3C bug.

[Nicolae Brinza (nbrinza)]

There is no bug in the serializer. Markos' statement is incorrect. You can run the test without the serializer -- run the test with the "apitest" and do not use the "-r" option. It will bypass the serializer entirely. You will see that the result contains the xmlns:e attribute declaration. Hence, this has nothing to do with the serialization.

There is no magic in the correctness of our result, and it needs no fancy explanation. The xmlns:e attribute is present in the root element of the original test XML file and it was simply erroneously omitted by whoever wrote the test result.

Since it is a bug in the XQFTTS and it has nothing to do with anything else, you should definitely be the one to do it.

[Paul J. Lucas (paul-lucas)]

Re-assigned per conference-call discussion.

[Nicolae Brinza (nbrinza)]

I have opened a bug with the W3C: http://www.w3.org/Bugs/Public/show_bug.cgi?id=11718 .

[Nicolae Brinza (nbrinza)]

Fixed in the XQFTTS.

[Paul J. Lucas (paul-lucas)]

This problem affects all the following tests. The bug was renamed to reflect this. The latest XQFTTS from CVS is apparently not fixed.

FTPrimary-FTSelection-q1
FTPrimary-FTWords-FTTimes-q1
FTPrimary-FTWords-q1
FTSelection-FTTimes-q1a
FTSelection-FTTimes-q1b
FTSelection-FTTimes-q1c
FTSelection-FTTimes-q1d
FTSelection-FTTimes-q1e
FTSelection-FTTimes-q2a
FTSelection-FTTimes-q2b
FTSelection-FTTimes-q2c
FTSelection-FTTimes-q2d
FTSelection-FTTimes-q2e
FTSelection-FTTimes-q3a
FTSelection-FTTimes-q3b
FTSelection-FTTimes-q3c
FTSelection-FTTimes-q3d
FTSelection-FTTimes-q3e
FTSelection-FTTimes-q4a
FTSelection-FTTimes-q4b
FTSelection-FTTimes-q4c
FTSelection-FTTimes-q4d
FTSelection-FTTimes-q4e
FTSelection-FTTimes-unconstrained-q1a
FTSelection-FTTimes-unconstrained-q1b
FTSelection-FTTimes-unconstrained-q1c
FTSelection-FTTimes-unconstrained-q1d
FTSelection-FTTimes-unconstrained-q1e
FTSelection-FTTimes-unconstrained-q2a
FTSelection-FTTimes-unconstrained-q2b
FTSelection-FTTimes-unconstrained-q2c
FTSelection-FTTimes-unconstrained-q2d
FTSelection-FTTimes-unconstrained-q2e
FTSelection-FTTimes-unconstrained-q3a
FTSelection-FTTimes-unconstrained-q3b
FTSelection-FTTimes-unconstrained-q3c
FTSelection-FTTimes-unconstrained-q3d
FTSelection-FTTimes-unconstrained-q3e
FTSelection-FTTimes-unconstrained-q4a
FTSelection-FTTimes-unconstrained-q4b
FTSelection-FTTimes-unconstrained-q4c
FTSelection-FTTimes-unconstrained-q4d
FTSelection-FTTimes-unconstrained-q4e
FTPrimary-FTWords-all-q1a
FTPrimary-FTWords-all-q1b
FTPrimary-FTWords-all-q1c
FTPrimary-FTWords-all-q2a
FTPrimary-FTWords-all-q3a
FTPrimary-FTWords-all-q4a
FTPrimary-FTWords-allwords-q1a
FTPrimary-FTWords-allwords-q1b
FTPrimary-FTWords-allwords-q1c
FTPrimary-FTWords-allwords-q2a
FTPrimary-FTWords-allwords-q4a
FTPrimary-FTWords-any-q1a
FTPrimary-FTWords-any-q1b
FTPrimary-FTWords-any-q1c
FTPrimary-FTWords-any-q2a
FTPrimary-FTWords-any-q3a
FTPrimary-FTWords-any-q4a
FTPrimary-FTWords-anyword-q1a
FTPrimary-FTWords-anyword-q1b
FTPrimary-FTWords-anyword-q1c
FTPrimary-FTWords-anyword-q2a
FTPrimary-FTWords-anyword-q2b
FTPrimary-FTWords-anyword-q3a
FTPrimary-FTWords-anyword-q4a
FTPrimary-FTWords-anyword-q4b
FTPrimary-FTWords-noanyall-q1a
FTPrimary-FTWords-noanyall-q1b
FTPrimary-FTWords-noanyall-q1c
FTPrimary-FTWords-noanyall-q2a

FTPrimary-FTWords-noanyall-q3a
FTPrimary-FTWords-noanyall-q4a
FTPrimary-FTWords-phrase-q1a
FTPrimary-FTWords-phrase-q1b
FTPrimary-FTWords-phrase-q1c
FTPrimary-FTWords-phrase-q2a
FTPrimary-FTWords-phrase-q3a
FTPrimary-FTWords-phrase-q4a

[Nicolae Brinza (nbrinza)]

Fixed.

[Paul J. Lucas (paul-lucas)]

I just tried the first test and it still fails for exactly the same reason.

[Nicolae Brinza (nbrinza)]

As I have previously said, it was fixed in the official XQFTTS. Try updating first.

[Paul J. Lucas (paul-lucas)]

I tried the test suite made by Sorin from the XQFTTS CVS and the tests still fail.

[Nicolae Brinza (nbrinza)]

I do not know when Sorin made the test suite. I have tried with the latest CVS and the tests are passing.