[Bug 6469] New: [FT] TestSuite issues


           Summary: [FT] TestSuite issues
           Product: XPath / XQuery / XSLT
           Version: Candidate Recommendation
          Platform: All
               URL: http://basex.org
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Full Text 1.0
        AssignedTo: jim.melton@acm.org
        ReportedBy: christian.gruen@gmail.com
         QAContact: public-qt-comments@w3.org

Hi Jim, hi all,

in the following you find some new comments on the XQFT test suite (updated on
Jan 24). Once more I'm sorry that I didn't have time to put each bug in a
single thread, but feel free to split the list. Here we go..

[1] ft-3.2-examples-q5.xq:

As the scoring algorithm is implementation defined, results will vary. A lower
scoring value than 0.8 might cover more implementations; alternatively, queries
with scoring could be to a minimum, as long as no reference scoring model is
offered (otherwise these test queries won't make too much sense).

[2] ft-3.3-examples-q1.xq:

Attribute value should be atomized to conform with the proposed result, e.g.
like this:

old: //book[. ftcontains "usability" occurs at least 2 times]/@number
new: data(//book[. ftcontains "usability" occurs at least 2 times]/@number)

[3] ft-3.4-examples-q1.xq:

Query yields a boolean, whereas the result contains elements

[4] ftstaticcontext-results-q1.txt:

Result file is incorrect (superfluous: line 377-end)

[5] ftstaticcontext-results-q2.txt:

Result file is incorrect (superfluous: line 280-end)

[6] ftstaticcontext-results-q6.txt:

All books expected as result ('approved'/'approve' -> 'approv')

[7] FTPrimary-FTWords-any-q4b.xq:

0 results expected (<paragraphs/>), as the matching elements don't contain any
of the phrases ("FTAnyallOption weekend" and "voting specifies").

[8] FTPrimary-FTWords-anyword-q4b.xq

9 results expected inside the <paragraphs> element (XQFT specs, 3.2: "[...] the
tokens from all the strings are combined into a single set. [...]").

[9] FTPrimary-FTWords-anyword-q2b_result.xml

Space character missing at line 9, col. 206

[10] FTWords/*.xml

..Numerous single spaces missing in the result files (alternatively, the input
document should be modified).

[11] FTPrimary-FTWords-phrase-q3a.xq - -q4a.xq

Refering to the specs ("the tokens from all the strings are concatenated in a
single sequence, which is considered as a phrase"), I wouldn't expect results
for "{ "FTAnyallOption", "containment" } phrase". Instead "{ "how",
"containment" } phrase" would yield a result.

[12] FTOr-badexpr1.xq

I was wondering that "FTOr with empty sequence" is not allowed; why is that?
Where is it stated in the specification?

[13] FTNot-q1.xq

This one returns all <book> elements in which the <para> element does not
contain the word "Ninja". In the result, even the books are listed which have
no <para> element; I would only expect 4 instead of 9 titles as result.

[15] FTNot-q2.xq / -q3.xq / -q4.xq / -q5.xq

"<title>No Bad Software</title>" is missing in the result

[16] FTMildNot-or1.xq

"<title>Ninja Coder</title>" is missing as its <para> element contains the word

[16] FTMildNot-or2.xq

Here I would also expect "<title>Ninja Coder</title>" in the result as
("usability" not in "ninja coder") yields true. If I should be wrong, please
tell me why..

[17] FTMildNot-not1.xq

I'd propose "<title>No Bad Software</title>" and "<title>Ninja Coder</title>"
as result; the proposed result would be correct without "ftnot" in the query.

[18] ftwildcard-q2.xq

<book number="2"> is missing

[19] ftwildcard-q5.xq

<book number="1"> expected as result (as it contains the word "task")

[20] FTOrder-q3.xq / FTOrder-q4.xq

should be swapped, otherwise results are wrong

[21] FTWindow-words1.xq

I wouldn't expect results here, as "physical" ftand "swift" have two words in
between: "swift application of physical". So isn't this a window of 4 (instead
of 2) words?

[22] FTWindow-complexwords2.xq

Why does "window 0 words" return results? Same for other queries in this
section: can a window of 1 word really contain two words?

[23] FTScope-q2.xq / FTScope-q4.xq

If I got it right, each <p> element will be evaluated separately in this query,
yielding 0 results:

  "This is a simple example." ftcontains ("simple" ftand "complex") different
sentence -> false
  "It is not complex." ftcontains ("simple" ftand "complex") different sentence
-> false

But the following query would in fact return true (assuming that "." is a
implementation-defined sentence delimiter):

  <p>This is a simple example. It is not complex.</p>
      ftcontains ("simple" ftand "complex") different sentence

[24] FTContent-q2.xq

...will return all books except for <title>The Blues</title> (otherwise, ftnot
must be removed)

[25] FTContent-and1.xq

0 results expected, as the query does only cover the first and last tokens, but
not all that occur: "The secret" ftand "nice" entire content

[26] FTNot-unconstrained-q1.xq

Only 4 results expected (see [14])

[27] FTNot-unconstrained-q3.xq / -q5.xq

<title>No Bad Software</title> missing in result.

[28] FTSelection-FTTimes-...

xlink attributes and namespaces missing (<nt
xmlns:xlink="http://www.w3.org/1999/xlink" xlink:type="simple"...>)

[29] ft-3.4.5-examples-q1.xq

true expected (or lowercase option missing)

This time, I've just looked at ~25% of the tests; seems that appr. one third of
the tests are somehow buggy.. IMHO, debugging would be much faster if the
queries and results were shorter and focused on the actual problem. In many
cases, simple string tests ( la 'A' ftcontains 'A', etc.) could suffice to
check potential implementation bugs. Anyway, enough smart suggestions..

Thanks for your efforts!

Christian, BaseX Team 

Configure bugmail: http://www.w3.org/Bugs/Public/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.

Received on Sunday, 25 January 2009 04:06:50 UTC