fn:format-number failures (at least 19)

[Sorin Marian Nasoi]

The following 19 test cases fail in "fn-format-number" test set:
"numberformat14,numberformat15,numberformat34,numberformat38,numberformat40,numberformat41,numberformat42,numberformat60a,numberformat60m,numberformat60o,numberformat60q,numberformat63,numberformat64,numberformat70,numberformat71,numberformat81,numberformat83,numberformat87,numberformat88"

The attached report was generated using:
./zorba -f -q ../../test/fots_driver/cli.xq -e mode:=run-test-sets -e testSetPrefixes:=fn-format-number -e verbose:=true -o results_fn-format-number.xml --indent --disable-http-resolution

[Paul J. Lucas (paul-lucas)]

paulbook<1>$ ./zorba -f -q ../../test/fots_driver/cli.xq -e mode:=run-test-sets -e testSetPrefixes:=fn-format-number -e verbose:=true -o results_fn-format-number.xml --indent --disable-http-resolution
'testSetPrefixes' was set to: [1]: fn-format-number
'testCasePrefixes' was set to: [1]:
'dependency' set to: [1]:
'verbose' set to: [1]: true
Cli command was set to: [1]: run-test-sets
Path to FOTS catalog.xml was set to: [1]: ../../build/test/fots/2011/QT3-test-suite/catalog.xml
Path to FOTSZorbaManifest set to : [1]: FOTSZorbaManifest.xml
Path to ExpectedFailures.xml set to: [1]:
</Users/pjl/src/flwor/zorba/repo/trunk/zorba/test/fots_driver/fots-driver.xq>:418,5: user-defined error [err:FODC0002]:
Please make sure the passed 'fotsPath' points to the exact location of the FOTS catalog.xml:
file:///Users/pjl/src/flwor/zorba/repo/trunk/zorba/build/test/fots/2011/QT3-test-suite/catalog.xml
 that the passed 'fotsZorbaManifestPath' points to a file in the same folder as cli.xq:
file:///Users/pjl/src/flwor/zorba/repo/trunk/zorba/test/fots_driver/FOTSZorbaManifest.xml
 that the passed 'expectedFailuresPath' points to the ExpectedFailures.xml file:
; raised at /Users/pjl/src/flwor/zorba/repo/trunk/zorba/src/runtime/errors_and_diagnostics/errors_and_diagnostics_impl.cpp:81

[Sorin Marian Nasoi (sorin.marian.nasoi)]

http://my.zorba-xquery.com/dokuwiki/doku.php?id=zorba_fots_testing

Beforehand please run:
make fots-import
make fots-activate-sets

[Paul J. Lucas (paul-lucas)]

I don't understand the results. For example, the output file says that the test "numberformat14" failed. First, it doesn't say what the expected test output is. Second, if I run the query by hand:

    xquery version "3.0";
    declare default decimal-format infinity="off-the-scale" ;

    format-number(1 div 0e0,'###############################')

I get the result:

    Infinity

which seems like it could be the correct answer.

[Sorin Marian Nasoi (sorin.marian.nasoi)]

Hi Paul,
first of all please read the thread "new FOTS bugs opened" on @zorba-dev.

Now, about your questions:
1) sry. but I uploaded the wrong report: that is fixed now
2) wrt. test "numberformat14"

Please check out the 'info' node in the detailed report:

<fots:info>
 <fots:query>[...]</fots:query>
 <fots:expected-result>
    <assert-string-value>off-the-scale</assert-string-value>
 </fots:expected-result>
 <fots:result>Infinity</fots:result>
 <fots:errors>'assert-string-value' returned: result different from expected result.</fots:errors>
</fots:info>

The 'expected-result' is an 'assert-string-value' with value 'off-the-scale'.
The actual result in inside 'result' node, meaning string 'Infinity' in this case.
The 'errors' node describes the errors returned when the assertions were run, in this case 'assert-string-value'.

Now, as you can see, the test case declares a decimal format (please see http://www.w3.org/TR/xquery-30/#id-decimal-format-decl).
In this case, this means that instead 'Infinity' one should return 'off-the-scale'.

Please send me an email if you have any more questions/remarks.

HTH.

[Paul J. Lucas (paul-lucas)]

Can you cite the spec that says that the expected result MUST be "off-the-scale"?

[Sorin Marian Nasoi (sorin.marian.nasoi)]

Paul I am not saying that the tests are correct, as already stated into one of the wiki's I created:
http://my.zorba-xquery.com/dokuwiki/doku.php?id=zorba_fots_driver#reporting_bugs

The test cases fail because one of the following 3 situations (or a combination of them):
- FOTS driver is wrong
- Zorba returns incorrect result
- test case in FOTS is wrong

I only triaged around 700 out of the 1011 failing FOTS tests and created some bugs assigned to the persons I thought know more about the code in question.

The bug was assigned to you because, IMHO, you know that code better than the rest, so you can make the decision.

In this particular case, if I am not missing something, Zorba returns incorrect result because:

According to:
http://www.w3.org/TR/xquery-30/#id-decimal-format-decl
"A decimal format declaration adds a decimal format to the statically known decimal formats, which define the properties used to format numbers using the fn:format-number() function"

and in
http://www.w3.org/TR/xquery-30/#dt-static-decimal-formats :
"[Definition: infinity specifies the string used for the infinity-symbol; the default value is the string "Infinity"]"

HTH

[Paul J. Lucas (paul-lucas)]

Your comment doesn't answer my question. I can not find anywhere in any XQuery specification where it says the result should be the literal string "off-the-scale". If I am correct and no such result MUST be returned, then I don't understand how you can think that Zorba returns the incorrect result.

[Sorin Marian Nasoi (sorin.marian.nasoi)]

The query in question is:

xquery version "3.0";
declare default decimal-format infinity="off-the-scale" ;
format-number(1 div 0e0,'###############################')

right?

The meaning of the second line:
declare default decimal-format infinity="off-the-scale" ;
according to the specs cited above is to replace the default value ("Infinity") used for the infinity-symbol with the string "off-the-scale".

What am I missing: are the cited specs optional ?

What do you mean by "I can not find anywhere in any XQuery specification where it says the result should be the literal string "off-the-scale""
I just pointed out the spec twice...

[Paul J. Lucas (paul-lucas)]

Never mind.

[Sorin Marian Nasoi (sorin.marian.nasoi)]

I do not get it ...

Please let me know: what did I miss ?