Open Formula test cases

[Marijn Kruisselbrink <mkruisselbrink@kde.org>]

Hello list,

The editor-revisions of the openformula spec nicely contain lots of test 
cases, but when I tried to write a script to create a .ods file that 
automatically tests these test cases, I came across a number of problems with 
the test cases:

(all comments related to http://www.oasis-
open.org/committees/download.php/40740/OpenDocument-v1.2-cd06-part2-rev02-
editor-revision.odt, which I think is the latest version of this document, but 
I'm not entirely sure).

The test cases for nearly all byte-access string functions are incorrect, the 
order or parameters is in some cases wrong, and in other cases the wrong 
function is called (or the expected output is wrong)

Several test cases assume test data that is different from the data described 
at the end of the document. This is of course not really wrong, but it is 
rather annoying for automated testing (some of the complex number tests are in 
this category).

The tests for IMCSCH and CSCH have an expected output that is not correct.

=IMREAL(IMSUM(COMPLEX(1;2); COMPLEX(2;3) )) has as expected value 2, that 
should be 3 of course. (the description of this test case seems to indicate 
that the formula is wrong though, not the expected result)

A number of testcases for ISOWEEKNUM and WEEKNUM don't have a result defined.

Several test cases incorrectly use , as an argument seperator instead of ; 
(NETWORKDAYS, WORKDAY are examples of this problem)

Several test cases use , as decimal point instead of . (some of the ACCRINT 
tests, one of the COUPDAYS, some of the CUMIPMT tests for example)

Some small typos in two of the COUPPCD tests:
in =COUPPCD( “2004-12-31”; “2007-01-01”; 2; 1 )=DATE(2004;1;7) the second date 
should be DATE(2004;7;1) (month and day in wrong order)
in =COUPPCD( “2004-12-31”; “2007-01-01”; 4; 1 )=DATE(2004;1;1) I think the 
second date should be DATE(2004;10;1)


 I can see how the test cases not being part of the 'real' spec, makes this 
not quite top-priority, but it would be nice to get this fixed anyway, so I 
hope I send this to the right list.

Regards,
Marijn Kruisselbrink - Calligra Tables maintainer