[redland-dev] Trying to compile redland on cygwin

Lauri Aalto laalto at gmail.com
Tue Nov 4 09:29:26 CET 2008


On Fri, Oct 31, 2008 at 11:29 AM, Martin Duerst <duerst at it.aoyama.ac.jp> wrote:
>>Looking at test-28-tuttle.ttl and test-28.out, test-28-turtle.ttl
>>seems to use a syntax close to N3, whereas test-28.out is expecting
>>something close to N-Triples. Also, test-28-turtle.ttl gives
>>'raw' numbers, whereas test-28.out expects double-quoted strings
>>followed by ^^<http://www.w3.org/2001/XMLSchema#decimal>. Also,
>>the order of the numbers is different, and some details of the
>>numeric notation are different.
>
> I commented out check-turtle-serialize-syntax in raptor/tests/turtle/Makefile
> for the moment.

Note that check-turtle-serialize-syntax compares to -out.ttl reference
files, not .out which are for parser tests. It's probably the same
end-of-line issue again.

> Moving one level higher, I got four errors in "Testing RDFA":
> 0094.xhtml   Triple counts: expected: 5, got 2
> 0101.xhtml   Triple counts: expected: 3, got 1
> 0102.xhtml   Triple counts: expected: 3, got 1
> 0103.xhtml   Triple counts: expected: 3, got 1
> But then I got the message "Ignoring expected failures", so
> I'm assuming these are not too much of a problem.

Yes.

> rasqal_decimal_test: Using double
> FAILED: (a+b)-b=1.234567888e9 expected 1.23456789e9
>  and two more immediately following.

This is expected if rasqal is built using no decimal library (gmp or
mpfr). Then it falls back to using regular double floating point
arithmetic which does not have enough precision for these test cases.
It can be safely ignored (and the test suite does not treat it as a
fatal error).

> Then things went south again. For "Testing SPARQL correct",
> it complained that it couldn't find XML/DOM.pm in @INC.
> I'm not exactly sure why I need Perl here, after all, we
> are not yet doing bindings, and I'm mostly interested in
> the Ruby binding. Any help appreciated.

The SPARQL test suite currently needs XML/DOM.pm for the check-sparql
perl script. Basically, XML::DOM::Parser is used for parsing reference
result files since the results cannot be compared byte-by-byte.

You can install the XML::DOM module and its dependencies from CPAN.

Lauri


More information about the redland-dev mailing list