[ANN] 4store version 1.1.2

[Steve Harris]

This is to announce the release of 4store version 1.1.2.

If is available for download as a tarball from http://4store.org/download/

A Mac DMG release will be made after the next Rasqal release, as this fixes some parser bugs.

It's recommended to use Rasqal 0.9.24 or newer with this release, to get all the features. If you use an older Rasqal then some of the tests will fail, though the system will be functional.

Release Notes:

* Added config file for configuring http server behaviour
(see http://4store.org/trac/wiki/SparqlServer#ConfigFile)

* Reverted fix for Content-Length bug, as it was causing serious side effects

* Fixed some memory leaks in various places, including one serious backend one

* Capped cache size in query engine

* Some debug tool improvements

* Some improvements to the CORS support

- Steve

[joseph padfield]

Hi,

I have just tried to build 4store version 1.1.2, with raptor2-2.0.0 and rasqal-0.9.24. All three configured and built without error, but then when I run "make test" in the 4store src folder I get the following error:

(cd query && pwd && ./setup.sh --autorun)

/home/joe/Downloads/4Store/4store-v1.1.2/tests/query

../../src/utilities/4s-backend-setup: error while loading shared libraries: libraptor2.so.0: cannot open shared object file: No such file or directory

../../src/backend/4s-backend: error while loading shared libraries: libraptor2.so.0: cannot open shared object file: No such file or directory

make[1]: *** [test-query] Error 127

make[1]: Leaving directory `/home/joe/Downloads/4Store/4store-v1.1.2/tests'

make: *** [test] Error 2

I have checked and libraptor2.so.0 exists and is readable:

/usr/local/lib$ ls -l libra*

-rw-r--r-- 1 root root 1934424 2011-02-28 16:02 libraptor2.a

-rwxr-xr-x 1 root root     981 2011-02-28 16:02 libraptor2.la

lrwxrwxrwx 1 root root      19 2011-02-28 16:02 libraptor2.so -> libraptor2.so.0.0.0

lrwxrwxrwx 1 root root      19 2011-02-28 16:02 libraptor2.so.0 -> libraptor2.so.0.0.0

-rwxr-xr-x 1 root root 1211184 2011-02-28 16:02 libraptor2.so.0.0.0

....

Am I missing something obvious ?

Thanks

Joe Padfield

--
You received this message because you are subscribed to the Google Groups "4store-support" group.
To post to this group, send email to 4store-...@googlegroups.com.
To unsubscribe from this group, send email to 4store-suppor...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/4store-support?hl=en.

[Steve Harris]

You may need to run "ldconfig" as root.

- Steve

[joseph padfield]

:-) 

Ok, tests running now, I had tried that before, but obviously not at the right time.

Thanks

Joe

[Steve Harris]

I've found that I needed to run ldconfig after installed raptor/rasqal recently too, but I'm not sure why. Often it's not necessary.

- Steve

[Karima]

Hi,

I have just tried to build 4store version 1.1.2, with raptor2-2.0.0
and

rasqal-0.9.25. And I had a problem with the lib rasqal-0.9.25.
So I have just tried to build 4store version 1.1.2, with raptor2-2.0.0

and
rasqal-0.9.24.
All three configured and built without error, but then when I

run "make test" in the 4store src folder I get the following bad
tests:
[FAIL] functions
[FAIL] post

Am I missing something ?

Thanks
Karima

[Karima]

Hi

There is the function COUNT in the tests "functions" ?

SELECT COUNT( ?a ) AS count WHERE {?a ?b ?c}
====>
400 Parser error
This is a 4store SPARQL server v1.1.2
parser error: syntax error, unexpected COUNT on line 1

Thanks
bye
Karima

[Steve Harris]

There is a COUNT test in the "count" test:

SELECT (COUNT(?a) AS ?count) WHERE { ?a ?b "Harris" } LIMIT 250000

Note the extra brackets.

- Steve

[Steve Harris]

Hm, I don't know why those would be failing.

There should be a file tests/httpd/results/functions-errs, tests/httpd/results/post-errs, and which will show why the test is failing.

There is an issue with older versions of glib, that a comparison is done wrong, which gives incorrect ordering in the functions test.

I don't know of any problems with the post test though.

- Steve

[Karima]

# cat functions-errs

13d12
< "17"^^<http://www.w3.org/2001/XMLSchema#integer>
15a15
> "17"^^<http://www.w3.org/2001/XMLSchema#integer>

# cat post-errs

13c13
< 201 imported successfully
---
> 411 content length required
16a17,19
> <test:a> <test:b> <test:c>
> <test:d> <test:e> <test:f>
> <test:d> <test:g> "text\nwith\tspaces"

# yum update glib
Loaded plugins: presto, refresh-packagekit
Setting up Update Process
No Packages marked for Update

--------------
I have just tried my old tests and I have for all response http 404 or
403.
I tested the simple query http get : http://localhost:8080/sparql/query=SELECT
?a WHERE {?a ?b ?c.}
==> 500 SPARQL protocol error This is a 4store SPARQL server v1.1.2

Hm,... strange -_-'

Bye
Karima

[Tim Schultz]

I actually can recreate the failed functions unit test as well with the exact same output in the functions-err file.

Don't have any issues with post however.  Looks like issue_6 also fails for me with "1d0 < 500 SPARQL server error".

Using RHeL5

- Tim

[Steve Harris]

On 2011-03-03, at 13:37, Karima wrote:

> # cat functions-errs
>
> 13d12
> < "17"^^<http://www.w3.org/2001/XMLSchema#integer>
> 15a15
>> "17"^^<http://www.w3.org/2001/XMLSchema#integer>

Yes, that's the glib problem.

> # cat post-errs
>
> 13c13
> < 201 imported successfully
> ---
>> 411 content length required
> 16a17,19
>> <test:a> <test:b> <test:c>
>> <test:d> <test:e> <test:f>
>> <test:d> <test:g> "text\nwith\tspaces"

Hm, I don't know what that is offhand.

> # yum update glib
> Loaded plugins: presto, refresh-packagekit
> Setting up Update Process
> No Packages marked for Update
>
> --------------
> I have just tried my old tests and I have for all response http 404 or
> 403.
> I tested the simple query http get : http://localhost:8080/sparql/query=SELECT
> ?a WHERE {?a ?b ?c.}
> ==> 500 SPARQL protocol error This is a 4store SPARQL server v1.1.2

Should be
http://localhost:8080/sparql/?query=SELECT...
to conform with the SPARQL protocol.

- Steve

[Steve Harris]

On 2011-03-03, at 14:13, Tim Schultz wrote:

I actually can recreate the failed functions unit test as well with the exact same output in the functions-err file.

Yep, will be the same glib version, RHEL 5's is old enough to have this problem. Need to look into a workaround.

Don't have any issues with post however.  Looks like issue_6 also fails for me with "1d0 < 500 SPARQL server error".

The fix for issue 6 has been backed out as it caused other problems. Martin G. is working on a fix, but I don't think it's been landed yet.

- Steve

[Nigel Wilkinson]

The function test fails for me as well. The exemplar file has the
following for one of the tests:
Query: SELECT (DAY(?d) AS ?dd) WHERE { <test:data> <test:bcdate> ?d }
ORDER BY ?d
?dd
"17"^^<http://www.w3.org/2001/XMLSchema#integer>
"23"^^<http://www.w3.org/2001/XMLSchema#integer>
"23"^^<http://www.w3.org/2001/XMLSchema#integer>

but if I change the query slightly to output ?d I get:
Query: SELECT (DAY(?d) AS ?dd) ?d WHERE { <test:data> <test:bcdate> ?
d } ORDER BY ?d
?dd ?d
"23"^^<http://www.w3.org/2001/XMLSchema#integer> "1969-05-23"^^<http://
www.w3.org/2001/XMLSchema#dateTime>
"23"^^<http://www.w3.org/2001/XMLSchema#integer>
"1969-05-23T23:05:00Z"^^<http://www.w3.org/2001/XMLSchema#dateTime>
"17"^^<http://www.w3.org/2001/XMLSchema#integer>
"2009-07-17T01:23:45+10:00"^^<http://www.w3.org/2001/
XMLSchema#dateTime>

Which looks to me to be the correct order for these dates.

- Nigel

[Steve Harris]

That's right, but the order is wrong in the first query, it should be 23, 23, 17.

I haven't yet figured out what causes the problem, but it seems to be related to the glib version.

I have 2.20 on my laptop, and that works fine, but 2.12 on one of the test boxes, and that gives the wrong answer.

- Steve

[Karima]

Hi

Steve, I have glib2-2.24.1-1.fc13.x86_64 on my laptop and I have the
same errors in the tests.
Is it another RPM ?
Karima

[Steve Harris]

Hum, interesting, it must not be related to glib then, I guess it was a coincidence. When I get a chance I'll look into it.

Either way, I wouldn't worry about it too much, it seems to be quite a subtle bug.

- Steve