[rules-users] Re: ok - index problems - got a real simple test to try !
Michael Neale
michael.neale at gmail.com
Tue Apr 24 03:14:09 EDT 2007
ignore that, wrong mailing list ;)
On 4/24/07, Michael Neale <michael.neale at gmail.com> wrote:
>
> Really truly ;) , try this with junit, and JR 1.2.2 (1.3 also does it):
>
>
> public void testIt() throws Exception {
> Repository repo = new TransientRepository();
> Session sess = repo.login(new SimpleCredentials("foo",
> "password".toCharArray()));
> Node root = sess.getRootNode();
>
> for (int i = 0; i < 10; i++) {
> root.addNode( "" + i + System.currentTimeMillis(),
> "nt:unstructured" );
> sess.save();
> }
>
>
> Query q = sess.getWorkspace().getQueryManager().createQuery(
> "select * from nt:unstructured", Query.SQL);
> QueryResult qr = q.execute();
> NodeIterator ni = qr.getNodes();
>
>
> //check we get at least something back from the query
> assertTrue(ni.hasNext());
> assertNotNull(ni.next());
> assertNotNull( ni.next());
> assertNotNull(ni.next());
> }
>
> And start it up, try it about 10 times in a row, at least one time you
> will see a failure (ie it won't return anything from the query) - and then a
> subsequent time it just wont start, wait a few seconds, and then it will
> work again.
>
> Very very wierd, but somewhat relieved its not my own confused code, this
> is as simple as I can make it. I can only imagine there is some index sync
> issue still at play. Unfortunately, there is no way to wrap this up as a
> proper test, as it needs to startup fresh each time to reliably reproduce
> this (seperate JVM process each time).
>
> (and this will happen on any OS, I thought it was just linux, but have
> tried it on mac, windows etc).
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/rules-users/attachments/20070424/a10b7039/attachment.html
More information about the rules-users
mailing list