6:11 a.m.
Hi,
I stumbled across an issue in Drools 5.1.1 that is still present in
Drools 5.4.0-Final, and which I raised (with a patch) as JBRULES-3630.
This issue is a blocker for what I need to do with Drools - the
correctness (or not) of the patch will determine whether what I am
trying to do is viable. Can anyone comment on this, please? Is this
really a bug in Drools, or have I completely misunderstood the
documentation instead?
Thanks,
Chris
1:53 a.m.
If you have a patch for review, please submit it as a pull request at github. Details of
how to this are here:
http://docs.jboss.org/drools/release/5.4.0.Final/droolsjbpm-introduction-...
Mark
On 22 Sep 2012, at 12:11, Chris Rankin <rankincj(a)googlemail.com> wrote:
Hi,
I stumbled across an issue in Drools 5.1.1 that is still present in
Drools 5.4.0-Final, and which I raised (with a patch) as JBRULES-3630.
This issue is a blocker for what I need to do with Drools - the
correctness (or not) of the patch will determine whether what I am
trying to do is viable. Can anyone comment on this, please? Is this
really a bug in Drools, or have I completely misunderstood the
documentation instead?
Thanks,
Chris
_______________________________________________
rules-dev mailing list
rules-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-dev
2:02 a.m.
I reviewed the patch, it would be nice if there was a unit test too for this - to ensure
we catch any future regressions.
Mark
On 23 Sep 2012, at 07:53, Mark Proctor <mproctor(a)codehaus.org> wrote:
If you have a patch for review, please submit it as a pull request at
github. Details of how to this are here:
http://docs.jboss.org/drools/release/5.4.0.Final/droolsjbpm-introduction-...
Mark
On 22 Sep 2012, at 12:11, Chris Rankin <rankincj(a)googlemail.com> wrote:
> Hi,
>
> I stumbled across an issue in Drools 5.1.1 that is still present in
> Drools 5.4.0-Final, and which I raised (with a patch) as JBRULES-3630.
> This issue is a blocker for what I need to do with Drools - the
> correctness (or not) of the patch will determine whether what I am
> trying to do is viable. Can anyone comment on this, please? Is this
> really a bug in Drools, or have I completely misunderstood the
> documentation instead?
>
> Thanks,
> Chris
> _______________________________________________
> rules-dev mailing list
> rules-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-dev
7:24 a.m.
On Sun, Sep 23, 2012 at 8:02 AM, Mark Proctor <mproctor(a)codehaus.org> wrote:
I reviewed the patch, it would be nice if there was a unit test too
for this - to ensure we catch any future regressions.
I can do that. You presumably need this patch against the "trunk",
although this issue is apparently present in all releases since 5.1.1.
Cheers,
Chris
4:06 a.m.
Chris,
I tried the reproduce your issue on the master branch with the very simple
test case (practically identical to the one you reported on the jira
ticket) that I am pasting below, but it works for me. I didn't try the same
on the 5.1.1 though. Anyway can you please confirm that my test should be
enough to reproduce your problem? Because, if so, I believe that, at least
on the master, this bug should be already fixed.
Thanks,
Mario
@Test
public void testClassLoader() throws Exception {
// JBRULES-3630
File jar = new
File("drools-compiler/src/test/resources/org/drools/compiler/xml/changeset/changeset.jar");
assertTrue(jar.exists());
ClassLoader classLoader = URLClassLoader.newInstance(new URL[]{
jar.toURI().toURL() }, getClass().getClassLoader());
Resource changeSet =
ResourceFactory.newClassPathResource("changeset1Test.xml", classLoader);
KnowledgeBuilderConfiguration conf =
KnowledgeBuilderFactory.newKnowledgeBuilderConfiguration(null, classLoader);
KnowledgeBuilder builder =
KnowledgeBuilderFactory.newKnowledgeBuilder(conf);
builder.add(changeSet, ResourceType.CHANGE_SET);
}
On Sun, Sep 23, 2012 at 2:24 PM, Chris Rankin <rankincj(a)googlemail.com>wrote:
On Sun, Sep 23, 2012 at 8:02 AM, Mark Proctor
<mproctor(a)codehaus.org>
wrote:
> I reviewed the patch, it would be nice if there was a unit test too for
this - to ensure we catch any future regressions.
I can do that. You presumably need this patch against the "trunk",
although this issue is apparently present in all releases since 5.1.1.
Cheers,
Chris
_______________________________________________
rules-dev mailing list
rules-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-dev
6:14 a.m.
On Mon, Sep 24, 2012 at 10:06 AM, Mario Fusco <mario.fusco(a)gmail.com> wrote:
I tried the reproduce your issue on the master branch with the very
simple
test case (practically identical to the one you reported on the jira ticket)
that I am pasting below, but it works for me. I didn't try the same on the
5.1.1 though. Anyway can you please confirm that my test should be enough to
reproduce your problem? Because, if so, I believe that, at least on the
master, this bug should be already fixed.
Thanks,
Mario
@Test
public void testClassLoader() throws Exception {
// JBRULES-3630
File jar = new
File("drools-compiler/src/test/resources/org/drools/compiler/xml/changeset/changeset.jar");
assertTrue(jar.exists());
ClassLoader classLoader = URLClassLoader.newInstance(new URL[]{
jar.toURI().toURL() }, getClass().getClassLoader());
Resource changeSet =
ResourceFactory.newClassPathResource("changeset1Test.xml", classLoader);
KnowledgeBuilderConfiguration conf =
KnowledgeBuilderFactory.newKnowledgeBuilderConfiguration(null, classLoader);
KnowledgeBuilder builder =
KnowledgeBuilderFactory.newKnowledgeBuilder(conf);
builder.add(changeSet, ResourceType.CHANGE_SET);
}
Mario,
Can you also test the following, please?
// Assuming that your change set contains resource
"classpath:data/IntegrationExampleTest.xls"
URL resource1 =
getClass().getClassLoader().getResource("data/IntegrationExampleTest.xls");
Assert.assertNull(resource1);
URL resource2 = classLoader.getResouce("data/IntegrationExampleTest.xls");
Assert.assertNotNull(resource2);
The point is that the JAR that you are dynamically loading must *not*
already exist on the classpath. My patch definitely applies on the
master branch, and so I strongly doubt that this issue is fixed there.
Cheers,
Chris
8:45 a.m.
Thanks Chris,
you're right, I found my mistake: I had a file in the classpath with the
same name of the one contained in the jar, so it was loading that file
instead of the one in the jar.
I just changed the name of the file and then I've been able to reproduce
that issue. I also checked that the fix you suggested works perfectly.
I pushed that fix on the master (but using my test since it is a bit easier
and doesn't require a change in the pom.xml) and backported it to the 5.4.x.
Mark, should I backport it to other (older) releases?
Mario
8:52 a.m.
Hi,
Thanks for pulling the fix. I think it should apply fine to the 5.1.1
branch too (which is what I need to use at work). I don't suppose you
could push out a drools-core-5.1.2 as well, please?
Thanks,
Chris
9:46 a.m.
On 23 Sep 2012, at 13:24, Chris Rankin <rankincj(a)googlemail.com> wrote:
On Sun, Sep 23, 2012 at 8:02 AM, Mark Proctor
<mproctor(a)codehaus.org> wrote:
> I reviewed the patch, it would be nice if there was a unit test too for this - to
ensure we catch any future regressions.
I can do that. You presumably need this patch against the "trunk",
although this issue is apparently present in all releases since 5.1.1.
please
against trunk. We don't generally do community patch releases. But the fix will go
into the next product release, but there is a core team to take are of this.
http://blog.athico.com/2011/04/drools-jbpm-community-versus-product.html
Mark
Cheers,
Chris
_______________________________________________
rules-dev mailing list
rules-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-dev
12:18 p.m.
On Mon, Sep 24, 2012 at 3:46 PM, Mark Proctor <mproctor(a)codehaus.org> wrote:
We don't generally do community patch releases. But the fix will
go into the next product release,
but there is a core team to take care of this.
OK, thanks. Who do I need to ask to get a drools-core-5.1.2 release
with this fix, please? I'm not sure we're ready to switch to
drools-5.4.1.Final yet, but my core won't work without the patch.
Cheers,
Chris
4471
days inactive
4474
days old
9 comments
3 participants
participants (3)
-
Chris Rankin -
Mario Fusco -
Mark Proctor