To resolve the forking problem I suggest we remove the solder maven artifact resolver and
use the one from shrinkwrap, which is what i18n uses.
Sent from my iPhone
On Aug 23, 2011, at 6:52, Marek Schmidt <maschmid(a)redhat.com> wrote:
I have managed to run (more or less, with 1 failure and 4 errors) the new Solder
testsuite on AS7 (the jbossas-managed-7 profile)
with the following changes:
1. removing the <forkMode>always</forkMode> from
testsuite/internals/jbossas/pom.xml as it breaks the MavenArtifactResolver.
The other option for this issue is to hack the MavenArtifactResolver to deal with the way
surefirebooter works with forking, see the commented out parts in
2. add the "Dependencies: org.jboss.logging,org.jboss.logmanager" to the
seam-logger jar manifest, as no solder test archive gets deployed on AS7 without it.
3. Upgrade arquillian.version and arquillian.jboss7.version to 1.0.0.CR4, as I get
exception with CR1 for some strange reason.
org.jboss.seam.solder.test.resourceLoader.ResourceLoaderTest are broken because the
"targetContainerAdapterClass" hack doesn't work anymore.
On 08/22/2011 11:48 AM, Shane Bryzak wrote:
> Guys and girls,
> In the interest of moving things forward, I'm going to shortly start
> working on the beta 2 release. There are still some issues that we're
> working through with the multi-container testsuite configuration,
> however with some help I've been able to resurrect the existing tests in
> Solder to run in the latest version of Arquillian and I'm confident that
> I can get the tests in the other modules passing successfully without
> too much trouble. I would still like us to press on with the test suite
> work, however I don't want it to hold up the release schedule.
> Since the Beta1 release there have been a couple of important changes,
> which I'll quickly summarise here:
> 1) seam-parent is now at the top of the food chain and no longer
> inherits its configuration from weld-parent. This change is good for a
> couple of reasons, first of all the disassociation from the Weld project
> will help to reinforce that Seam is CDI implementation-independent.
> Secondly it gives us total control over configuration for the Maven
> plugins that we use, allowing us to update the plugin versions (which
> I've already done, to the latest version of each plugin) and fine tune
> the plugin configuration parameters for Seam.
> 2) I've decided against moving the logging feature out of Solder and
> into its own module for now, as I discovered that it was too tightly
> coupled to Solder to "de-tangle" it easily. I have however refactored
> the logging API classes into their own package, org.jboss.seam.logging
> which I think will be more intuitive for our users. I believe that the
> Solder reference documentation still needs to be updated in response to
> the changes made for SOLDER-102
) - Ken, could you please
> confirm? If so, would you be willing to update the reference docs to
> reflect the logging changes?
> I am aiming for the Beta2 release this coming weekend, which means with
> our 20+ modules now I need to get started pretty soon. If anyone has
> any spare cycles and is able to help out with this, it would be greatly
> appreciated. Each module needs to have its tests fixed, logging
> refactored to use the new logging API package, and its distribution
> thoroughly checked for correctness/completeness. If you'd like to
> volunteer for a module or two, please catch me on IRC and we can work it
> out there.
> seam-dev mailing list