Complete enforcer setup in odlparent-lite

We need to specify the plugin version and ignore it in m2eclipse.

Change-Id: Ied2769679be6e58f0299120b513693d66d2185e5
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Add maven-antrun-plugin to m2e lifecycle mapping

Change-Id: Ib28a8b6d8e576186673ca1fa7637f0eda8c84659
Signed-off-by: Lorand Jakab <lojakab@cisco.com>

Move maven-enforcer-plugin from odlparent to odlparent-lite

That way it's detected earlier, e.g. Maven version mismatch is already
flagged up to end-users on their root parent project (which has
odlparent-lite as parent), and only a little later on the first Java
project (which have odlparent as parent).

Change-Id: I9c285fc45ffc0a82e8375a2b77dc89da19238033
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bump maven-javadoc-plugin version from 2.10.3 to 2.10.4

Not just for fun, but because I've just figured, through trial and
error, that this fixes a weird problem seen in the failed builds of
https://git.opendaylight.org/gerrit/#/c/46206/ - the older 2.10.3 seems
to have some bug which causes it to choke on finding references between
(two different kinds of) generated code in JavaDoc gen. and the 2.10.4
upgrade magically fixes it (I've locally tested this).

Note that we seem to have our maven-javadoc-plugin version defined in
two places, which is a bit of a shame (and very confusing at first).  I
wanted to get this bump merged in ASAP, with minimal side effect, so in
this commit have just changed both places.  A follow-up commit which
unifies this into a single place for the next time around would be cool.

Change-Id: Ie2c35f4a7dea357eed5ce8b72c9e757dbdf85565
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

odlparent/pom.xml: Parameter tasks is deprecated

remove maven warning by replacing tasks with target
> [WARNING] Parameter tasks is deprecated, use target instead

Change-Id: I6cf82b1760e711de60dd836024b6e67239214c2c
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>

Upgrade jsonassert 1.2.3 -> 1.3.0

Change-Id: Icf7717624601182e96ae21821d50131e059b63ac
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Upgrade webcohesion enunciate 2.1.1 -> 2.6.0

Enunciate 2.2.0 adds support for Spring Web annotations and fixes a
number of bugs:
https://github.com/stoicflame/enunciate/issues?q=milestone%3A2.2.0+is%3Aclosed

Enunciate 2.3.0 fixes a number of bugs:
https://github.com/stoicflame/enunciate/issues?q=milestone%3A2.3.0+is%3Aclosed

Enunciate 2.4.0 includes many bug fixes and enhancements:
https://github.com/stoicflame/enunciate/issues?q=milestone%3A2.4.0+is%3Aclosed

Enunciate 2.5.0 includes minor bug fixes and support for Spring's
@ModelAttribute:
https://github.com/stoicflame/enunciate/issues?q=milestone%3A2.5.0+is%3Aclosed

Enunciate 2.6.0 includes support for Bean Validation annotations and
Jackson annotations, and fixes a number of bugs:
https://github.com/stoicflame/enunciate/milestone/13?closed=1

Change-Id: Ic61354408b64089d228cb89a7f13e11874ecf5d5
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bump enforced Maven version to 3.3.9+

Also, remove enforcer section from karaf-parent,
as odlparent:odlparent already sets strong enough requirement.

Change-Id: I6952b1a5bfd98180ea9f6fc598029163780ffa9a
Signed-off-by: Vratko Polak <vrpolak@cisco.com>

Work around apparent Checkstyle bug re. JavaDoc in package-info.java

We have earlier agreed to, and aligned our Checkstyle rules accordingly,
not enforce JavaDoc presence, just validate its content if it's present.
However, we are still enforcing every package-info.java (of which we
have almost none..) to have a JavaDoc.

While experimenting with plastering package-info.java into EVERY package
everywhere (to have an annotation @NonNullByDefault package in it..),
I don't want to put in bogus empty JavaDoc into every package just to
satisfy Checkstyle, that would be ridiculous.

Related to https://github.com/checkstyle/checkstyle/issues/3416

Change-Id: I397763f9e09bc7db009ade5da1a5b35bdc518169
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bug 6790: use non-blocking /dev/urandom

Set Karaf up to use /dev/urandom (where available) instead of
/dev/random, to avoid blocking (and failing incoming SSH connections)
in low entropy situations.

Bug: 6790
Change-Id: I17c5681151a4bb6c61def952a7898b018d3ece86
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Revert "Bump akka to 2.4.10"

This reverts commit 6ef4fcc3b02cb261afa0fd6fb615e4f5f9002d95. (The upgrade breaks controller.)

Change-Id: I6af4777350c0254f95b2acc52a5657f54091984a
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bump akka to 2.4.10

Not so much it would impact us, but it is still good to keep
up with upstream.

Change-Id: Idaea34037e190378abe821be015506083306cd0a
Signed-off-by: Robert Varga <rovarga@cisco.com>

Add org.apache.karaf.shell.console

This is useful for projects implementing Karaf CLI commands.

Change-Id: Id08c52d66a1992ba1645fb1691be4f967088d82f
Signed-off-by: Stephen Kitt <skitt@redhat.com>

BUG-6341: add Cassandra and Coda Hale Metrics

(io.dropwizard is Metrics' new home. These releases are
backwards-compatible.)

Change-Id: Id8bfc25548238b806f10771b86f478249d84faef
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Enable Blueprint annotation support

This enables injection annotations everywhere. To activate Blueprint
annotation support (which generates an autowire.xml Blueprint file),
projects need to add

    <build>
        <plugins>
            <plugin>
                <groupId>org.apache.aries.blueprint</groupId>
                <artifactId>blueprint-maven-plugin</artifactId>
            </plugin>
        </plugins>
    </build>

and

    <dependency>
        <groupId>javax.inject</groupId>
        <artifactId>javax.inject</artifactId>
    </dependency>

The generated XML file is moved to org/opendaylight/blueprint so it is
loaded alongside other Blueprint descriptors.

Change-Id: I4423f691ca5c475e8d14f5210d61855a7ddc2328
Signed-off-by: Stephen Kitt <skitt@redhat.com>
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bug 6730: Upgrade Apache Shiro to the latest compatible release

Upgrade Apache Shiro to the latest release.  1.X stream contains
only compatible changes.

Change-Id: I3b89215842f469bea21fba3ca86589501cd516bc
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>

BUG-6712: fix bin/shell's classpath

bin/shell is duplicated from Karaf, but needs ${karaf.version} instead
of ${project.version} to define its classpath.

The correct long-term fix is to drop bin/shell and rely on upstream's
version (https://git.opendaylight.org/gerrit/45546); this patch is
intended for cherry-picking to Boron.

Change-Id: I91658bc9068119aa29d0b3a93d9fce53cfe9d9c6
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Intro. Google Guice DI with JSR-250 Extensions from Mycila

See https://github.com/mycila/guice/ & https://github.com/google/guice

See https://github.com/mycila/guice/issues/11 re. <version>4.0.rc1

This is already used in aclservice-impl for the first end2end test.

Intentionally <dependencyManagement> only with <scope>test, for now, as
it's not required at RT, yet.  (When it will be, then we'll look at OSGi
MANIFEST.MF, and a Karaf feature; later.  For the time being the
intention is to use Guice for more end2end style tests.)

Change-Id: Ia4c99a958ee0a5d687a311936693501a2a68c36d
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Upgrade karaf from 3.0.7 to 3.0.8

Change-Id: Ieb3e4424a84332d6bd4c1d12e62e6eba50440874
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>

Enable JSR-305 annotations everywhere

Change-Id: I1f0b62a629f206957cdcd928d1f0b5ae8cb97ec9
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Upgrade gson 2.3.1 -> 2.7

Change-Id: I09c99491b259939603639cdea165253ce2848ade
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bump akka-companion netty to 3.10.6

This is the latest bugfix release, possibly even
improving performance. Changelog is at
https://github.com/netty/netty/issues?q=milestone%3A3.10.6.Final

Change-Id: I67232fec57864cad01f8702f61e910ad098ed6f7
Signed-off-by: Robert Varga <rovarga@cisco.com>

Bug 6278: Move opendaylight-karaf-empty to odlparent

Move opendaylight-karaf-empty from controller to odlparent.  As with
moves of other artifacts, the following process will happen:

1) Copy opendaylight-karaf-empty into odlparent and adjust the groupId.
   This patch handles the copy work.

2) Change opendaylight-karaf-empty in controller to derive from
   odlparent's opendaylight-karaf-empty

3) Change references to use the artifact added in #1.

4) Deprecate and remove controller's opendaylight-karaf-empty.

This patch just handles the initial groundwork outlined in #1.

Change-Id: I824e61280784409194840ba54bdc0a6a10f0368f
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>

Remove a few unneeded braces

This avoids maven picking up certain properties (i.e., JAVA_HOME)
and replacing with build time value.

Change-Id: I5a7b0a7e5c8698cfb1469420b8bb358857c3e449
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>

Upgrade Karaf 3.0.6 -> 3.0.7

https://issues.apache.org/jira/secure/ReleaseNote.jspa?projectId=12311140&version=12334639
has the details; in particular this should fix
https://issues.apache.org/jira/browse/KARAF-4502 (the SSH key
negotiation failure).

This upgrade also pulls in newer versions of Aries Blueprint.

Change-Id: I475281cc9e422bfdd7574c23457259133593966d
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bug 6341 Remove copy of dependencies from karaf-parent

It appears that a lot of bloat is based on the manner in which the transition of
integration/distribution/distribution-karaf was transitioned to using karaf-parent.

There-in lies a tale.  Way back in Helium days... we used a 'dependency' hack for populating system/.
It basically just copied all the dependencies into system/ from the whole shebang.  This had two issues:

a)  Massive overinclusion , as you don't need most of the maven dependencies.
b)  Maven only ever copies *one* version of a thing via the maven dependency mechanisms... so *any* version skew borks you.

To fix this, in Li timeframe I wrote karaf-plugin.  Karaf-plugin walks the features.xml files
and makes sure that we have the needed bundles so all possible features can run.  This radically
reduces the overinclusion, as you are only pulling in the stuff you actually need.
It also solves the version skew problem... which is important because while we don't *want* version skew... it happens.

Because we were at the end of Li, and I did not wish to possibly break someone's
project local distribution, I *only* deployed karaf-plugin to integration/distribution, *not* to karaf-parent.

Since then, others have done good work on karaf-parent (and migrated it to odlparent).
Many thanks to them.  As part of that work, integration/distribution was migrated to use
karaf-parent (again, a good thing).

*But*... this suddenly caused integration/distribution to start doing massive overinclusion again (not good) :(
This appears to account for around 100mb of our size growth since Be.

The correct solution here IMHO would be:

a) To remove the dependency copying behavior of karaf-parent
b) Have the karaf-parent use karaf-plugin

Note: If you do #a without #b... you will likely break the project local distributions.

Sadly, doing #a and #b puts us back in the late breaking change that might impact the
project local distributions place I decided to avoid in Li.  However,
we now have a couple of releases of experience with the karaf-plugin, and thus the risk is somewhat lower.

This patch does #a and #b.

PatchSet #5: Fix karaf-plugin to also install bundles listed in startup.properties as well.

Change-Id: Ie4d99e86cf364fd4fc6d7d99622991133fa3e006
Signed-off-by: Ed Warnicke <eaw@cisco.com>
Signed-off-by: Robert Varga <rovarga@cisco.com>

Bump versions by 0.1.0 for next dev cycle

Change-Id: I5d0c92a382129f35ed2cefe407720a30d1e1332b
Signed-off-by: Thanh Ha <thanh.ha@linuxfoundation.org>

Add a workaround for Karaf not recognizing Windows 10

* Adjusts Karaf(Equinox) properties on startup to allow running on Windows 10
* This fix will be obsolete once we update Karaf to a version
   with w10-compatible Equinox
* Equinox commit needed for a permanent fix:
   https://github.com/eclipse/rt.equinox.framework/commit/5ebd8e732dcca11efc03ac4721e1909565caf679

Change-Id: Ide7cf4600a83dd379a7bb2811735e1980dc0887c
Signed-off-by: Martin Dindoffer <martin.dindoffer@pantheon.tech>

Add more logging from karaf

When a feature fails to install, we would really like
some context to the failure. This option will result
in an additional warning with exception being logged.

Change-Id: I44d44f1905e9b2f400dc31b33473b16d839b9874
Signed-off-by: Robert Varga <rovarga@cisco.com>

Add relativePath to karaf poms

Change-Id: I4176f2fdb4ed57d5c69c2b2eac01765c53ddf0e2
Signed-off-by: Vratko Polak <vrpolak@cisco.com>

Add an explicit comment about CPD being brain-dead

This is a follow-up to the reverted CPD commit. Appropriate
configuration is added to mdsal's binding parent.

Change-Id: I38616d2e1b7e4340d31bb977d07c2e7858a352ca
Signed-off-by: Robert Varga <rovarga@cisco.com>

Remove tcpmd5 even from comments

Change-Id: I5a4510d964b017871004c0096f2114775e113db4
Signed-off-by: Vratko Polak <vrpolak@cisco.com>

Remove a few unneeded braces

This avoids Maven picking up certain properties in some circumstances
(e.g. ${PWD} getting replaced by the current directory at
build-time).

Change-Id: I33e664605d9b715295fbc0214fa384f8c1177be6
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Revert "Do not run CPD on anything in target/generated-sources"

This breaks the lispflowmapping build.

This reverts commit ef60163bb91025f24881897e8916902187207d02.

Change-Id: I55715d0796620fe95371728babf4935909112db8
Signed-off-by: Lorand Jakab <lojakab@cisco.com>

BUG-6027: merge upstream realname fix

This patch syncs the startup scripts with Karaf 3.0.6 and adds the
upstream fix to handle symlinks (see
https://git-wip-us.apache.org/repos/asf?p=karaf.git;a=commit;h=5bbaf71814df514b112cbe8cd73ba57fefdc869f
for details).

Once we upgrade to a version of Karaf which includes this fix, then
client, shell, start, status and stop should be deleted.

Change-Id: I82efe6d29206c37bc46a08bc367342d63dea51c9
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Do not run CPD on anything in target/generated-sources

From layering perspective odlparent does not know anything
about mdsal, hence it must not make any special handling
for its behavior.

What it can do is to say that we simply do not CPD any
source that is generated -- which is actually the correct
thing to do.

Change-Id: If325f2c1da00fdba8f337f6b75fb860ac86797fa
Signed-off-by: Robert Varga <rovarga@cisco.com>

Add odl-triemap feature

The TrieMap bundle is being packaged by multiple projects.
provide a separate package so it can be reused.

Change-Id: If29bec69ef4b06fd32346d150a378c69be0b222d
Signed-off-by: Robert Varga <rovarga@cisco.com>

Add maven-cpd-plugin to m2e lifecycle mapping

Change-Id: Ie564815a81a28da608d5e1ef8577df97706a3c7a
Signed-off-by: Lorand Jakab <lojakab@cisco.com>

Add missing properties for startup scripts

The Karaf startup scripts use a number of Maven properties; this patch
copies the values used in Karaf 3.0.6.

Change-Id: I4d2a3e0ae98955c32cc07af2713f6c759a8c4fb5
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Do not specify MaxPermSize

In Boron we only support Java 8, which does not need
this deprecated option. Remove it.

Change-Id: Iad00670c2298ac105fa078d26e6520bd0bdf832b
Signed-off-by: Robert Varga <rovarga@cisco.com>

Bug 6187: Force more maven resolution in karaf-plugin

Especially concerning level 1 feature repository dependencies.
This fixed Bug 6187 in my local testing.

+ Improve logging.

Change-Id: Ie973e3669510061afd59a6c9d22f7a9161f0bd13
Signed-off-by: Vratko Polak <vrpolak@cisco.com>

Eliminate karaf.resources.version

This property is only used once, inline it in dependency
declaration.

Change-Id: I618ead3ad2a4232e8f41283ddd68649c78f2601d
Signed-off-by: Robert Varga <rovarga@cisco.com>

Bug 6278: Delete incorrect start script

This version doesn't work, the version provided by Karaf does.

Change-Id: I694e926d1917951708ed9740b51ad38f4b826de9
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Remove duplicate karaf-branding declaration

scope=compile is the default, hence elimite the duplicate
which emits a maven warning.

Change-Id: Id352a6a11008effec98371d8f4e48ec95df60b41
Signed-off-by: Robert Varga <rovarga@cisco.com>

Include README and CONTRIBUTING & PROJECT_INFO.yaml files in built JARs

Bug: 4134
Change-Id: I455c815270a2f71f12ecd78727461e4403009850
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Enable the license checker on src/test/** files

Should be merged once all the Boron projects merged their patches first.
See:
https://git.opendaylight.org/gerrit/#/q/status:open++topic:fix-license

Change-Id: I067c5034bf5b9897d030700f3956a0cbc2abce06
Signed-off-by: Thanh Ha <thanh.ha@linuxfoundation.org>

Bug 6278: Copy karaf-parent from controller to odlparent

As discussed in the MD-SAL call, there is an architectural need to move
karaf-parent from the controller project to the odlparent project.  This
is particularly useful for karaf upgrades, since right now a bump in karaf
version within odlparent requires a rebuild of controller to reflect the
change in karaf-parent, and our build jobs are not set up to support such
a process.

The move process will be handled in multiple steps:

1) Copy karaf-parent, karaf-branding and opendaylight-karaf-resources to
odlparent.  All three of these should belong in odlparent.  All three must
be moved since karaf-parent depends on the latter two artifacts.  Since
controller depends on odlparent (and not the other way around), they must
be moved upstream to odlparent.  [THIS PATCH]

2) Have controller's karaf-parent derive from odlparent's karaf-parent.
This preserves the ability for downstream consumers to derive from the
controller karaf-parent in the interim, while allowing changes to odlparent's
karaf-parent to be recognized since controller does not need to be rebuilt.

https://git.opendaylight.org/gerrit/#/c/42649/

This also involves removing karaf-branding and opendaylight-karaf-resources
from the controller project, since they are no longer needed.  There are two
consumers that need to be patched:
lispflowmapping: https://git.opendaylight.org/gerrit/42647
vtn: https://git.opendaylight.org/gerrit/42648

3) Change all downstream projects to utilize odlparent's karaf-parent.  This
is future work and will be done in several patches.

4) Remove controller's karaf-parent once we feel all downstream consumers
are using the odlparent's karaf-parent.

Change-Id: If2cdb2796a0649dbea5e289d35a95647eb3bdb60
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Upgrade pax-url-aether 1.5.0 -> 2.4.7

In particular this removes an old dependency on httpclient 3.1 which
suffers from CVE-2012-6153 and CVE-2014-3577 (SSL MITM attacks).

Add a few missing dependencies (which used to be pulled in by
pax-url-aether).

Explicitly scope feature-test's dependencies.

Change-Id: I9430f2f646ea050359a9c52bd978d8fb6c6a01e5
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Activate PMD CPD by default

The new Copy-Paste detector which I recently introduced in
c5cfbc434448b646fd93a89e0928111e2a4d500c and announced on
https://lists.opendaylight.org/pipermail/odlparent-dev/2016-July/000443.html
would probably be of more value if it was actually on by default, and
not just dormant in an pluginManagement doing nothing? ;-)  Note again
that this will never break a build - just log; unless enforcement is
enabled (same model as for Checkstyle).

Change-Id: I07b90dfb5a9c3f724eba77d2b87e08c25699c8b6
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

maven-compiler-plugin javac -parameters

Quote e.g.
https://docs.oracle.com/javase/8/docs/technotes/tools/windows/javac.html:
"Stores formal parameter names of constructors and methods in the
generated class file so that the method
java.lang.reflect.Executable.getParameters from the Reflection API can
retrieve them."

The XtendBeanGenerator successfully uses this Java feature, and in order
to be able to use this across all projects, this should be the new
default setting in odlparent.

AFAIK some IDEs can now also leverage this to offer better default
method parameter name e.g. on @Override.  Dito AFAIK some other
libraries, such as e.g. Jackson JSON stuff, may also be able to benefit
from this and need less annotations, if present.  I do remember that
e.g. Spring Framework can leverage this for it's REST thing - it saves
you from having to say: void foo(@ParamName("bar") int bar) kind of
thing.

I can't think of any real downside.. I'm guessing that it leads to an
ever so slight increase in .class and thus JAR file size (logical), but
that seems a very small disadvantage, and fair price to pay, for the
benefit.

Change-Id: I24015bd00b777c0a68fcf17316a13a86dd1251df
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Add sensible defaults for metadata required for Maven Central deployment

Change-Id: I15d09744aede2e54b5c4f63167777ab6b98600eb
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Checkstyle now also covers src/test/java & src/test|main/resources

But license headers are not enforced in src/test/java - not because that
wouldn't be nice, but just because it is way too much work to fix up
everything - I've tried, and then realized that's not realistic.

Change-Id: I66bcca1c8a54cb8a242579884a33032eba3ee100
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Checkstyle new rule ConstantName

Change-Id: I2497c871880139142fc79cb05b677eb2629dce94
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Checkstyle IllegalThrows with ignoreOverriddenMethods

The default is true, and for good reason; it was stupid of me to
originally have used ignoreOverriddenMethods = false.

This is because when you @Override a method, you cannot do anything
anymore about its throws clause - you have to comply with what you are
overriding.

Based on doing some real world Checkstyle clean up work and noticing
that this was wrong.

Change-Id: Ifbfd677074d7b84cbea52f5ca520791d648def75
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Add NoCleanIfFailure option to SingleFeatureTest

This will as shown much clearer errors in case of SFT failures; e.g.
"java.lang.IllegalStateException: Can't install
feature odl-integration-compatible-with-all/0.5.0-SNAPSHOT:
Could not start bundle mvn:org.opendaylight.snmp4sdn/snmp4sdn/0.4.0-SNAPSHOT
in feature(s) odl-snmp4sdn-snmp4sdn-0.4.0-SNAPSHOT: The bundle
"org.opendaylight.snmp4sdn_0.4.0.SNAPSHOT [504]" could not be resolved.
Reason: Missing Constraint: Import-Package: org.opendaylight.controller.sal.packet; version="[0.7.0,1.0.0)""

as shown on https://lists.opendaylight.org/pipermail/odlparent-dev/2016-July/000467.html

Change-Id: I5b5df8807e87c76975598a0f3311010b781227c6
Signed-off-by: Yakir Dorani <yakir.dorani@hpe.com>

Option to performance profile SingleFeatureTest

with Java Mission Control FlightRecorder

via -Dkaraf.featureTest.profile=true

Change-Id: I7450f44bb4543718253bda742d111c5050d5bf2b
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Quick (-Pq) should skip running tests, but not building them

Change-Id: Ie537fc006265da0492f8bf620a73280301af91e2
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Intro. Google Truth into dependencyManagement

see http://google.github.io/truth/.  This seems neater than Hamcrest in
particular for tests on Iterables (Collections).  For example, in
Hamcrest:

assertThat(Arrays.asList("foo", "bar"), contains("foo", "baz")); // OK

assertThat(Arrays.asList("foo", "bar"), contains(Arrays.asList("foo",
"bar"))); // NOK! :-( because contains only works for vararg not List

whereas with Google Truth:

assertThat(Arrays.asList("foo", "bar")).
containsExactlyElementsIn(Arrays.asList("foo", "baz"));

==> AssertionError: Not true that <[foo, bar]> contains exactly <[foo,
baz]>. It is missing <[baz]> and has unexpected items <[bar]>

also see http://google.github.io/truth/comparison

Change-Id: Iac121b302179e64c8cbd339b6f027582abfdd822
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Increase enforced Maven minimal version from 3.1.1 to 3.2.5 (no upper)

as discussed on [odlparent-dev] thread "Re: Introducing PMD Copy/Paste
Detector (CPD)"

Change-Id: I6c1e32029ba993849a368d832a48a442cd396053
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Introduce PMD Copy/Paste Detector (CPD)

Take II - is using <excludeRoots><excludeRoot> better? (We tried this
before and had to revert, despite it passing on Jenkins; very strange.)

Like Checkstyle, this by default only reports to the console during
(local..) builds, but does not enforce it - builds will NOT fail with
this. Then projects ready to "opt-in" for CPD enforcement can use:
<properties> <pmd.cpd.fail>false</pmd.cpd.fail> </properties>

Change-Id: I01de6ce01fe5f7020aae47ee2e9325cdc14e00e4
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Checkstyle java.nio.charset.StandardCharsets instead of com.google.common.base.Charsets

http://blog2.vorburger.ch/2016/06/how-to-use-checkstyle-to-enforce-using.html

Change-Id: Ibdc5520a7700f548e735c569ee8a3c1c9426dcd1
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Revert "Introduce PMD Copy/Paste Detector (CPD)"

This reverts commit 68afa051f302f5f67aedd42bf1ab2f4cc20d75c7.
It breaks build in yangtools: checkstyle fails to parse
configuration and cannot assign String to File[].

Change-Id: I5f9025c1dd79c19f10e8c17a56e270d17cbda043
Signed-off-by: Robert Varga <nite@hq.sk>

Upgrade Felix dependencies

* Dependency Manager 3.1.0 -> 4.3.0
* Dependency Manager Shell 3.0.1 -> 4.0.4
* File Install 3.1.6 -> 3.5.4
* Web Console 4.2.0 -> 4.2.14

Change-Id: I7d499669cf98e0761b92f7cce6b870ebc3f1d7c5
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Introduce PMD Copy/Paste Detector (CPD)

Like Checkstyle, this by default only reports to the console during
(local..) builds, but does not enforce it - builds will NOT fail with
this. Then projects ready to "opt-in" for CPD enforcement can use:

  <properties>
    <pmd.cpd.fail>false</pmd.cpd.fail>
  </properties>

BTW: The "minimumTokens" threshold may take some tuning - let's start
with these approx. 10 line rule, and then adjust it as we see how this
plays out in practice on a number of projects.

Change-Id: If862b8a16db5eed391bb85165987a4302ae44540
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Mark slf4j-api as provided

Since Karaf provides its own SLF4J encapsulation, it might make sense
to simply identify slf4j-api as provided.

Change-Id: I42f2854eec608155b9013ec6756053ea85104ba6
Signed-off-by: Stephen Kitt <skitt@redhat.com>

akka-testkit should be scoped to test

This is a testing framework for testing actor systems,
it should never be carried in production.

Change-Id: I32cd9af660463ee3d981acde29ff635a518b7a88
Signed-off-by: Robert Varga <rovarga@cisco.com>

Checkstyle remove duplicate UpperEll (minor)

Change-Id: I4a557c4b2dc629f6d0544c24b09aa09ea53b673e
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Introduce hamcrest-library, going beyond hamcrest-core, to the stars..

http://www.vogella.com/tutorials/Hamcrest/article.html
https://code.google.com/archive/p/hamcrest/wikis/Tutorial.wiki

Change-Id: If042c21216c62509c37176df647369b1fa2757c5
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Relax Checkstyle for JavaDoc, not mandatory for annotations

Because in at least one instance a project (ovsdb/library;
https://git.opendaylight.org/gerrit/#/c/40967/) has adopted Checkstyle
99% but then disabled enforcement because of lack of JavaDoc in a few
annotation types.  This is the wrong way to go about this IMHO; I'd
rather we relax this arbitrary requirement (that even annotations do not
have to have JavaDoc, like everything else), and be able to enforce
Checkstyle in more projects.

Change-Id: I42d1c81aa5ff511c272d39752340dddc5e3b0636
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Intro. Xtend dependencyManagement and Maven plugin

Xtend <http://xtend-lang.org> is already used in ODL's MDSAL project,
and I'd like to start using it in other projects as well (for writing
test code, for now).  In order to keep the version of Xtend used
uniform, and to keep child pom.xml shorter re. the xtend-maven-plugin
configuration, it would thus make sense to have this in the ODL parent.

Change-Id: I8356ef07772d3d487b688b5f6789c23a9b698bd2
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Add guava-testlib to dependencyManagement

Guava Testlib has a neat EqualsTester(), and other test helpers.

Change-Id: I843b3f724b545107f0ecf6e3651f2a583fbca820
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Add skip for pmd & cpd to Q

Even though PMD / CPD aren't actually used yet, I'm planning to propose
gradually introducing it, and having this already in Q will help not
forgetting it, and people not complaining.  It does not hurt having this
in early.

Change-Id: Ib6021da46129b5e5b9d6c2e79475fbf03ddda199
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Checkstyle rulez for Exception Handling Anti-Patterns

* http://checkstyle.sourceforge.net/config_coding.html#IllegalCatch
* http://checkstyle.sourceforge.net/config_coding.html#IllegalThrows
* http://checkstyle.sourceforge.net/config_design.html#MutableException

and TBD:
* http://sevntu-checkstyle.github.io/sevntu.checkstyle/apidocs/com/github/sevntu/checkstyle/checks/coding/AvoidHidingCauseExceptionCheck.html
* http://sevntu-checkstyle.github.io/sevntu.checkstyle/apidocs/com/github/sevntu/checkstyle/checks/coding/EitherLogOrThrowCheck.html

Includes fixing "Catching 'Exception' is not allowed" in
SingleFeatureTest.

Change-Id: I716bd43573c4431e0276141acfb3eec2195d5aa6
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Checkstyle enforce no redundant modifiers

see
http://checkstyle.sourceforge.net/config_modifier.html#RedundantModifier

Change-Id: I711de6ea9dc374435af5ba7403eef53ff2a99ce8
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Increase SingleFeatureTest timeout from 5 to 10 min.

Bug: 6156
Change-Id: I1f0a887e2e5fbaad9a934558cd13a28ab43ec434
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

SingleFeatureTest PerRepoTestRunner logging

Clarified; description includes feature name which is what you want to
show here, instead of just features.xml. System.out exceptionally used
because, whatever I did, I could never see that log otherwise..

Change-Id: I1cb1873e7eda3dcf2643888251f49cafec4e4dea
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

maven-enforcer-plugin requireJavaVersion 1.8.0

but we should probably move this to odlparent-lite, not just odlparent,
so that it already works for <packaging>pom roots.

Change-Id: Ide367d2697341e791bf1b9516a8937d1833cb35b
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Checkstyle System.out violations only in Java not in TXT etc.

In Java sources, still do not permit them even commented out.

Change-Id: If878b9e6954a62cb4a21496e6bb27962d3f59423
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bump akka to 2.4.7

The regression introduced in 2.4.6 has been fixed,
bump akka to latest.

Change-Id: I4b569a7f53af4dd77f7211ae9088983d16010f89
Signed-off-by: Robert Varga <rovarga@cisco.com>

Upgrade Maven plugins

(maven-surefire-plugin is left out because 2.19.1 breaks our tests.)

build-helper-maven-plugin 1.9.1 -> 1.10
 * 1.10: http://www.mojohaus.org/build-helper-maven-plugin/github-report.html

maven-bundle-plugin 3.0.0 -> 3.0.1
 * http://www.mail-archive.com/users%40felix.apache.org/msg16914.html

maven-clean-plugin 2.6.1 -> 3.0.0
 * 3.0.0: http://mail-archives.apache.org/mod_mbox/maven-announce/201510.mbox/%3C20151023060827.136D6188B5%40minotaur.apache.org%3E

maven-compiler-plugin 3.3 -> 3.5.1
 * 3.5.1: http://mail-archives.apache.org/mod_mbox/maven-announce/201602.mbox/%3CCAK2FUK3FfTSZSeQybK_4KXhFbtRNZ1uo6wxEbfnnyy5MCh1DsQ%40mail.gmail.com%3E
 * 3.5: http://mail-archives.apache.org/mod_mbox/maven-announce/201601.mbox/%3CCAK2FUK1kNz9di_SAUDBO2VjD3XLEL%3D0XLJkLkg68Q7x4ChJh-w%40mail.gmail.com%3E

maven-jar-plugin 2.6 -> 3.0.0 (pending
https://git.opendaylight.org/gerrit/40175)
 * 3.0.0: http://mail-archives.apache.org/mod_mbox/maven-announce/201605.mbox/%3C20160515194703.8B01819730%40minotaur.apache.org%3E

maven-project-info-reports-plugin 2.8.1 -> 2.9
 * 2.9: http://mail-archives.apache.org/mod_mbox/maven-announce/201603.mbox/%3C56DC1611.1050409%40apache.org%3E

maven-release-plugin 2.5.2 -> 2.5.3
 * http://mail-archives.apache.org/mod_mbox/maven-announce/201510.mbox/%3CCALhtWkezhk9ykpxzx-1T7qiyt7VZ88Nn1R%2BsSz9iK%2Buv3r%2B4aA%40mail.gmail.com%3E

maven-resources-plugin 2.7 -> 3.0.1
 * 3.0.1: http://mail-archives.apache.org/mod_mbox/maven-announce/201606.mbox/%3C20160608173951.D512D19E18%40minotaur.apache.org%3E
 * 3.0.0: http://mail-archives.apache.org/mod_mbox/maven-announce/201605.mbox/%3C20160521184859.A223F18263%40minotaur.apache.org%3E

maven-shade-plugin 2.4.1 -> 2.4.3
 * 2.4.3: http://mail-archives.apache.org/mod_mbox/maven-announce/201601.mbox/%3CCAKuVzBS2BeJoxXi3Yotbbyy5kGJemVSSCNr3%2BYW3ytz51G2wbQ%40mail.gmail.com%3E
 * 2.4.2: http://mail-archives.apache.org/mod_mbox/maven-announce/201510.mbox/%3C20151029055337.CB8AC18CD5%40minotaur.apache.org%3E

maven-site-plugin 3.4 -> 3.5.1
 * 3.5.1: http://mail-archives.apache.org/mod_mbox/maven-announce/201604.mbox/%3C7199325.eTch8dpi9I%40herve-desktop%3E
 * 3.5: http://mail-archives.apache.org/mod_mbox/maven-announce/201602.mbox/%3C1518487.5YI36Fm8fR%40herve-desktop%3E

maven-source-plugin 2.4 -> 3.0.0
 * 3.0.0: http://mail-archives.apache.org/mod_mbox/maven-announce/201602.mbox/%3C20160218044420.1FAD3185C6%40minotaur.apache.org%3E

propertymavenplugin 1.0-alpha-2 -> 1.0.0
 * http://www.mojohaus.org/properties-maven-plugin/github-report.html

Change-Id: I7e6a04f98d37d3beb1adc0b270f00c9bc323c0ef
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Update plugin property info

checkstyle.runtime.version, gmaven.plugin.version,
jar.plugin.version, and maven.release.version aren't used outside
odlparent, remove them.

Change-Id: Ic492b2f8ac01736c32485503c5418381423a650a
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Change Checkstyle allowedAbbreviationLength from 3 to 4 ("managerUUID")

Change-Id: I617a42df79b201d1def106cc98988c376bff74e0
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Fix Checkstyle warnings, and make them errors for the future

Change-Id: I8d98314114b3de7038e178cee82acc495a96c7e5
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Add apache xerces declaration

Dealing with Xerces is fun, except when we pull in wrong version of
xml-apis. As turns out, 1.4.01 is actually newer than 2.0.2 (based
on search.maven.org). Adjust the xml-apis version and declare xerces
dependency.

Change-Id: Iecdd82ab181768281d820d862378ffe30530cfd1
Signed-off-by: Robert Varga <rovarga@cisco.com>

Add mvn install -DaddInstallRepositoryPath=.../karaf/system

http://blog2.vorburger.ch/2016/06/maven-install-into-additional.html

Change-Id: Ia3a347250e22dbecef9558875e889a5a9fd2c5e3
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Checkstyle Import order as per original ODL convention on Wiki

see discussion on dev@lists.opendaylight.org email thread "Import
ordering: Going Wiki or Checkstyle way?"

Change-Id: I4568c68bea6a0aaeb358eb133d7201327a413b4d
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Use ${project.remoteProjectRepositories} in karaf-plugin

This Change is inspired by https://issues.jboss.org/browse/JBASMP-40

Without this, https://git.opendaylight.org/gerrit/#/c/40432/1 fails,
as karaf-plugin picks repositories defined for itself,
and not for the project it is supposed to handle.

Change-Id: I49fb3dd9b8d268ce3d52cae79bd2bbb0f8a06db4
Signed-off-by: Vratko Polak <vrpolak@cisco.com>

Checkstyle JavaDoc presence not enforced, but if present, check style

Change-Id: Ic924345a670e2938ee12b7c4f41bdc382aad2409
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Revert "Bump maven-failsafe-plugin version from 2.18.1 to 2.19.1, to solve NoClassDefFoundError"

This reverts commit fb0e070b9c331b4487ee476673f5104e0d4f12d9.

Change-Id: Iecebc9322bf63c45355600536a66dae94cd58d8a
Signed-off-by: Robert Varga <rovarga@cisco.com>

Bump maven-failsafe-plugin version from 2.18.1 to 2.19.1, to solve NoClassDefFoundError

required for https://git.opendaylight.org/gerrit/#/c/39484/

[INFO] ------------------------------------------------------------------------
[INFO] Reactor Summary:
[INFO]
[INFO] archetypes-parent .................................. SUCCESS [  2.431 s]
[INFO] odl-model-project .................................. SUCCESS [  9.970 s]
[INFO] opendaylight-configfile-archetype .................. SUCCESS [  3.365 s]
[INFO] distribution-karaf-archetype ....................... SUCCESS [ 24.224 s]
[INFO] opendaylight-karaf-features-archetype .............. SUCCESS [02:12 min]
[INFO] opendaylight-startup-archetype ..................... FAILURE [05:31 min]
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 08:23 min
[INFO] Finished at: 2016-06-14T14:25:26+02:00
[INFO] Final Memory: 50M/1025M
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal org.apache.maven.plugins:maven-failsafe-plugin:2.18.1:integration-test (failsafe-integration-tests) on project opendaylight-startup-archetype: Execution failsafe-integration-tests of goal org.apache.maven.plugins:maven-failsafe-plugin:2.18.1:integration-test failed: There was an error in the forked process
[ERROR] java.lang.NoClassDefFoundError: projects/basic/project/basic/it/target/test-classes/it/pkg/it/BasicIT (wrong name: it/pkg/it/BasicIT)
[ERROR] at java.lang.ClassLoader.defineClass1(Native Method)
[ERROR] at java.lang.ClassLoader.defineClass(ClassLoader.java:763)
[ERROR] at java.security.SecureClassLoader.defineClass(SecureClassLoader.java:142)
[ERROR] at java.net.URLClassLoader.defineClass(URLClassLoader.java:467)
[ERROR] at java.net.URLClassLoader.access$100(URLClassLoader.java:73)
[ERROR] at java.net.URLClassLoader$1.run(URLClassLoader.java:368)
[ERROR] at java.net.URLClassLoader$1.run(URLClassLoader.java:362)
[ERROR] at java.security.AccessController.doPrivileged(Native Method)
[ERROR] at java.net.URLClassLoader.findClass(URLClassLoader.java:361)
[ERROR] at java.lang.ClassLoader.loadClass(ClassLoader.java:424)
[ERROR] at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:331)
[ERROR] at java.lang.ClassLoader.loadClass(ClassLoader.java:357)
[ERROR] at org.apache.maven.surefire.util.DefaultScanResult.loadClass(DefaultScanResult.java:131)
[ERROR] at org.apache.maven.surefire.util.DefaultScanResult.applyFilter(DefaultScanResult.java:95)
[ERROR] at org.apache.maven.surefire.junit.JUnit3Provider.scanClassPath(JUnit3Provider.java:137)
[ERROR] at org.apache.maven.surefire.junit.JUnit3Provider.invoke(JUnit3Provider.java:88)
[ERROR] at org.apache.maven.surefire.booter.ForkedBooter.invokeProviderInSameClassLoader(ForkedBooter.java:203)
[ERROR] at org.apache.maven.surefire.booter.ForkedBooter.runSuitesInProcess(ForkedBooter.java:155)
[ERROR] at org.apache.maven.surefire.booter.ForkedBooter.main(ForkedBooter.java:103)
[ERROR] -> [Help 1]

Change-Id: I9d80dc095b21d1ef33e4da214d2399d778a902ac
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Introduce "mvn -Pq install" to just build JAR, but no tests, QA etc.

@see #also http://memory-alpha.wikia.com/wiki/Q ;)

Bumped maven-source-plugin from 2.4 to 3.0.0 just because
maven.source.skip appears to be broken with 2.4 (it gen. the
*-sources.jar anyway) but with 3.0.0 it's correctly "Skipping source per
configuration."

Change-Id: I77e521c60f8390df7b2e15182f1715e1a92db901
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Upgrade Apache Shiro 1.2.4 to 1.2.5

Fixes include the following:
[SHIRO-443] - SessionValidationScheduler created multiple times, enabling it is not thread safe
[SHIRO-462] - Authentication exceptions are swallowed
[SHIRO-467] - Authentication exception gets swallowed
[SHIRO-550] - Randomize default remember me cipher
[SHIRO-504] - Java 8 support
[SHIRO-516] - Explicitly specify the version of aspectjtools to avoid build warning
[SHIRO-562] - WildcardPermission calls String.trim() twice in setParts()

Release notes are located here:

https://issues.apache.org/jira/secure/ReleaseNote.jspa?projectId=12310950&version=12332994

Change-Id: Ic4a2342abecc31d7beb1b204adfba03bda78d3a8
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>

Revert "Upgrade Netty 4.0.37.Final -> 4.1.1.Final"

This reverts commit 34a2b5c5d7d09af6dad600f97332e05c65526022. Merged accidentally.

Change-Id: I9128e09fe147bb7059ce2c168f818715a0566338
Signed-off-by: Robert Varga <rovarga@cisco.com>

Upgrade Netty 4.0.37.Final -> 4.1.1.Final

See also https://lists.opendaylight.org/pipermail/odlparent-dev/2016-April/000269.html

Change-Id: I0403a36eebaa348c6ee1e52585a89eb2f6eb9ad4
Signed-off-by: Lorand Jakab <lojakab@cisco.com>

Fix use of BAD BAD BAD BAD static class initializer code block

This clarifies a problem we run into when doing the PAX Exam upgrade.
Fixing this will give us a much clearer error message the next time
anything related to this goes wrong.

Change-Id: I296d7475bf6fbb6390f0c08dafabd42fa1a9891d
Signed-off-by: Michael Vorburger <vorburger@redhat.com>
Signed-off-by: Robert Varga <rovarga@cisco.com>

Upgrade Netty 4.0.36.Final -> 4.0.37.Final

Fixes CVE-2016-4970, among other thing, see [1] for more.

[1] http://netty.io/news/2016/06/07/4-0-37-Final.html

Change-Id: I34cc70dbf5e9d12f8d3246cfcb4daea353635074
Signed-off-by: Lorand Jakab <lojakab@cisco.com>

Move 3 Checkstyle from controller/checkstyle to odlparent

These used to be in
controller/opendaylight/commons/checkstyle/src/main/resources/controller/checkstyle.xml,
but should better be here in odlparent odl_checks.xml:

* http://checkstyle.sourceforge.net/config_misc.html#UpperEll
* http://checkstyle.sourceforge.net/config_coding.html#EmptyStatement
* http://checkstyle.sourceforge.net/config_coding.html#EqualsHashCode

Change-Id: Ia6bded1dadc2e07620f8372f9d9323ec933299be
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Move Checkstyle module MethodName up in XML.

There must be order. Always.

Change-Id: I37fc452bac88267e2e8d745c82e109b95cda7e07
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Revert "Upgrade Pax Exam 4.4.0 -> 4.9.1"

This reverts commit 223a8cf3cb62c7873f3ba3de15728be06ae39010. It breaks
the features tests in yangtools.

Change-Id: Iac1614e003f89c0ff2f336718412f0a504134262
Signed-off-by: Robert Varga <rovarga@cisco.com>

Merge "Upgrade Pax Exam 4.4.0 -> 4.9.1"