Bump netty to 4.1.15

http://netty.io/news/2017/08/25/4-0-51-Final-4-1-15-Final.html

Change-Id: Icc6415af8d73ab22f75fceb9f7d3fd574d3ebd0a
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Add net.sf.jung dependency declaration

Downstreams are using an old revision backaged from controller,
which is completely out-dated. Add a dependency declaration and
a proper feature.

Change-Id: I77b1e42aa8a21d94f96a967c0c21af497a65f18e
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Checkstyle AvoidHidingCauseExceptionCheck

http://sevntu-checkstyle.github.io/sevntu.checkstyle/apidocs/com/github/sevntu/checkstyle/checks/coding/AvoidHidingCauseExceptionCheck.html

Change-Id: I37d22c88869911f26daad023d66c5559a3cd1fa8
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

CS new FinalClass & HideUtilityClassConstructor

* http://checkstyle.sourceforge.net/config_design.html#FinalClass
* http://checkstyle.sourceforge.net/config_design.html#HideUtilityClassConstructor

Change-Id: I4ab4d4550c2bba4e25a19cae7a0c20aa1241757e
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Remove & forbid mockito-all + findbugs:jsr305

via the maven-enforcer-plugin

as all projects should only use findbugs:annotations, to avoid duplicate
classpath entries from FindBugs annotations which are present in both
:jsr305 as well as :annotations (which is more recent, bigger, and
maintained)

https://git.opendaylight.org/gerrit/#/q/topic:bug/7663, which replaces
all current existing usages of findbugs:jsr305 by :annotations in ODL
project, is the 1st pre-req. to this change.

https://git.opendaylight.org/gerrit/#/c/47337/, which changes the
dependencyManagement re. FindBugs, is the 2nd pre-req. to this change.

Related to Bug 7663 & Bug 7662.

Change-Id: Ifcd9c833af37fe52182dbc8604978fb82cdbcba7
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Add karaf.shell.core dependencyManagement

see also https://github.com/apache/karaf/pull/338

Change-Id: Ia1e732dfe5b83b5795bf022d8bdd89417f0c0d62
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

BUG-7767: Bump servlet-api to 3.1.0

This reverts commit 63c56783199d6fba319fc92373c1714dc8816d8e,
reinstating servlet-api 3.1.0. Also bump H2, which breaks unless
upgraded.

Change-Id: Icb141c95812b2d30a6a23801386a8e931c34b5be
Signed-off-by: Robert Varga <rovarga@cisco.com>

Bump guava to 23.0

https://github.com/google/guava/wiki/Release23

Change-Id: I1efaea3a901f93e4a95aabf13b7dacd4f544aad9
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Upgrade jolokia to 1.3.7

https://github.com/rhuss/jolokia/releases/tag/v1.3.7

Change-Id: I6a1174ba6a9621b66d495d92d87f05a4e75d0083
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Bump jettison to 1.3.8

https://github.com/jettison-json/jettison/compare/jettison-1.3.7...jettison-1.3.8

Change-Id: Ie1f260e7673d10822c1cc1f43eba32775f8810c9
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Remove equinoxSDK381 artifact declarations

These should not be used anywhere, the only downstream of these
is openflowplugin, which does not really need it.

Change-Id: I0c06aa5a951af78002a9e17545f610fe96d58589
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Bump maven-bundle-plugin 2.4.0/3.0.1 -> 3.3.0

Change-Id: I4f4909f69f57fcdfbe703a91251fd2c038d777ab
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bump depends-maven-plugin 1.3.1 → 1.4.0

Change-Id: I72ecfa4178042ce7bf994f717edad1e32a70823f
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Configurable path to place SFT heap dumps

Change-Id: Ib9e56f9147f972b4b28db2e7db7426402adb220e
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
(cherry picked from commit 3b3dc36e0af2d848a044f2a5841e10fcb3258e20)

Make SingleFeatureTest max heap size configurable

Currently, CSIT jobs of -only- type use 2GB of heap and CSIT
of -all- type use 3GB. But we are seeing occasional failures
in distribution, especially odl-integration-all feature can take
more heap than usual.

Address the TODO item by allowing maximum heap size to be specified
as a system property -- thus can be overridden by downstreams in
their pom.xml.

Change-Id: I84adda01f3bd20180c6d81d3da807d38b5af6fa1
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>
(cherry picked from commit 1aa0ac18166b3183e7733b50612f9706efa3c8b9)

Move karaf-util dependency

from bundles-test to features-test

The new karaf-util artifact introduced in
https://git.opendaylight.org/gerrit/#/c/61293/ should not be a
dependency of bundles-test but rather of features-test IMHO.

This is because bundles-test actually not have a compile time dependency
to the CustomBundleUrlStreamHandlerFactory from karaf-util; only
features-test (in PerRepoTestRunner).

However, bundles-test is also used in infrautils' ready.  Bumping to
odlparent 2.0.4 (which includes this) fell over with "...
odl-infrautils-ready ... Unable to resolve
org.opendaylight.odlparent.karaf-util...", because of the
<Embed-Dependency>bundles-test-lib in ready-impl.

We could add karaf-util to the <Embed-Dependency> of ready-impl to work
around this, at least temporarily to unblock, but the root cause is that
the dependency is wrong.

Change-Id: I031fbbc8fb9d8eebba3550ef0683b8363397d8a8
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bump netty to 4.1.14

http://netty.io/news/2017/08/03/4-0-50-Final-4-1-14-Final.html

Change-Id: Iff5310e47cfbfd4e12bb33ac38cebc1fe6c20ccc
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Eliminate triemap declaration

Triemap is forked and delivered in yangtools, this was a private
stop-gap build.

Change-Id: Ife6f1b29e28d7fea2e80b49dc6670d92458d3240
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Bump scala to 2.12.3

http://scala-lang.org/news/2.12.3

Change-Id: I39b18bd3abbd8aeaa6693f098599f9362cfe0023
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Remove opendaylight-karaf-empty scope=provided

Having a scope=provided breaks features-test, as it fails to
pull in its required resources, leading to SFT failure.

Change-Id: I1203f224f72f0678c7ccab976d4eb1ae530bc774
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>
(cherry picked from commit 3e11458209eee6a61cc9837928a016e82bab8866)

BUG-8923: use karaf-empty for SFT

Since we are excluding bouncy castle from generated features (and
we want to assume it is installed), we cannot run SFT against
vanilla karaf. This patch adjusts features-test to pull in our
pre-build opendaylight-karaf-empty instead of vanilla upstream.

Change-Id: Ic45ff713e90c0be7c232e0a1cee36b27c83efe95
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>
(cherry picked from commit 70f33da04cfdd2e9ddea9a322da5e0b9f6610106)

BUG-8923: create karaf-util

We need to use opendaylight-karaf-empty for SFT, but doing so
directly leads to a circular dependency, as karaf-plugin needs
sft (which needs karaf-empty, which needs karaf-plugin).

Resolve the dependency cycle by breaking out the single class
karaf-plugin requires into a separate artifact.

Change-Id: Id0649dd8203087ed82c1b936428cb6d9295f2957
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>
(cherry picked from commit 5d3c00c2b98de5ebb4a7eda6e5cc0734c86b74f8)

BUG-8923: Bouncy Castle is provided

We need to exclude the Bouncy Castle artifacts from generated feature
descriptors, since they’re provided in our startup configuration.

Change-Id: I26814b09efa37d12c1e567e5622f509eea9241be
Signed-off-by: Stephen Kitt <skitt@redhat.com>
(cherry picked from commit 580bee99d255788f8e0f879075b8a12cb92c607f)

BUG-8923: package javassist in a feature

Javassist is an optional dependency to netty and is required by
binding mdsal. This can lead to re-wires when mdsal is pulled in
after anything using odl-netty-4 is installed.

Change-Id: Ia5135dfd27dd6227d9897d1a2ab24e0078673d8c
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>
(cherry picked from commit e1b90dbbcdf55251192c2bf5b34c5a6f61e58590)

Fix tagGz property

karaf.archiveTagGz should be karaf.archiveTarGz, as used in the plugin
configuration.

Change-Id: I7173180a0e602867bd6a9c5b9fe1ac41a0a4ab6a
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bump immutables to 2.5.6

General house-keeping to keep us uptodate.

Change-Id: Ic8ce978f8274701677997ac2987ec483816d7842
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Add commons-text declaration

commons-lang3's WordUtils is deprecated in favor of commons-text
version. Downstreams should be able to migrate without having to
deal with versions.

Change-Id: I592f92b79372a21a8ada3225d05bed410a00363b
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Bump javassist to 3.21.0GA

https://github.com/jboss-javassist/javassist/compare/rel_3_20_0_ga...rel_3_21_0_ga

Change-Id: I56af3a4a7ecacd919300e5c2a0332e749b1ed708
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Bump akka to 2.5.3

http://akka.io/blog/news/2017/04/13/akka-2.5.0-released
http://akka.io/blog/news/2017/05/02/akka-2.5.1-released
http://akka.io/blog/news/2017/05/24/akka-2.5.2-released
http://akka.io/blog/news/2017/06/19/akka-2.5.3-released

Also leaks sun.net from the JDK, as it is required now :(

Change-Id: I65ab6b1a2024084281774cfb17ea1bdd54cd1848
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Bump jackson to 2.8.9

Version 2.3.2 is really ancient, bump it to 2.8.9.

Change-Id: I92b5aa15c49fe0742f0069a88328e700d2e04f82
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Add stax2-api declaration to match jackson

stax2-api is useful for others even without a provider, but
we cannot use 4.0.0 (as we'd like) because jackson has not upgraded
to it.

Change-Id: I69c15a360c2624fb842d4512accad67769de5d1d
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Add Javadoc links for third parties

Add some links for third party libraries declared as dependencies in
odlparent-lite so that those links are automatically added to the Maven
javadoc:aggregate or javadoc:javadoc goals. The list is not
comprehensive, it is just an starting point to continue adding more
links.

The projects/modules/aggregators referencing odlparent-lite will inherit
these links.

The <detectLinks> flag is set to false until [1] is resolved.

[1] https://issues.apache.org/jira/browse/MJAVADOC-427

Change-Id: I56a6eabe0ba7b9f32ffed7748ab13a79aa6bc2e1
Signed-off-by: David Suarez <david.suarez.fuentes@gmail.com>

Revert "Add an origin ID from the MDC into ..."

"... Karaf logs" until we agree on terminology etc.

This reverts commit 03609ff917ea5a187664614252a3e777108cb2ab.

Change-Id: I947fb2e44622f33e71614a66cf2ded816a61b26c
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Remove logback.xml

as well as the reference to it in etc/custom.properties

as discussed on https://lists.opendaylight.org/pipermail/odlparent-dev/2017-July/001262.html

Change-Id: I4423d3950e3d5f23be6d59423ef8b39abd0866a8
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Add an origin ID from the MDC into Karaf logs

see https://git.opendaylight.org/gerrit/#/q/topic:mdc related changes

and read https://logback.qos.ch/manual/mdc.html
or https://logging.apache.org/log4j/2.x/manual/thread-context.html

This change will show the OID in all logs, for example like this:

2017-07-27 23:14:16,034 | WARN  | 0000000000000 | ion(1)-127.0.0.1 |
InfrautilsIT                     | 93 -
PAXEXAM-PROBE-41ce6a29-abaf-44c0-8044-71d3e2bf99da - 0.0.0 | Ho, ho - do
you also see this funky Origin ID at the <== start of this log message:
0000000000000

If there is no originID available, then the new column will just be 13
empty spaces (for correct padding, like all other columns in the log),
like this:

2017-07-27 23:14:16,554 | ERROR |               | 0e6-c784b3cb4f5b |
core                             | 6 - org.apache.karaf.features.core -
4.0.9 | FrameworkEvent ERROR - org.apache.karaf.features.core

Change-Id: If18e1d35d71da5bd5c0fee5f2212915cccb0992e
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

BUG-7464: do not package upstream triemap

yangtools carries a reworked fork with a large number of bugfixes
and it is no longer using upstream. Remove odl-triemap, letting
downstreams consume the yangtools fork.

Change-Id: I5c04c5bed193594a57d4845f857ab62637b70ce4
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Bump versions to 3.0.0-SNAPSHOT

We are making backwards-incompatible changes, hence we need to bump
major version.

Change-Id: I1bbd7d55f3a4f1a8cb31e03be9e9215a7bbca4dd
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Bug 8771: add bcpkix to the startup bundles

This will avoid bundle refreshes due to wiring in bcpkix.

Change-Id: Ia7ceb119d96c8ffa642e3d427e3357c2b7e750d3
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bump scala to 2.12.2

This updates scala runtime to 2.12 and adjusts features accordingly.

Change-Id: I1cffaa9c4e7ef2e615d41e125d4d75a377acf537
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Bump netty to 4.1.13

http://netty.io/news/2017/03/10/4-0-45-Final-4-1-9-Final.html
http://netty.io/news/2017/04/30/4-0-46-Final-4-1-10-Final.html
http://netty.io/news/2017/05/12/4-0-47-Final-4-1-11-Final.html
http://netty.io/news/2017/06/09/4-0-48-Final-4-1-12-Final.html
http://netty.io/news/2017/07/06/4-0-49-Final-4-1-13-Final.html

Change-Id: I950a06bcb8e70f07803912a6d7c2f55eb67a0394
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Upgrade Jacoco Listeners 2.4 -> 3.8

Change-Id: I951403569556ded05380bb65e5e30addb643dc64
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Remove outdated src/site documentation

see https://lists.opendaylight.org/pipermail/odlparent-dev/2017-July/001218.html

Change-Id: Ie1580e11fa449ed38864960c868c4930f847e291
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Control karaf archive by parameters

    - add properties karaf.archiveZip and karaf.archiveTagGz
    - by default all archives are being created as before

See also: BUG-8765

Change-Id: I303bc2966d3436b7080d2ac47c7b2c735e537a40
Signed-off-by: Michal Rehak <mirehak@cisco.com>

${project.build.directory} in maven-bundle-plugin

Instead of hard-coding "target/" as directory name.
for the LICENSE file. This fixes the following probem:

   mvn -Pq,ide clean install

[INFO] --- maven-bundle-plugin:3.0.1:bundle (default-bundle) @ ... ---
[ERROR] Bundle ... : Input file does not exist: target/classes/LICENSE
[ERROR] Error(s) found in bundle configuration

Locally verified that with this fix, the problem shown above does not
happen anymore - and the LICENSE file is still in bundle JARs.

PS: In order to reproduce the problem above (and test this fix),
you (obviously) need to 'rm -rf target/ target-ide/' first... ;)

Change-Id: I313478ec95c9a6e647a595418bd92d45bbb04a8d
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bug 8799: populate repo first

To avoid overwriting patched files, we populate the local repository
before overwriting files with contents from
opendaylight-karaf-resources. This ensures that our patched files
survive karaf4-parent processing into projects’ distributions.

Change-Id: I4c00b493d50e56b117ec3c56ac3ffe2fa7814516
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bug-8798: ensure script permissions are correct

Since https://git.opendaylight.org/gerrit/59486 script permissions
aren’t preserved (from a tarball); this patch enforces permissions on
all the shell scripts.

Change-Id: Ie4ed6b31b20302ea925aeefc472e1c09a3295a1d
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Disable SFT on feature repos by default

Testing the features in repos can cause issues when they are
mutually-incompatble. Since we test features individually anyway, we
can safely disable SFT by default. skip.karaf.featureTest can be set
to false for projects which want to re-enable this.

Change-Id: I02ec0094985ccc5e7cc1f404808b156883c25911
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bump BC to 1.57

https://www.bouncycastle.org/releasenotes.html

Change-Id: I802837879550de85af80858bd1c62c8028d04f03
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Add Bouncy Castle to startup bundles

We rely on the provider being available, wire it into startup
properties to get that going.

Change-Id: Iee3a386a2d7f6e36caaaa03e846511ad4a213bbe
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Bug 8677: Do Karaf patch-ing earlier and only once

This changes the application of those patches over Karaf to happen once
during the opendaylight-karaf-resources build (as part of odlparent),
instead of in (each child of) karaf4-parent, as it used to be.

This is better because it:

 (a) solves mysterious issues with patch sometimes failing, see
https://lists.opendaylight.org/pipermail/odlparent-dev/2017-June/001136.html

 (b) avoids problems for Windows users (where there is no "patch")

 (c) makes Karaf local distro builds in projects (a bit) faster.

Change-Id: I16d912c6bc0d60295a787994a602e2753b37bfb2
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Remove unused k4.0.7 patches

Change-Id: I956ad203e1f8d17b5e5da61f44590e789ad705d8
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bump versions to 2.1.0-SNAPSHOT

Change-Id: I87425ce991c5277f60074121132254104704d813
Signed-off-by: Thanh Ha <thanh.ha@linuxfoundation.org>

Bug 8571: copy src/main/assembly

The contents of src/main/assembly need to be copied to
target/assembly, overwriting any conflicting file from Karaf.
integration/distribution needs this to be able to add custom scripts
and replace configuration files.

(Arguably this should be part of integration/distribution, but it
can’t be — at least not in a maintainable fashion.)

Change-Id: Ifcce82862d350b87853727e2c77140adee164886
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bump org.apache.aries.blueprint.core version

From 1.4.2 to latest 1.7.1, because in
org.opendaylight.controller:blueprint version 1.7.1 was already used
anyway, and indeed this appears to be the version used in our current
version of Karaf, and it seems wiser to declare this in only a single
place, to avoid possible version discrepancy problems re. this artifact
in the future.  So let's just bump this one, and drop the version in
controller:blueprint in a follow-up change.

Change-Id: Ib332fb88eab11a376e93ea90304e2a22115b3e85
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Remove '4' suffix in karaf[4]-plugin

Change-Id: I05c6241213d761f4764f7b7db8accb2e100d638c
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Drop odl-guava-18

Change-Id: I545bcacb92d7ede5c95a1f004869d3d35f0438ca
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Clean up POMs

The main point of this patch is to remove duplicates in the artifacts
POM. The patch also uses ${project.version} where appropriate.

Change-Id: I0554adb79bb7060bf0d45dbd3349440b7f808a7c
Signed-off-by: Stephen Kitt <skitt@redhat.com>

BUG-7446: switch advertized guava to 22.0

This is the build-time switchover patch to have downstream
pick up guava 22.

Change-Id: Ic547fb93ba718d51e6be6f6400b259beff83b8b8
Signed-off-by: Robert Varga <rovarga@cisco.com>
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Add antlr4 dependency declaration

This is used in yangtools mostly, but will be useful for anyone
dealing with lexers/parsers.

Change-Id: I5e1b232ca7357a6895162427d35736020859f266
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Bump Xtend 2.11 → 2.12

https://eclipse.org/Xtext/releasenotes.html#/releasenotes/2017/05/25/version-2-12-0

Change-Id: I538fed69d39af0c493d0eda6642601b1a28a3c7d
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bump to 2.0.0-SNAPSHOT

Since we’ve introduced breaking changes, we need to bump to
2.0.0-SNAPSHOT. (Thanks to Michael for pointing out the error of my
ways!)

Change-Id: Ic79385b234526ca0ab72180dbe2664351ed680ab
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bump git-commit-id-plugin 2.2.1 → 2.2.2

Change-Id: Id73ca8bee3b8b048ad3f277b2bc40dced19c6b37
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Remove '4' suffix in applicable artifact names

During the Karaf 4 migration, while both Karaf versions coexisted it
was necessary to have a different artifactId for Karaf 4 features and
feature repositories produced by odlparent. After Karaf 3 is removed,
this is no longer necessary, and this patch removes the '4' from
applicable directory, artifact and property names.

Change-Id: Ib6446a86a75a1a4b9637e73b71aca640a22807f4
Signed-off-by: Lorand Jakab <lojakab@cisco.com>

Remove all Karaf 3 Artifacts (now Karaf 4, only)

Change-Id: I4e60779869d8d4cb4760c1327f4abf238be98ddd
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bump PMD plugin

Change-Id: Ica5d0d9e0751ee058ed97d0f06a5704c2223f7ca
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Use groupId and version from parent

To keep things consistent inherit artifact group identifiers and
versions from their parent. Use the same version everywhere.

Note that checkstyle and license bundles are getting a version bump to
1.9.0-SNAPSHOT as a result.

Change-Id: I3640869b59cc1e452155e59d19a77e1abacd5454
Signed-off-by: Thanh Ha <thanh.ha@linuxfoundation.org>
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bump logback version

Bumps to latest 1.2.2.

Change-Id: I02a0bf1210197d4b6e59d5ba8eba17e9e5ee42e9
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Bump odlparent for next dev cycle

Change-Id: Iae69f160b89e59ade8f3efade03f2c18d3ba9806
Signed-off-by: Thanh Ha <thanh.ha@linuxfoundation.org>

Bug 8685: upgrade jolokia from 1.3.3 to 1.3.6

Changes are described in depth here:
https://jolokia.org/changes-report.html

Change-Id: Ifad9f981d893c686616c5b8ee509da997db5ed37
Signed-off-by: Ryan Goulding <ryandgoulding@gmail.com>

Add more pax-exam artifacts

These artifacts are used by downstream to run tests and need to
declare them as dependencies.

Change-Id: I1f41d44d13513f1abe8f3ae9a8663cd3665ef067
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Add Jackson dataformat XML

Change-Id: Ic3aa82a376765e0871369fe119b82d21ca80a319
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Update commons-lang3 3.5 → 3.6

Release notes:
http://www.apache.org/dist/commons/lang/RELEASE-NOTES.txt

Change-Id: Ib0b83a16d93600b746e8a7e938d9415912d398ea
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Remove <prerequisites><maven> avoid mvn WARNING

[WARNING] The project ... uses prerequisites which is only intended for
maven-plugin projects but not for non maven-plugin projects. For such
purposes you should use the maven-enforcer-plugin.

See https://maven.apache.org/enforcer/enforcer-rules/requireMavenVersion.html

Change-Id: I196af33c5a4fc80d6720add92171d5de58d2224c
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Remove @Deprecated bundles4test TestBundleDiag

as infrautils.ready does not use this anymore, now

Change-Id: Idd26ebd139d3ddb7ed8c8ef65c8301497f9158d4
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bump Checkstyle from 6.16 to 7.6.1

This (likely) requires some tweaks in projects enforcing CS;
needs testing using https://wiki.opendaylight.org/view/EvolvingCheckstyle

Motivated by having had to fix
https://git.opendaylight.org/gerrit/#/c/54978/

Change-Id: Ibef1cbb06e8e3d28a44b687562ebfc400d403b32
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Fix aggregator-parent relativePath

Fixes aggregator-parent's incorrect relative path to odlparent-lite.

Change-Id: Ibab245e65e6ce0583c6c58c219a40ac7dd2c3c7a
Signed-off-by: Thanh Ha <thanh.ha@linuxfoundation.org>

Bug 4290 Add missing org.osgi.service.event

Because apparently we need this, always (without it
opendaylight-karaf4-empty does not start anymore, following c/58053).

Change-Id: Icac86b301bda8136448ec45d3567fe728efc7f10
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

BUG-4219: fix karaf CLI reload

Controller's blueprint extenstion is pulling in an optional
dependency of pax-logging, causing it to reload, effectively
restarting everything. Wire osgi-service before starting up
to eliminate this.

Change-Id: Idf0170d805e98362e7f3560db7f166c68a18d19d
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Remove remote repositories

Our ODL Karaf distributions should never download artifacts from
maven.org or repository.springsource.com (we prepare an offline Maven
repo under the system/ directory in our Karaf distribution during
build).

This also adds the missing Apache Aries Quiesce API to system/ to fix
another problem; this may always have gotten downloaded at runtime, but
we never noticed this problem before removing the remote repos (above),
re. "Could not find artifact
org.apache.aries.quiesce:org.apache.aries.quiesce.api:jar:1.0.0"

I've noticed this while trying to work offline on the plane to SFO for
the DDF! ;)

Change-Id: Ib8d1390c463d22130e3dc46d27ac2951a808a8d6
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Relax Checkstyle RightCurly rule for LITERAL_DO

The change [1] try to avoid checkstyle forcing this
"...} while (condition)" statement to be split in two lines by
removing the LITERAL_WHILE from odl_checkstyle, but the correct
token should have been LITERAL_DO.

LITERAL_WHILE avoids forcing this [2]:
 while(line !=null)
 { // We do not want this
   process(line);
   line = in.readLine();
 }

LITERAL_DO avoids forcing the following [3]:
 do
 { // We do not want this
   x = rand.nextInt(10);
 } // neither this
 while(x < 5);

[1] https://git.opendaylight.org/gerrit/#/c/55264/
[2] http://checkstyle.sourceforge.net/apidocs/com/puppycrawl/tools/checkstyle/api/TokenTypes.html#LITERAL_WHILE
[3] http://checkstyle.sourceforge.net/apidocs/com/puppycrawl/tools/checkstyle/api/TokenTypes.html#LITERAL_DO

Change-Id: I2054e44f140e7b92480da954c45169a31931d683
Signed-off-by: David Suarez <david.suarez.fuentes@ericsson.com>

Better gitignore rule for vim tmp files

Old one just caught the first backup, this catches additional.

Change-Id: I01802aeee8b578c6cac25bcb626e06e3d194098d
Signed-off-by: Daniel Farrell <dfarrell@redhat.com>

Bug 8415 related: Make TestBundleDiag a lib

This is the alternative to c/56848 requested by Robert.

We're changing this so that we can consume bundles-test-lib in the
infrautils.ready implementation as a service for both Karaf v3 and v4.

As a (VERY nice!) side effect, this allows us to remove the duplicated
Java code we had between bundles-test and bundles4-test.  That was,
originally, intended to only be like it still is for a short while, but
with the Karaf 4 situation being what it is, this code duplication is
still there - and the respective code has started diverging a bit; so
unifying this in the new bundles-test-lib is great.

Change-Id: I0fd2c954fdc4a6c84505bef51729d15e4ba2888c
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Pin requests to <2.16

The new requests library pulls in a chardet dependency that is in
conflict with coala packages. Pin the version of requests to ensure
we're pulling in a supported version.

Copied change by Thanh.

Change-Id: I3b29f2ab809a1373a9e148120db9f004f3e3fba2
Signed-off-by: Daniel Farrell <dfarrell@redhat.com>

Pin coala to 0.11

coala-bears 0.11 just released 5 hrs from this patch author date. It
breaks the build because we were pinning coala. It seems coala and
coala-bears versions must match exactly so pin both versions for good
measure.

Change-Id: I044b283be027ec04930ff956583b16daa11bc950
Signed-off-by: Thanh Ha <thanh.ha@linuxfoundation.org>

Pin coala version to <0.11

I seems the new 0.11 version breaks us due to dependency conflicts.
Let's pin the version and migrate when we have time to troubleshoot.

Change-Id: I29b31fb12580e54d9a1761e11d3f75998ff7f43e
Signed-off-by: Thanh Ha <thanh.ha@linuxfoundation.org>

Bug 8176: Bump Pax Exam (and Pax URL) versions

to solve NoRouteToHostException on Fedora-based Linux

see https://ops4j1.jira.com/browse/PAXEXAM-740, and
http://blog2.vorburger.ch/2017/04/java-local-rmi-using-jdks-inetaddress.html

I've locally verified this does indeed fix the original problem.

Change-Id: Iaa4a7c74dc887b2105a80604afea498bb7deac60
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

blueprint-maven-plugin file move made conditional

This makes the move of the generated blueprint XML file conditional,
because we should not unconditionally hard code the move of the
blueprint XML files generated by the blueprint-maven-plugin.

This move from OSGI-INF/blueprint/autowire.xml to
org/opendaylight/blueprint should instead be conditional on the presence
of hand written XML in src/main/resources/org/opendaylight/blueprint
(which, typically, uses the ODL extensions).

This is because it should still be possible to use the @Singleton and
@Inject annotations with the XML generation by blueprint-maven-plugin in
bundles which do not depend on the MDSAL (e.g. in infrautils; I actually
have this case & problem in the real world).

Note that the gen. autowire.xml never uses the ODL extensions (unless we
do more work in the future to generate ODL extensions from annotations;
in which case the move should be made conditional on the XML content
actually containing/sing ODL extensions, not location).

Of course, normal ODL instance will always have the ODL controller mdsal
blueprint integration active.  But e.g. in ITs (e.g. in infrautils) it
seems "silly" to have to depend on the full mdsal JUST to get the
blueprint integration working JUST because we hard-coded and enforced
the move here.

This change has been tested by running a "mvn clean package" on e.g.
genius/interfacemanager/interfacemanager-impl, which is a typical ODL
bundle with a src/main/resources/org/opendaylight/blueprint/*.xml with
xmlns:odl="http://opendaylight.org/xmlns/blueprint/v1.0.0” as well as
some @Singleton in *.java, and verified that in the resulting JAR as
expected there are x2 XML in org/opendaylight/blueprint, namely the
handwritten say interfacemanager.xml as well as the generated
autowire.xml - both with and without this change; so no regressions.
Likewise, IFF there is no src/main/resources/org/opendaylight/blueprint,
then (only) the autowire.xml stays in OSGI-INF/blueprint (that's new,
and the goal of this change).

Change-Id: Id283078e0547e41caa0a714fe0f0fd70a048966c
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bug 8483: Fix maven-dependency-plugin to v3.0.1

In the (vain?) hope that this may prevent NPEs; see bug.

Change-Id: I21d355d82d68625d00651336e928fb8c0d2f35f7
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Introduce new aggregator-parent POM

Just to avoid having that <skip>true for maven-deploy-plugin and
maven-install-plugin copy/pasted all over ODL for each *-aggregator.

This also makes changes as appropriate to use it in odlparent itself.

The intention is to start shifting to using this in other projects too.

Having a new parent artifact for all aggregators MAY also come in handy
when I start looking in earnest at our JavaDoc generation.

Q & TODO: Some, but not all, of the typical *-aggregator projects also
have a maven-site profile with maven-javadoc-plugin - should we move
that here as well? I'm hesitant because that sometimes has
<inherited>false .. which I'm not very familiar with - it would not do
anything with <inherited>false in a <parent>, would it? So do we include
it in aggregator-parent, but without <inherited>false ?  Or forget about
it and just start with a simple aggregator-parent as is proposed here
for now (and refine JavaDoc related stuff later) ?

Change-Id: I33fba54051b5d3e5fec145d8ae68785bb66c8725
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bug 8415 related: TestBundleDiag API improved

required for https://git.opendaylight.org/gerrit/#/c/56749/

The implicit throws org.awaitility.core.ConditionTimeoutException and
java.lang.AssertionError were not great.. an AssertionError is special,
and more importantly Awaitility's ConditionTimeoutException is an
implementation detail.

Awaitility is actually bundled as a private OSGi embedded JAR-in-JAR,
which is a bit like shaded fat über JARs with relocation in non-OSGi
Java built by the maven-shade-plugin or maven-assembly-plugin; so it was
wrong to throw an exception from it to a consumer of this utility.

Therefore, this introduces a new SystemStateFailureException instead.

Change-Id: I7b48f5d39258a62c38a3fb8c98a6a35b97360276
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Bump akka to 2.4.18

http://akka.io/news/2017/05/02/akka-2.4.18-released.html
https://github.com/real-logic/Agrona/releases/tag/0.9.5
https://github.com/real-logic/Aeron/releases/tag/1.2.5

Change-Id: I2a19b3b2b1d35dba28e729938d6c579d55930681
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

javax.inject dependencyManagement <optional>true

Because at production run-time in Karaf we do not actually need these
annotations - they are only consumed by the blueprint-maven-plugin to
generate XML.  (In component tests with Guice we do need them at
runtime, but then infrautils.inject.guice.testutils enforces
<optional>false via c/56724.)

Thanks to FELIX-2809, resolution:=optional actually gets added to the
OSGi MANIFEST.MF by the maven-bundle-plugin if in the POM a dependency
is marked as <optional>true.

NB that even though javax.inject.Inject & Co. have
@Retention(RetentionPolicy.RUNTIME), this does work - because in Java
annotations can be absent at runtime when loading classes (contrary to
other linked classes; annotations are handled specially, see JLS).

This should be safe & backwards compatible to change, because it just
means we don't have to have javax.inject at runtime anymore, but for
those existing features which include a wrap:mvn:javax.inject already,
it normally should not cause any disruption either.

see also https://ops4j1.jira.com/browse/PAXCDI-230

Change-Id: I0246e5d150fd018a26fd3a9ca46b2c96b430bc94
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Introduce pax-cdi-api into dependencyManagement

This enables @Inject @OsgiService instead of <reference> in XML; see
https://wiki.opendaylight.org/view/BestPractices/DI_Guidelines#.40OsgiService_.26_.40OsgiServiceProvider

see c/56723 for full details re. why it is <optional>true

Change-Id: I1aa5f314cb561dbbd82cbd34c4812e0e1a803a4b
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

Add pax-url-link & pax-exam-link-assembly

into dependencyManagement of odlparent's POM.

I'm experimenting with using http://veithen.github.io/alta/ to use Pax
URL's link: protocol and (eventually) completely do away with the mvn:
URLs from pax-url-aether (which is a stupid approach, not by us/ODL but
originally by Pax Exam).

Merging this isn't urgent yet, as I can just put the version into the
POM of where I'm experimenting with this, but ultimately it would be
good to have this in odlparent; just thought I'd raise it early.

Change-Id: Ic372ad5db0db9c26bbfce2f14592408365a1d5d7
Signed-off-by: Michael Vorburger <vorburger@redhat.com>

skip.karaf.featureTest in single-feature-parent

Historically, the skip.karaf.featureTest property could be used to
skip SingleFeatureTest; that was lost in the migration to Karaf 4.
This patch restores the functionality in single-feature-parent.

Change-Id: Id6ed63129d7b9d7284356ca665152b321fc9ca16
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Fix pax-web's commons-collections

pax-web-4.3.0 can still pull in commons-collections-3.2.1, fix that.

Change-Id: Id0e0a678fcc54b45a864b69aed8ba7d1345bb756
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>
(cherry picked from commit 60cadf07b66747fd5e9823edcb8ef61b1f33b804)

Align pax-web and karaf standard features

Karaf's standard feature uses ow2/asm-all 5.0.4, while pax-web
is installing 5.0.2. Patch pax-web to use the same version, saving
around 230k.

Change-Id: Ifb29d1b3d3a54d4d55818795289b3eb0e357a826
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>

Use odl4-guava in odl4-akka-leveldb

Our Karaf 4 features should only use other Karaf 4 features.
odl4-akka-leveldb-0.7 ended up pulling odl-guava-18 instead of
odl4-guava-18... This patch fixes that, and changes odl4-guava-18 to a
dependency in odl4-akka-leveldb-0.7.

Change-Id: I1206579b51adaa159e0d629db3fbdbf03fcce455
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Update dependencies for Scala 2.11.11

Change-Id: Ibe0cfd9975d3f8e5d97ae040dba12f5532cce3c1
Signed-off-by: Stephen Kitt <skitt@redhat.com>

Bump scala to 2.11.11

Scala-2.11.11 is the last release in the 2.11 train, update to it.

Change-Id: I2d053d10b35b55efff0359f4303a66e8099e810b
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>