Add OME model jobs to the devspace stack

[sbesson]

For the upcoming testing of the transform code-generation work.

This PR:

• add two more jobs preceding the Bio-Formats build for 1- merging all PRs opened against ome-model and 2- building the JARs using Maven
• removes the bf slave in favor of the generic testintegration for running all Java related build jobs
• adds Findbugs installation to the slave Dockerfile as required by the Bio-Formats job

[sbesson]

Currently tested in the context of ome/ome-model#14. With the corresponding branches allowing to couple the various versions, a low-level model PR can be propagated up the stack to the OMERO build for investigation.

One interesting side issue is that the Bio-Formats Ant build currently needs to be disabled as it cannot access the JAR built by the model jobs. I see two possible ways:

• either remove the bf container and unify all the builds under a single java container including all requirements for the Java stack
• or share the Maven repository across containers

[joshmoore]

No immediate objections to sharing .m2 across the containers. We'll just need to put some thought into not having 2 jobs write the same sub-directories at the same time. However, that being said, I don't necessarily see an advantage in having a bf container for itself, but you might.

[sbesson]

At least at the level of Bio-Formats, having Ant and Maven writing concurrently into the same shared directory is known to cause corruption. The ci-master strategy (like Trigger) is to call these 2 jobs successfully.

On the other side, while the bf container made sense in the context of the original introduction in #53, the model decoupling forces us to rethink the assumptions made at that time. I am fine with removing it in favor of the more generic java build slave. The requirements addition should be minimal (findbugs). Note this would likely count as a "breaking" change i.e. 0.4.0.

[sbesson]

Note the latter commit should implement #59 (comment), was successfully deployed on the transforms devspace and restored the success of the BIOFORMATS-ant. As noted above, if accepted, this proposal should likely constitue a 0.4.0 release as the structure of the containers is affected and the corresponding devspace Ansible role will need to be modified accordingly.

[sbesson]

For some unkown reason we are back to the encoding issue that I thought #7 had fixed. Using either this branch of the last tag of devspace:

[omero@seb-extendedmetadata ~]$ docker run -it devspace_testintegration bash
[omero@79fa5b547aa1 linux]$ locale
locale: Cannot set LC_CTYPE to default locale: No such file or directory
locale: Cannot set LC_MESSAGES to default locale: No such file or directory
locale: Cannot set LC_ALL to default locale: No such file or directory
LANG=en_US.UTF-8
LC_CTYPE="en_US.UTF-8"
LC_NUMERIC="en_US.UTF-8"
LC_TIME="en_US.UTF-8"
LC_COLLATE="en_US.UTF-8"
LC_MONETARY="en_US.UTF-8"
LC_MESSAGES="en_US.UTF-8"
LC_PAPER="en_US.UTF-8"
LC_NAME="en_US.UTF-8"
LC_ADDRESS="en_US.UTF-8"
LC_TELEPHONE="en_US.UTF-8"
LC_MEASUREMENT="en_US.UTF-8"
LC_IDENTIFICATION="en_US.UTF-8"
LC_ALL=
[omero@79fa5b547aa1 linux]$ exit
[omero@seb-extendedmetadata ~]$ docker run -it devspace_nginx bash
[omero@302a15c89215 nginx]$ locale
LANG=en_US.UTF-8
LC_CTYPE="en_US.UTF-8"
LC_NUMERIC="en_US.UTF-8"
LC_TIME="en_US.UTF-8"
LC_COLLATE="en_US.UTF-8"
LC_MONETARY="en_US.UTF-8"
LC_MESSAGES="en_US.UTF-8"
LC_PAPER="en_US.UTF-8"
LC_NAME="en_US.UTF-8"
LC_ADDRESS="en_US.UTF-8"
LC_TELEPHONE="en_US.UTF-8"
LC_MEASUREMENT="en_US.UTF-8"
LC_IDENTIFICATION="en_US.UTF-8"
LC_ALL=

[sbesson]

Closing for now as this conflicts with the master branch.