GitLab

Add a simple comment, just so that the pipeline is reran

FIXED separating gitlab runner

See merge request !109

debian stable and packaging now have their own runners on two different
VM hosts

Feature/vtk7 compatibility

See merge request !107

It seems that the "--rerun-failed" option has some bug
see for instance https://gitlab.kitware.com/cmake/cmake/issues/17767
and https://gitlab.kitware.com/cmake/cmake/issues/16314

CMake 3.0 does not have string(APPEND...)

These macros are used to specify how and why some automatic tests
need to be disabled

see https://gitlab.kitware.com/cmake/cmake/issues/17767

It seems that it is ok to use quoted test name when using
the WILL_FAIL property, but not when using the DISABLED
property (cmake bug?)

This merge is necessary in order to get the gitlab ci bug fix
to this branch

When a test fails or is expected to fail on a specific plateform
or a specific version of a dependency (VTK, ITK...) it should not
be marked as WILL_FAILED, it should just be disabled.

bug/xvfb-erratic-access

See merge request !105

Bug/xvfb erratic access

Closes #80

See merge request !104

Also tentative: do not listen to tcp or local unix

Bug/gitlab runner breaks ci

Closes #79

See merge request !103

So that it is exported to next bash script

gitlab-runner 11.7 fixes  a bug on windows. From now on,
CI_PROJECT_DIR uses \ (backslash) instead of / (forward slash)
Even if displayed otherwise when printing the bash environment,
CI_PROJECT_DIR (and therefore our custom PROJECT_SOURCE_DIR, 
variable, which is created in .gitlab-ci.yml from CI_PROJECT_DIR)
contains backslash.
When used from inside a bash shell (such as configure.sh), this
confuses ctest. Backslashes have to be replaced by forward slashes.

Resolve "change the Q_PLUGIN_METADATA(IID ... string in cep wizzard"

Closes #76

See merge request !101