On backend side headers are asserted in tests:

 * X-Forwarded-For
 * X-Forwarded-Proto
 * X-Forwarded-Port
 * Host

In order to pass cleanly X-Forwarded-For from the frontend to the backend,
it's passed as X-Forwarded-For-Real in case of cached slaves.

Noted problem with IPv6 endpoint was used, as in this case 6tunnel IP would
be used.

QUIC is not used at all, and became superseded by HTTP/3

Customized configuration support is not used since introduction of
Caddy software, so there is no need to support it anymore.

This reverts commit ef5aca54.

software-shared.cfg is used on master from scalability tests, let's keep
it for now.

Install setuptools_scm before jsonschema (which has it in setup_requires) so that the version can be controlled by buildout

See merge request !768

We don't need a section here, we inherit this from jupyter stack

See merge request !756

Since e5687602 it's instance
responsability to generate a fonts.conf if they need fonts, but we
realized to late that if there is no /etc/fonts/fonts.conf on the
machine firefox does not start.

In recent ERP5 software release, we have a $FONTCONFIG_FILE, but on
older versions we don't.

To keep compatibility with old ERP5 software release, we generate a
default FONTCONFIG_FILE that will be used as fallback when
$FONTCONFIG_FILE environment variable is not set.

As we have kvm with a section deployed using
collective.recipe.shelloutput, if collective.recipe.shelloutput isn't in
the SR folder anymore this section can't be uninstalled (as the
uninstall method is defined in the egg).

The output of buildout for this error was :

Installing switch_softwaretype.
Uninstalling frontend-port.
While:
  Installing.
  Uninstalling frontend-port.
  Installing recipe collective.recipe.shelloutput.
  Getting distribution for 'collective.recipe.shelloutput'.
Error: We don't have a distribution for collective.recipe.shelloutput
and can't install one in offline (no-install) mode.

The cause was a7a36453

See merge request !700

now that we have a patch to support gcc 8 we should support all known
versions of gcc

gcc-5.5 compiles a bit slower python and we are observing failures in
ERP5 performance tests.

Fix build errors on gcc 8:

   ../glibmm/threads.h: In member function 'GPrivate* Glib::Threads::Private<T>::gobj()':
   ../glibmm/threads.h:661:29: error: cannot convert 'GPrivate' {aka '_GPrivate'} to 'GPrivate*' {aka '_GPrivate*'} in return
      GPrivate* gobj() { return gobject_; }
                                ^~~~~~~~

This backport fixes from
https://github.com/GNOME/glibmm/commit/37d57ae9572b7d74aa385a30313eceae7f2d3fce
plus changes to generated files, as found in
https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=85e6e0e04ae19670b071820947d941e521d1200e
or
https://src.fedoraproject.org/rpms/glibmm24/c/ecc4276486706c0ebc55aff3ac592499dc9df2a9?branch=master

This includes a workaround to fix ZSTD detection.

This allows to test the minimal setup of the configuration.

The setup is split, as slaves are now not required.

Some test methods are changed to reflect the reality of instantiation.