And the upgrade from one to another

This fixes a regression from merge request !729 , surykatka command was
no longer installed

- not all SR was using Cython 0.28.2
- with an upcoming upgrade of pycocotools, matplotlib>=2.1.0 will be required

in some webrunner, there are 2 keys in equeue.db:

```
>>> print(db.keys())
['/srv/slapgrid/slappart18/bin//runner-importer', '/srv/slapgrid/slappart18/bin/runner-importer']
```

As you can see, the importer script path changed. But we don't care
about this, we always want to have the latest (= the newest) backup date.

See merge request !734

in some webrunner, there are 2 keys in equeue.db:

```
>>> print(db.keys())
['/srv/slapgrid/slappart18/bin//runner-importer', '/srv/slapgrid/slappart18/bin/runner-importer']
```

As you can see, the importer script path changed. But we don't care
about this, we always want to have the latest (= the newest) backup date.

Because of a bug in `slapos/testcase`, ldd check was not properly reporting several problems, because several components were using system libraries.

These are the changes to profiles, adding the missing rpath and other changes so that components do not use system libraries either by explicitly providing the libraries so that the configure script does not use system development libraries if they are available ( like for zstd ) or by disabling the features at compile time for features we don't use ( like the `--without-icu` for boost )

See merge request !732

Fix this random failure we sometimes have, like [here](https://nexedijs.erp5.net/#/test_result_module/20200414-44DF7D23/11) by adding a promise so that test does not try to connect to a nginx that's not started yet

```
======================================================================
ERROR: test (test.TestJSTestNode)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "parts/slapos-repository/software/jstestnode/test/test.py", line 60, in test
    '%sjio/test/tests.html' % (connection_dict['nginx'], ), allow_redirects=False)
  File "eggs/requests-2.20.0-py2.7.egg/requests/api.py", line 75, in get
    return request('get', url, params=params, **kwargs)
  File "eggs/requests-2.20.0-py2.7.egg/requests/api.py", line 60, in request
    return session.request(method=method, url=url, **kwargs)
  File "eggs/requests-2.20.0-py2.7.egg/requests/sessions.py", line 524, in request
    resp = self.send(prep, **send_kwargs)
  File "eggs/requests-2.20.0-py2.7.egg/requests/sessions.py", line 637, in send
    r = adapter.send(request, **kwargs)
  File "eggs/requests-2.20.0-py2.7.egg/requests/adapters.py", line 516, in send
    raise ConnectionError(e, request=request)
ConnectionError: HTTPConnectionPool(host='::1', port=9443): Max retries exceeded with url: /jio/test/tests.html (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7f9b01f4b310>: Failed to establish a new connection: [Errno 111] Connection refused',))

```

See merge request !733

Several software use `slapos` command and expect this command to have networkcache support enabled, which is achieved in buildout by having `slapos.libnetworkcache` in the egg "working set" when generating the slapos script.

Introduce a dedicated section to install `slapos` command and use it in softwares which includes a slapos, testnode and slaprunner. Test softwares like slapos-sr-testing also need a slapos command, but they generate one with development version of slapos.core.

Before these changes, several software were installing `slapos` indirectly by installing `extra-eggs` from monitoring stack. The monitoring stack was also modified so that it uses a section with a prefix (`monitor-eggs`) and users of monitoring stacks were also modified to not use this `extra-eggs` part from monitoring stack, but install eggs through their own section.

Similar problem also existed with `slapos.toolbox`, which is installed by several softwares and stacks, for different reasons (because it contains lots of utilities), now that buildout detects parts installing the same path, only one parts can install the scripts. For this, define a section `slapos-toolbox` that software and stack should install, instead of listing the egg as `eggs` in a `zc.recipe.egg` section, because that would install the scripts again.

See merge request !729

When we have a section to generate a python interpreter with eggs built
in, don't name this `extra-eggs` because the meaning is not clear.

This section installs slapos.toolbox eggs with dependencies and install
scripts

Despite the name, this section was inherited from monitor, but now that
we fixed monitor to prefix the sections, this section is no longer
extended from monitor, but defined in softwares

Also build the binaries as static so that they don't use system
libraries.

To prevent using system libraries

Otherwise it will try to use development system package and we will have
different results depending on wether that package is installed or not.

Note that because of a typo in the environment variable (this makefile
uses $LDFLAGS, slapos zlib was not effective so far)

slap_connection is deprecated

This prevent naming conflicts in softwares extending and is cleaner.

Also other cleanups:
 - install an interpreter with a prefix, for similar reason
 - list monitor scripts explicitly, so that we don't have collisions
   with other sections installing same eggs (eg. slapos.core)
 - use ${buildout:bin-directory} instead of guessing it

This section installs a slapos command with network cache enabled, for
softwares using an embedded slapos.

Only install required eggs, drop unused [testnode] section