this is not really related and we are considering checking software
directly in slapos node software.

this is used in a type comment

Show exactly what type is used for a request, it makes debugging easier.

By default proxy checks if software type is None, then it's setting it to
RootSoftwareInstance, but if it comes as "" (empty string) it's kept as is.

So for consistency with proxy set default software type, so that proxy can
manage it correctly.

Respect semantics of "verbose" vs "debug", to make it possible to run tests in
debug mode but without verbose messages.

Also adjust "Starting", this message was not clear

In SlapOS software release tests we often have to start a process or store
temporary files in a directory for the lifetime of the test. Cleaning up the
resources is a bit error prone and source of code duplication.

This introduce a registry of resources that are automatically created and freed
by the test framework, to simplify tests.

scan installed eggs and check if they have known vulnerabilities listed on
https://github.com/pyupio/safety-db

In a scenario of an instance with child instances which does not properly
support destruction, a scenario like this happens:

 1. first report detect that it's destroying a root partition but child
partitions are still started and request destruction of these child instances
 2. child partitions are destroyed
 3. root partition is destroyed

For this case we need to run 3 times

This also makes development easier, because we can make a temporary
change to skip setupModule not to reinstall software when we are working
only on the test and we don't need to recompile the software.

This should make investigation of errors on testnode easiers, any logged
messages by test framework goes to testcase.log

Setup logging at the very beginning, to also have messages logged during
initialization

Also adjust a bit the logged messages, we don't need to log everytime we
snapshot a file or folder, this cause the log to be very big, but it can
be interesting to log when we start and stop checking software, so that
we see how long this took.

Once testnodes are updated and always set SLAPOS_TEST_LOG_DIRECTORY, we
no longer need to have a default case of creating a snapshot folder
automatically, which was causing difficulties when running tests locally
as we had to remove this folder manually between test executions.

When this test fail we had incorrect traceback

We observe that software sometimes fail due to transiant errors such as
DNS/network errors, so increase the default number of retries to build
software from 2 to 10.

Sometimes also instance is failing for monitoring or other errors that
might become OK if we retry more, so increase the default number of
retries for instance from 10 to 20.

This reverts commit 3160c589.

pushed by mistake. This is unfinished but it was wrong because we copy
this log file to snapshot directory before we are done writing to it.

buildout profiles and .installed.cfg, shared parts signatures

This reverts commit 8a3c8aa3 because we
still have complaints for binaries that are linked, directly or not, to
libgfortan (e.g. openblas). Adding zlib to the rpath of these binaries
would be annoying.

See slapos@6db91afc

the rpath we set in LDFLAGS is not honored here, but we can ignore this
as rpath is needed on executables and not important on shared objects.

libgfortran.so is a shared object that happens to be executable, so it
appears in our ldd check, but we are not executing it directly so it
should be OK.

if previous test fail to build, compile dir is there for inspection. We
don't want to check binaries in there.

This supports the case where some paths used for shared parts are
symlinks

Empty paths confuse the ldd check, because empty prefix always look
valid, so executables using system libraries will be considered as
valid.

This should prevent such confusing tracebacks when building software fail:

    INFO Finished software releases.
    ERROR:slapos.testing.testcase:Error during request destruction
    Traceback (most recent call last):
      File "parts/slapos.core-repository/slapos/testing/testcase.py", line 615, in _cleanup
	cls.requestDefaultInstance(state='destroyed')
      File "parts/slapos.core-repository/slapos/testing/testcase.py", line 680, in requestDefaultInstance
	cls._instance_parameter_dict)
    AttributeError: type object 'SlapOSInstanceTestCase for ...' has no attribute '_instance_parameter_dict'

This reverts commit 49ec3ece.

That commit revealed that all the rpath check was not effective and that
we have in fact lots of components which use system libraries.

We'll fix rpath on slapos profiles and only then we'll fix this check,
to prevent test failures (even if the test failures show a real problem)

This supports the case where some paths used for shared parts are
symlinks

Slapos is started the first time the StandaloneSlapOS is instanciated,
then, in each class teardown we stop slapos, so the second time
setUpModule is executed, the slapos is not started yet.

Start it one more time in setUpModule to ensure it's started in this
scenario.

When running with shared parts from slaprunner, we'll have path with
double slashes like:

    /srv/slapgrid/slappart3/srv//runner//shared/libexpat/468993ea3777c1a342102bc1c4b43425/lib/libexpat.so.1

because slaprunner has some paths with double slash.

Our condition was only checking executables, which missed some non
executable libraries.

Following up on nexedi/slapos.core!172 (comment 96033)

/reviewed-on nexedi/slapos.core!180

During teardown, we sometimes have errors that processes are still
running after being stopped, something like:
    
    Fault: <Fault 91: 'STILL_RUNNING: TestTelegraf-0'>

This is because we only retried `slapos node report` two times, but
`slapos node report` tells supervisor to stop processes without
waiting and when processes takes time to stop, this is sometimes not
enough. Increase to retry up to 10 times by default.

/reviewed-on nexedi/slapos.core!178

Except a short whitelist of system libraries, we should have always set
rpath so that the slapos version of the library is used.

We had issues with components which where compiled with slapos version
of openssl but linked without rpath. At runtime they were using system
openssl.

Rewrite our complex grep shell pipeline in python, with a function
parsing ldd output so that we can check not only for not found dynamic
libraries, but also for libraries that are found outside of slapos
library dirs.

This also fixes problems that this grep pipeline did not support paths
with spaces in their names.

Fix long lines, incorrect type annotations and adjust a bit to make mypy
happy.