This reverts commit 85e6e268.

This reverts commit 38cc7f9c.

Add a test where the binary is compatible but the signature is unverified

Avoid catching interruptions and system calls during exception handling

Get all binaries for the requested SR
Indicate which are verified by the certificate list

Leave string interpolation to the logging method

Use `gcc -dumpmachine` instead of `platform.machine()` to discriminate
architecture

Use libnetworkcache to list only binaries with verified signature

See merge request nexedi/slapos.core!352

See merge request nexedi/slapos.core!346

See merge request !347

Improvements:

 * allow to destroy any kind of instance
 * damage software release to avoid change propagation
 * provide more information in the UI for the OSS Operator
 * stabilise field and action naming
 * allow to rename and destroy only stopped instances

For the OSS Operator it's important to have a safe way to rename and stop
instances, especially in context of the Instance Tree management.

This action is more advertised then rename and destroy, which has much
more drastic consequences.

Fix a missing dependency in 1.7.0 on python2

Minor version was increased because "slapgrid: Process promises with
instance python" brings full support of using slapos node on python3
also when using software on python2

Support a syntax like:

      slapos request \
        --node=computer_guid=local \
        --parameters-file=~/request.json \
        ERP5 \
        https://lab.nexedi.com/nexedi/slapos/raw/1.0.145/software/erp5/software.cfg

to request an instance with parameters in ~/request.json file.

This also teach slapos.core about the instance schema, which makes it possible to
understand the serialisation, so that we no longer need to use `_ = json.dump(params)`
to encode parameters.

This also emits a warning when requesting with parameters which do not match the schema.


See merge request nexedi/slapos.core!215

Because json is unicode in python2, pretty-printing the parameters would
display something like {u'foo': u'bar'}. Encode these unicode strings on
python2, so that it looks like {'foo': 'bar'} on both python2 and python3

This is more readable