Signed-off-by: NCleber Rosa <crosa@redhat.com>

Signed-off-by: NCleber Rosa <crosa@redhat.com>

The fabric-2.0.0 is a complete rework of fabric using different
approach and will require many changes to our code. Let's keep using the
1.x.y version for now.

The fabric3 is still only based on 1.x.y branch but let's also limit
that one as it is based on the original fabric so it's expected that
they might eventually come-up with 2.x.y py3 fork.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

* https://github.com/avocado-framework/avocado:
  get_slot_from_sysfs: accomodate more type of slot names in pci utils

Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

* https://github.com/avocado-framework/avocado:
  utils.process: Ignore decode-errors when logging the output
  utils.process: Simplify GDBSubProcess docstring
  utils.process: Use the `astring.ENCODING` by default

It might happen that user runs process with encoding that is ritcher
than system one. This is fine by the process as internally it stores
encoded bytes object, but the drainer might fail attempting to log the
output.

Those failures should not be critical as, under described circumstances,
the unavailability of the output is expected and should not prevent the
execution. Let's simply replace those characters using 'replace' (the
'xmlcharrefreplace' might fail with ASCII coder) and keep processing the
data.

To simplify error handling new "astring.to_text" argument "errors" is
added.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Most of the arguments in GDBSubProcess are ignored, let's list all with
a simple generic note to be quickly recognized.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Recently we introduced the "preferred" encoding, which should be better
default than just `sys.getdefaultencoding`. Let's use it in the process
module.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Signed-off-by: NCleber Rosa <crosa@redhat.com>

Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

These short exception detail variables are just one of the many
"invalid-name" violations we have.  Better have some fixed than
none.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Signed-off-by: NCleber Rosa <crosa@redhat.com>

And not "asset_file", because that references the file with the
content itself, and not the file containing its hash.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Which helps us to better document, also in the avocado.Test class
itself the hash algorithm that will be used by default.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Instead of raising a generic exception about not being able to fetch
the file, let's tell the user upfront that the protocol is not
supported.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

By using a common interface, and choosing the function to use based on
the protocol (aka scheme) some duplicate code can be removed.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

This improves readability, removing on indentation level and bringing
the exception raised when no writable cache dir exists, (much) closer
to the code that checks for them.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The asset fetchers cache directories are system wide, and not test
specific.  They're available at the test only for convenience.

They relate much more to the the other data directories, and deserve
to be visible by users alongside the other data directories.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The author probably meant "search" and not "find".
Signed-off-by: NCleber Rosa <crosa@redhat.com>

It adds no real value on top of hashlib.new(), only imposing virtual
limitations that do not map the real list of available hash
algorithms.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The parameter for the hash algorithm is "algorithm", and not "method".
Signed-off-by: NCleber Rosa <crosa@redhat.com>

This patche takes care of Slot format for ibm and openpower machines
* S0001
* Slot1
* SLOT1
* Backplane USB
* U78CB.001.WZS07CU-P1-C9-T1 (should return U78CB.001.WZS07CU-P1-C9)
* Onboard USB

Respective unit test has also been added test_utils_pci.py
Signed-off-by: NAbdul Haleem <abdhalee@linux.vnet.ibm.com>
Signed-off-by: NNarasimhan V <sim@linux.vnet.ibm.com>

Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

* https://github.com/avocado-framework/avocado:
  Update documentation to reflect relocated examples.
  Update avocado.conf to reflect relocated example tests.
  SPEC file: Relocate examples to doc directory to prevent inclusion of python version-specific byte-compiled files in package.

Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

* https://github.com/avocado-framework/avocado:
  Add methods to get/set cpuidle states

Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

* https://github.com/avocado-framework/avocado:
  selftest: Fix line-wrap

Add methods to get/set cpuidle states
in host and selftests for the method.
Signed-off-by: NSatheesh Rajendran <sathnaga@linux.vnet.ibm.com>

Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

* https://github.com/avocado-framework/avocado:
  Added method to get NUMA nodes with memory

This patch adds method to get list of nodes with memory on given
machine.
Signed-off-by: NHarish <harish@linux.vnet.ibm.com>

Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

* https://github.com/avocado-framework/avocado:
  Add comma_separated_ranges_to_list utility

Signed-off-by: NCleber Rosa <crosa@redhat.com>

Fedora 28 changed the URL of the Cloud Images, from "CloudImages" to
simply "Cloud".  Let's adapt the name depending on the version
requested.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Patch adds comma_separated_ranges_to_list utility to return a list
of integers from comma separated range values
Signed-off-by: NHarish <harish@linux.vnet.ibm.com>

Signed-off-by: NCleber Rosa <crosa@redhat.com>

Signed-off-by: NMerlin Mathesius <mmathesi@redhat.com>

Signed-off-by: NMerlin Mathesius <mmathesi@redhat.com>

On some architectures, Python will return a long on
data_structures.DataSize conversions.  That makes the identity
check assertion fail, because:

   >>> long(0) == int(0)
   True

But:
   >>> long(0) is int(0)
   False

Reference: https://kojipkgs.fedoraproject.org//work/tasks/1223/26881223/build.logSigned-off-by: NCleber Rosa <crosa@redhat.com>

python version-specific byte-compiled files in package.
Signed-off-by: NMerlin Mathesius <mmathesi@redhat.com>

Signed-off-by: NCleber Rosa <crosa@redhat.com>

Signed-off-by: NCleber Rosa <crosa@redhat.com>