Develop release #123
Closed
Develop release #123
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
It makes pylint happy to run on test directories. It makes it possible to put code that both plugins and tests use in a directory that they can access. Signed-off-by: mulhern <amulhern@redhat.com>
Make a separate target, for convenience. Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
For hypothesis and for vim. Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Change pytest required version to 2.3. Signed-off-by: mulhern <amulhern@redhat.com>
It is not reliably raised. Signed-off-by: mulhern <amulhern@redhat.com>
Updated rhbz#1255191 with this info. Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
This reverts commit 5a86fa2. Better not to skip these tests.
Signed-off-by: mulhern <amulhern@redhat.com>
Hypothesis is not supported for 3.2. See: https://github.com/DRMacIver/hypothesis/iss Signed-off-by: mulhern <amulhern@redhat.com>
Remove remaining bits of udev_database plugin. Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
This gives fewer distracting skip messages, because the skipped tests get filtered out during input selection by hypothesis. This results in fewer tests being assembled and fewer skipped. Runtime is more variable, because of hypothesis database. Should expect it to be consistently longer on Travis CI by a factor as large as 2. Some of the @given() annotations could be usefully abstracted once it is certain that they are correct. Signed-off-by: mulhern <amulhern@redhat.com>
There was too much indentation in the main test file.
Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Increase the number of examples in case there are other interesting exceptions. Signed-off-by: mulhern <amulhern@redhat.com>
Some links that udevadm reports are not actually links to their device node, but instead to related device nodes in a multipath group. Signed-off-by: mulhern <amulhern@redhat.com>
It is more reasonable to do this, since this appears to be an unfixed regression. Signed-off-by: mulhern <amulhern@redhat.com>
Read the version information from a special version file. Signed-off-by: mulhern <amulhern@redhat.com>
See justification here: http://blog.ionelmc.ro/2014/05/25/python-packaging/.
This serves as a record of what we claim does not require a checker. It cuts down the number of distracting SKIP messages.
Filter blacklisted methods in strategy in order to avoid distracting SKIP messages. Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
See rhbz#1263351. Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
…hesis. Signed-off-by: mulhern <amulhern@redhat.com>
Should have gone in with commit 9e2fe0a. Signed-off-by: mulhern <amulhern@redhat.com>
Previously, ran test on a selection of operators that hypothesis sampled. Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
min_satisfying_examples is a soft bound; it will not cause a failure if the list sampled from is shorter than the bound. Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Change the exceptions raised slightly: Raise DeviceNotFoundByNumberError on bad subsystem value. More robust, because libudev seems to handle it, and ValueErrors should really just be used by Python libs. Add a DeviceNotFoundByFile error and use in Devices.from_device_file(). Signed-off-by: mulhern <amulhern@redhat.com>
Verify that pathlike sysnames always fail to be looked up properly, as well as that simple names always get looked up properly. Signed-off-by: mulhern <amulhern@redhat.com>
These tests will fail, because of rhbz#1263441. Signed-off-by: mulhern <amulhern@redhat.com>
Should have gone in with commit ca9b27d. Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
This seems like the proper place, since both methods assume that the origin of the identifying string is known. Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Closes: #101. Signed-off-by: mulhern <amulhern@redhat.com>
Closes: #104 Resolves: rhbz#1267584 A general requirement of a mapping is that there is a value for every key, but this is not really the case for attributes. This removes some methods and changes the semantics of some other methods. * __contains__ used to return True only if there was a value corresponding to the key. Now, it returns True if the key is what libudev calls available. It no longer reads a value for the key, which is probably closer to the correct semantics. * __getitem__ used to raise a KeyError for any key that had no value, regardless of whether the key was an available attribute or not. Now it raises a KeyError only when the attribute is not among the available attributes, and returns None when the attribute is among the available attributes but has no value. This is fairer than raising a KeyError even if the key is available, which is a bit counter-intuitive. * asstring() used to assume that the attribute always had some value, now it also handles the None case. It no longer possesses the following concrete methods of Mapping: __eq__ __ne__ get() items() iteritems() iterkeys() itervalues() keys() values() The only behavior that would have given consistancy with a Mapping interface would be to only return keys that have a corresponding value. It is, however, not recommended to read attributes without knowing exactly what they correspond to, as reading an attribute can have an effect. Second commit, replaces first idea. Stop treating Attributes as any sort of mapping from keys to values. It turns out that the libudev methods used to get the keys and those that look up the values have a very uncertain relationship to each other, and the best that you can say is that the intersection between the set of keys and the strings for which libudev will succesfully look up a value is almost certainly non-empty. This change forces getting rid of the __getitem__ method in the Attributes class, because if it is still around then the 'in' operator will pass int values to it. Change the method to one called 'lookup' to emphasize a little more that this is not simply an access to readily available data using a known useful key. Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Get rid of traversal entirely. Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
Signed-off-by: mulhern <amulhern@redhat.com>
…side. There will probably be more at some time, so the better abstracted the better. Do the same for the deprecated QUDevMonitorObservers, for clarity. Signed-off-by: mulhern <amulhern@redhat.com>
* Do not make DeviceNotFoundError a subtype of LookupError. * Use abstract super classes. * Add a new exception for a ValueError. Signed-off-by: mulhern <amulhern@redhat.com>
|
Supersed by develop-release-0.18 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.