- 12 6月, 2017 1 次提交
-
-
由 Cleber Rosa 提交于
New optional plugins are now available, and some in RPM package format. Let's include them where plugin RPM packages are listed. Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
- 08 6月, 2017 6 次提交
-
-
由 Amador Pahim 提交于
Signed-off-by: NAmador Pahim <apahim@redhat.com>
-
由 Lukáš Doktor 提交于
Currently we just pickle the varianter object and hope we'll be able to load it with other Avocado version. It usually works, but is not really scalable. This patch adds a way to serialize Varianter into a list of dicts and serializes it using json, turning the unsafe objects to strings. Then it adds a way to load Varianter status using this serialized state. This procedure is lossy as it always assumes we are in non-debug mode and it replaces the variants with safe representations. As a benefit one does not need the original plugins to be able to reuse/replay the variants as Avocado loads the representation using fake classes. Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
由 Lukáš Doktor 提交于
Currently we dump args using pickle, which tries to pickle all objects. This could fail between versions or even between instances (eg. with different set of plugins). Let's just use safe json serialization replacing unsafe objects with None to avoid unnecessary crashes while still being able to make the safe information available. Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
由 Lukáš Doktor 提交于
The variants file is called multiplex even though we changed the terminology long ago. Let's support both names for now and slowly move to the correct name. Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
由 Lukáš Doktor 提交于
We had already split the mux into mutliple files. Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
由 Amador Pahim 提交于
Signed-off-by: NAmador Pahim <apahim@redhat.com>
-
- 07 6月, 2017 5 次提交
-
-
由 Lukáš Doktor 提交于
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com> * https://github.com/avocado-framework/avocado: Added support for POWER9 cpu arch identification
-
由 Lukáš Doktor 提交于
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com> * https://github.com/avocado-framework/avocado: loader: recursive test discovery loader: add '__init__.py' to path when path is a dir
-
由 Amador Pahim 提交于
This patch adds to the loader the ability to recursively discover tests when the docstring `
🥑 recursive` is specified in the test class. This recursive discovery will walk all the way up in the recursion chain, adding the parents test methods to the child class information. Reference: https://trello.com/c/GtFlhcOjSigned-off-by: NAmador Pahim <apahim@redhat.com> -
由 Amador Pahim 提交于
In `_find_avocado_tests()`, `path` is a python source code file which will be processed by `ast.parse()`. If it's a directory, `ast.parse()` will fall in exception. This patch makes the `path` to point to `path/__init__.py` when `path` is a directory. Signed-off-by: NAmador Pahim <apahim@redhat.com>
-
由 Lukáš Doktor 提交于
Currently when the target file does not exists the FileLoader reports: "Does not look like an INSTRUMENTED test, nor is it executable" which can be a bit misleading. Let's change it to: "File not found ($locations)". Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
- 06 6月, 2017 2 次提交
-
-
由 Satheesh Rajendran 提交于
Added support for POWER9 cpu arch identification Signed-off-by: NSatheesh Rajendran <sathnaga@linux.vnet.ibm.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
- 01 6月, 2017 1 次提交
-
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
- 31 5月, 2017 4 次提交
-
-
由 Lukáš Doktor 提交于
We rely on garbage collector to run __del__s before we check all tmpdirs were cleaned. Let's force-run gc.collect() and (just to be sure) run sync. Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
由 Lukáš Doktor 提交于
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com> * https://github.com/avocado-framework/avocado: VM Runner: use default on option parser for port and username VM Runner: do not keep parser as attribute Remote Runner: use default on option parser for port and username Remote Runner: do not keep parser as attribute Remote and VM runner: remove configured attribute assignment
-
由 Lukáš Doktor 提交于
The suite order means the order in which the variants will be applied to tests. Right now we run all variants of first test, then the same for the second, third, ... tests. The new option "--execution-order" allows to alternatively specify "tests-per-variant" mode where we run first variant of all tests, then second variant, third, ... Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
- 30 5月, 2017 11 次提交
-
-
由 Lukáš Doktor 提交于
We need optional plugins in ReadTheDocs, but they are not installed by default. Let's use `make develop` there as it should do and install normally when outside ReadTheDocs environment. Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
由 Lukáš Doktor 提交于
Currently the yaml_to_mux plugin is a core plugin without properly defined requirements. Let's separate it to optional_plugins, define dependencies and also build it separately in RPM. Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
由 Lukáš Doktor 提交于
With more and more optional plugins it could be useful to build their API documentation in order to be able to refer to it from the main documentation. Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
由 Lukáš Doktor 提交于
Let's report full result on doc build failure to be able to see what is wrong. Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Instead of current, let's stick to the default of using the word default, and also the option value from optparse itself. Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
由 Cleber Rosa 提交于
Signed-off-by: NCleber Rosa <crosa@redhat.com>
-
- 29 5月, 2017 6 次提交
-
-
由 Amador Pahim 提交于
This patch adds the missing definition of `MODE_0644` in `LoaderTestFunctional` class. Signed-off-by: NAmador Pahim <apahim@redhat.com>
-
由 Lukáš Doktor 提交于
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com> * https://github.com/avocado-framework/avocado: selftests: remove redundant test test: always execute tearDown()
-
由 Lukáš Doktor 提交于
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com> * https://github.com/avocado-framework/avocado: avocado list command: also show tag statistics avocado list command: also list test tags
-
由 Amador Pahim 提交于
Signed-off-by: NAmador Pahim <apahim@redhat.com>
-
由 Amador Pahim 提交于
Signed-off-by: NAmador Pahim <apahim@redhat.com>
-
由 Amador Pahim 提交于
Signed-off-by: NAmador Pahim <apahim@redhat.com>
-
- 26 5月, 2017 4 次提交
-
-
由 Cleber Rosa 提交于
The format of the docstring directives has not been formally described so far. Because at least two different features are built on that, and possibly more will come, it's a good idea to define a format. This introduces a stricter format, basically adding a number of requirements or rules. To make the following description more consistent let's call the "
🥑 " part of the docstring directive "the marker" and what follows it the "content". 1) One or more spaces are required between the "🥑 " marker and the content. The following is a valid docstring directive: "🥑 enable" While this is invalid: ":avocado:enable" 2) A more limited set of characters are now allowed in the content, namely one from the following set: [a-zA-Z0-9_:,=]. The following is a valid docstring directive: "🥑 foo=bar,foz=baz:extra" While the following is invalid: "🥑 foo=bar!" 3) The first character of the content must be alphanumeric, so the following is a valid docstring directive: "🥑 foo=bar" While the following is invalid: "🥑 =bar" 4) An end of string (or end of line) is now required after the content (which itself may not contain white spaces). The following is a valid docstring directive: "🥑 enable" While the following is invalid: "🥑 enable FOO" Signed-off-by: NCleber Rosa <crosa@redhat.com> -
由 Lukáš Doktor 提交于
When the test does not produces the early_status in 60s or when the test pushes the final status and not exits we SIGKILL the test, leaving us with no trace on where it might hang. Let's use SIGTERM first and wait for 1s to see whether the process won't be able to give us traceback, otherwise nuke it with SIGKILL as before. Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
由 Lukáš Doktor 提交于
The SystemExit exception does not produce the traceback, let's use RuntimeError instead to be able to tell where the test was interrupted. Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>
-
由 Abdul Haleem 提交于
Added a function to get the linux page size for the given kernel, it returns a byte value from 'getconf PAGESIZE' command. systems can have 4k, 64k page sizes, accordingly decide to start a memory tests. Signed-off-by: NAbdul Haleem <abdhalee@linux.vnet.ibm.com>
-