##// END OF EJS Templates
threading: Add factory that creates curl session for each thread....
threading: Add factory that creates curl session for each thread. When creating a repo with an import url pointing to another repo on the same enterprise instance we call the vcssserver to check the url. The vcsserver then calls to enterprise to verify the url. This leads to two threads using the same cur session.

File last commit:

r1:854a839a default
r243:dae685be default
Show More
spec-by-example.rst
75 lines | 1.9 KiB | text/x-rst | RstLexer
project: added all source files and assets
r1
.. _test-spec-by-example:
==========================
Specification by Example
==========================
.. Avoid duplicating the quickstart instructions by importing the README
file.
.. include:: ../../../acceptance_tests/README.rst
Choices of technology and tools
===============================
`nix` as runtime environment
----------------------------
We settled to use the `nix` tools to provide us the needed environment for
running the tests.
`Gherkins` as specification language
------------------------------------
To specify by example, we settled on Gherkins as the semi-formal specification
language.
`py.test` as a runner
---------------------
After experimenting with `behave` and `py.test` our choice was `pytest-bdd`
because it allows us to use our existing knowledge about `py.test` and avoids
that we have to learn another tool.
Concepts
========
The logic is structured around the design pattern of "page objects". The
documentation of `python-selemium` contains a few more details about this
pattern.
Page Objects
------------
We introduce an abstraction class for every page which we have to interact with
in order to validate the specifications.
The implementation for the page objects is inside of the module
:mod:`page_objects`. The class :class:`page_objects.base.BasePage` should be
used as a base for all page object implementations.
Locators
--------
The specific information how to locate an element inside of the DOM tree of a
page is kept in a separate class. This class serves mainly as a data container,
it shall not contain any logic.
The reason for keeping the locators separate is that we expect a frequent need
for change whenever we work on our templates. In such a case it is more
efficient to have all locators together and update them there instead of having
to find all locators inside of the logic of a page object.