Skip to content

Test runner GUI for Google-Test instrumented test applications

License

Notifications You must be signed in to change notification settings

tomzox/gtest_gui

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

GtestGui

GtestGui is a test-runner with graphical user-interface for test applications using the GoogleTest framework.

GtestGui was created mainly for removing the need for fiddling with Gtest command line options, for tracking the status of long-running test campaigns and for extracting individual test cases' traces from large trace output files.

The GoogleTest framework is widely known for its C++ Unit-Testing support via the GMock feature. But one can also use just the test framework for defining test cases, scheduling test executions, and collecting verdicts. Even using only the framework saves some implementation effort, but more importantly also has the advantage of presenting a well-known common interface to developers.

GtestGui is designed especially for component tests, integration tests and other higher-level testing, where the number of tests is smaller compared to unit-testing, but individual test cases run longer and tests may have to be run repeatedly (e.g. when test scenarios are not fully reproducible due to timing between threads.) But it can also be used for unit-tests with "flakiness" (as the gtest-parallel command line tool calls it) which show sporadic failures and thus require high repetition count for proving stability.

GtestGui supports this by allowing to schedule test cases concurrently across multiple CPUs. For very long-running campaigns, the "clean traces of passed tests" option allows storing only traces of failed tests to disk. Status of ongoing tests is presented in form of a top-level pass/fail summary and progress bar, as well as a running log of test case verdicts, which can be filtered and sorted (e.g. to show only failed test cases.) Traces of individual test cases are accessible while the test campaign is still running either in a small preview window below the test log, or via double-click which exports them to an external trace browser. (Trowser is recommended for this purpose.)

GtestGui provides access to Gtest command line options directly from the main window. Most importantly, test case filters can be entered manually, or via drop-down menu listing all test cases, or the test case list dialog. Other supported options are the test repetition count, failure limit, shuffling execution order, running disabled tests and breaking on failure or exceptions. The latter will on POSIX systems produce a core dump that can be pre-analyzed via the result log's context menu by extracting stack traces. Finally, an option allows running tests under valgrind and automatically detecting issues that were found for the result log.

GtestGui can also be used for analyzing pre-existing trace output files by adding the trace file names on the command line when starting.

screenshot of main window

GtestGui has a mature feature set that can support C++ development in all development phases: During initial code construction with test-driven design, it supports running the same sub-set of test cases over and over again. After code completion, it supports getting a quick overview of stability across the complete set of test cases and which kind of failures occur. And it supports final quality checking in long-running, high-repetition test campaigns, optionally using valgrind or sanitizer builds.

Installation

GtestGui is available as package "mote-gtest-gui" ("Module tester's Gtest GUI") on PyPi. It can be installed using pip3:

   pip3 install mote-gtest-gui

Make sure that the path where pip3 installs scripts is in your search path (e.g. on UNIX $HOME/.local/bin). Then you can run GtestGui simply by executing gtest_gui.

Alternatively, you can run GtestGui without installation by cloning the repository and then running gtest_gui.py in the top-level directory. For integration with the Trowser trace browser either place trowser.py somewhere in the path, or create a symbolic link to the path where it is located as gtest_gui/trowser, or specify the full path to the executable in configuration.

Usage

You can either specify the path of your test application executable file directly on the command line, or select one via a file browser afterward (see the Control menu). Alternatively, or additionally, you can specify one or more trace output files on the command line, which will be parsed for test case results and loaded into the result list in the GUI.

In summary, most common usage is as follows:

    gtest_gui my_test_application.exe

On the MS-Windows platform, start GtestGui via pythonw.exe to avoid getting a console window.

Documentation

For a detailed description, please refer to the manual page