Calligra/autobuild: Difference between revisions
| Line 40: | Line 40: | ||
Usage: make benchmark | Usage: make benchmark | ||
Output: only on stdout, no files | Output: only on stdout, no files | ||
For running benchmarks, see: | |||
http://labs.qt.nokia.com/2009/11/05/analyzing-benchmark-results-with-bm/ | |||
== CSTester == | == CSTester == | ||
Revision as of 09:19, 8 February 2011
Autobuild
The goal of the autobuilder is to build Calligra continuously and run tests after Calligra has built. There are four kinds of tests currently:
- unittests (written using the QTest framework)
- benchmarks (also written using the QTest framework)
- roundtrip testing of the kofficetests document repository using massTester (TO BE MIGRATED)
- rendering testing of the kofficetests repository using CSTester (TODO)
Nightly packages for Linux can be delivered using the OpenSUSE build service (TODO)
Builds are run on:
cruncher1.kogmbh.net
a machine donated by KO GmbH. This machine is in certain cases also available to Calligra developers who want to test the behaviour of code on a machine with 12 cores (6 real with hyperthreading)
Results are available on:
http://my.cdash.org/index.php?project=Calligra
CDash can eat files like:
http://cyrille.diwi.org/tmp/calligra/Test.xml
Unittests
Usage: make test Outputdirectory: ${BUILD_DIR}/Testing/Temporary/ Files:
- CTestCheckpoint.txt
- CTestCostData.txt
- LastTest.log: full log including debug output
- LastTestsFailed.log: list of tests that failed
Benchmarks
Usage: make benchmark Output: only on stdout, no files
For running benchmarks, see:
http://labs.qt.nokia.com/2009/11/05/analyzing-benchmark-results-with-bm/