diff options
| author | Matúš Kukan <matus.kukan@collabora.com> | 2014-10-29 11:51:49 +0100 |
|---|---|---|
| committer | Matúš Kukan <matus.kukan@collabora.com> | 2014-10-29 17:41:54 +0100 |
| commit | e4e7f9d88e05fa610a72245c40f4e47f85db61ff (patch) | |
| tree | d28891fdc5bb13c94b51e315e15db319a44b0a8a /test/Library_test.mk | |
| parent | need to initialize mnRefCnt (diff) | |
| download | core-e4e7f9d88e05fa610a72245c40f4e47f85db61ff.tar.gz core-e4e7f9d88e05fa610a72245c40f4e47f85db61ff.zip | |
Improve framework for callgrind performance unit tests
Do not abort in tests which take longer then before.
Since 4f5f6d2444a24138c3d3d378771f87cb06427195 we have nice tool to
collect callgrind results, store them..
Also, make it not necessary to duplicate anything. Any test can be turned
into a performance test by using callgrindStart/callgrindDump. You only
need to duplicate the makefile name in gb_Module_add_perfcheck_targets.
So, if the test is run as part of 'perfcheck', it runs under callgrind,
otherwise it behaves like a regular unit test.
Change-Id: I7acbb855b1823c9a32ba126abb57dccc767ca239
Diffstat (limited to 'test/Library_test.mk')
| -rw-r--r-- | test/Library_test.mk | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/test/Library_test.mk b/test/Library_test.mk index 2c8d854417a2..df7bfd210a0a 100644 --- a/test/Library_test.mk +++ b/test/Library_test.mk @@ -19,6 +19,7 @@ $(eval $(call gb_Library_use_externals,test,\ boost_headers \ cppunit \ libxml2 \ + valgrind \ )) $(eval $(call gb_Library_use_libraries,test,\ @@ -40,6 +41,7 @@ $(eval $(call gb_Library_use_libraries,test,\ $(eval $(call gb_Library_add_exception_objects,test,\ test/source/bootstrapfixture \ test/source/diff/diff \ + test/source/callgrind \ test/source/xmltesttools \ test/source/htmltesttools \ test/source/mtfxmldump \ |