The new code had a successful build on the CI machine. The difference in the build time is less than I would have thought it would be. It looks like the tests are taking about 2 minutes less than they used to.[1]
I guess that’s something, but I’m not sure it’s different enough for a general time test. We could probably use the code in the test case on the JIRA[2] and test that selectRange takes less than 50-100ms or so. If a full recompose is triggered, it would take a full second or two. [1]http://apacheflexbuild.cloudapp.net:8080/job/flex-tlf/buildTimeTrend [2]https://issues.apache.org/jira/browse/FLEX-34769 On Dec 10, 2015, at 1:26 AM, Justin Mclean <justinmcl...@me.com> wrote: > Hi, > >> Anyone have thoughts on performance checking to ensure that future >> performance degradation is caught? > > Usually I just add a couple of unit tests that assert the time run < some > reasonable value. No idea how easy that is to do in this case. > > Thanks, > Justin
