Report per operation timings and drop the unused performance framework - #4230
Merged
vogella merged 1 commit intoAug 9, 2026
Merged
Conversation
Contributor
OpenCloseEditorTest measured opening and closing in one block and EditorSwitchTest measured twelve switches in another, so a change in a single operation could not be attributed. Both tests now time the individual operations and print min, median, 90th percentile and maximum via the new UIPerformanceTestUtil.reportTimings. Drop org.eclipse.test.performance from both tests. The performance database it reports to has not been configured for years, which made commitMeasurements and assertPerformance no-ops and left the tests unable to fail or to report anything. Its one measurement per block model is also what made the individual operations indistinguishable. Without it each operation is timed directly, and both tests assert that the editor actually opened and that the expected editor is active after a switch. Both tests warm up before measuring so that class loading and JIT no longer end up in the reported times, and both are bounded by time instead of a fixed iteration count. EditorSwitchTest drops from about 145 to 16 seconds while collecting more than twice the samples.
vogella
force-pushed
the
editor-performance-measurements
branch
from
August 8, 2026 04:39
5cfa04d to
db8a67e
Compare
vogella
changed the title
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The editor performance tests measured opening plus closing in a single block, and twelve editor switches in another, so a change in one operation could not be attributed, and nothing was printed at all. Both tests now time the individual operations and log min, median, 90th percentile and maximum, they warm up before measuring so that class loading and JIT no longer end up in the reported times, and they are bounded by time instead of a fixed iteration count.
Both tests also stop using org.eclipse.test.performance. The performance database it reports to has not been configured for years, so commitMeasurements and assertPerformance were no-ops and the tests could neither fail nor report anything, and its one value per measurement block model is what made the individual operations indistinguishable in the first place. In its place both tests now assert that the editor really opened and that the expected editor is active after a switch. EditorSwitchTest drops from roughly 145 to 16 seconds while collecting more than twice the samples. The bundle still requires org.eclipse.test.performance for the other tests, so the manifest is unchanged.
This is a first step towards measuring and then improving editor handling. Sample output of a local run: