[Webkit-unassigned] [Bug 184340] Perf dashboard should send a notification when a test group finishes
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Wed Jun 6 21:12:06 PDT 2018
https://bugs.webkit.org/show_bug.cgi?id=184340
Ryosuke Niwa <rniwa at webkit.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Attachment #342101|review? |review+
Flags| |
--- Comment #14 from Ryosuke Niwa <rniwa at webkit.org> ---
Comment on attachment 342101
--> https://bugs.webkit.org/attachment.cgi?id=342101
Patch
View in context: https://bugs.webkit.org/attachment.cgi?id=342101&action=review
> Websites/perf.webkit.org/public/api/test-groups.php:30
> + AND testgroup_needs_notification IS TRUE and testgroup_hidden IS FALSE");
Nit: and -> AND.
> Websites/perf.webkit.org/public/v3/models/test-group.js:43
> + notificationSentAt() { return this._notificationSentAt; }
This should probably return new Date if _notificationSentAt is set.
> Websites/perf.webkit.org/public/v3/models/test-group.js:61
> + await AnalysisTask.fetchById(this._taskId);
We should just return here.
> Websites/perf.webkit.org/public/v3/pages/chart-pane.js:123
> + createWithTestGroupCheckbox.onchange = () => repetitionCount.disabled = notifyOnCompletion.disabled = !createWithTestGroupCheckbox.checked;
Nit: Each assignment statement needs to be on its own line per our style guideline.
https://webkit.org/code-style-guidelines/#linebreaking-multiple-statements
> Websites/perf.webkit.org/server-tests/privileged-api-create-test-group-tests.js:1253
> + it('should be able to create a test group with needs-notification flag not set', async () => {
Nit: flag unset*
> Websites/perf.webkit.org/tools/js/analysis-results-notifier.js:18
> + this._messageConstructionRules = messageConstructionRules;
Why don't we add a check here to make sure each rule has either tests or platforms filter set,
and make sure they're also arrays of strings.
> Websites/perf.webkit.org/tools/js/analysis-results-notifier.js:80
> + if (rule.tests) {
> + if (!rule.tests.includes(test))
A simpler way to write this might be:
if (rule.tests && !rule.tests.includes(test))
return false;
if (rule.platforms && !rule.platforms.includes(platform))
return false;
return true;
and then do a check to make sure each rule has either tests or platforms set
as an early error check in the constructor.
> Websites/perf.webkit.org/tools/js/test-group-result-page.js:21
> + static _URLForAnalysisTask(analysisTask)
Nit: I think we usually use lowercase url.
> Websites/perf.webkit.org/tools/js/test-group-result-page.js:26
> + _buildCommitSetToResultsAMapAndBarWidthCalculator(testGroup, analysisResultsView)
I think it's better to simply call this _buildCommitSetToResultsMap.
> Websites/perf.webkit.org/tools/js/test-group-result-page.js:47
> + return [resultsByCommitSet, (value) => (value - minValue) / (maxValue - minValue) * 100];
Let's return a dictionary instead; e.g. {resultsByCommitSet, barWidthFromValue: (value) => ~}
> Websites/perf.webkit.org/tools/js/test-group-result-page.js:60
> + const formatter = metric.makeFormatter(4);
Can we extract the content of this loop as a helper function?
> Websites/perf.webkit.org/tools/js/test-group-result-page.js:80
> + for (const result of results) {
Can we extract this as a helper function as well?
> Websites/perf.webkit.org/tools/js/test-group-result-page.js:193
> + _constructBarGraph(width)
We should put this below render() function so that it reads naturally top-down.
> Websites/perf.webkit.org/tools/js/test-group-result-page.js:195
> + const barGraphPlaceholder = this.createElement('div',{class: 'bar-graph-placeholder'});
It's not necessary to create div each time. Just put it into the template and just set width instead.
> Websites/perf.webkit.org/tools/js/test-group-result-page.js:207
> + static get pageContent()
This is not right. Use contentTemplate.
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20180607/3e670f0f/attachment-0001.html>
More information about the webkit-unassigned
mailing list