REGRESSION(r212853): Comparisons to baseline no longer shows up

[WebKit-https.git] / Websites / perf.webkit.org / ChangeLog

2017-02-24  Ryosuke Niwa  <rniwa@webkit.org>

        REGRESSION(r212853): Comparisons to baseline no longer shows up
        https://bugs.webkit.org/show_bug.cgi?id=168863

        Reviewed by Joseph Pecoraro.

        The bug was caused by ChartStatusView's code not being updated to use TimeSeriesView's.
        Updated the code to use TimeSeriesView's methods to fix the bug.

        Also made InteractiveTimeSeriesChart's currentPoint to return a (TimeSeriesView, point, isLocked) tuple
        to consolidate it with lockedIndicator() to work towards making the baseline data points selectable.

        * browser-tests/time-series-chart-tests.js: Updated the test cases to use currentIndicator, and added
        test cases for newly added lastPointInTimeRange.

        * public/v3/components/chart-pane.js:
        (ChartPane.prototype.serializeState): Updated to use currentIndicator.
        (ChartPane.prototype._renderFilteringPopover): Ditto.

        * public/v3/components/chart-status-view.js:
        (ChartStatusView.prototype.updateStatusIfNeeded): Use currentIndicator for an interative time series.
        Fixed the non-interactive chart's code path for TimeSeriesView.
        (ChartStatusView.prototype._computeChartStatus): Modernized the code.
        (ChartStatusView.prototype._findLastPointPriorToTime): Deleted. Replaced by TimeSeriesView's
        lastPointInTimeRange.

        * public/v3/components/interactive-time-series-chart.js:
        (InteractiveTimeSeriesChart.prototype.currentIndicator):
        (InteractiveTimeSeriesChart.prototype.moveLockedIndicatorWithNotification):
        (InteractiveTimeSeriesChart.prototype._renderChartContent):
        (InteractiveTimeSeriesChart):

        * public/v3/models/time-series.js:
        (TimeSeriesView.prototype.lastPointInTimeRange): Added.
        (TimeSeriesView.prototype._reverse): Added. Traverses the view in the reverse order.
        * unit-tests/time-series-tests.js:

2017-02-23  Dewei Zhu  <dewei_zhu@apple.com>

        Rename 'commit_parent' in 'commits' table to 'commit_previous_commit'.
        https://bugs.webkit.org/show_bug.cgi?id=168816

        Reviewed by Ryosuke Niwa.

        Rename 'commit_parent' to avoid ambiguity in the coming feature.
        For exisiting database, run
            "ALTER TABLE commits RENAME commit_parent TO commit_previous_commit;"
        to update the database.

        * init-database.sql:
        * public/api/report-commits.php:
        * public/include/commit-log-fetcher.php:
        * server-tests/api-commits.js:
        (then):
        * server-tests/api-report-commits-tests.js:
        (then):
        * tools/sync-commits.py:
        (main):
        (Repository.fetch_commits_and_submit):
        (GitRepository._revision_from_tokens):
        * unit-tests/analysis-task-tests.js:
        (sampleAnalysisTask):

2017-02-23  Ryosuke Niwa  <rniwa@webkit.org>

        New sampling algorithm shows very few points when zoomed out
        https://bugs.webkit.org/show_bug.cgi?id=168813

        Reviewed by Saam Barati.

        When a chart is zoomed out to a large time interval, the new sampling algorithm introduced in r212853 can
        hide most of the data points because the difference between the preceding point's time and the succeeding
        point's time of most points will be below the threshold we computed.

        Instead, rank each data point based on the aforementioned time interval difference, and pick the first M data
        points when M data points are to be shown.

        This makes the new algorithm behave like our old algorithm while keeping it stable still. Note that this
        algorithm still biases data points without a close neighboring point but this seems to work out in practice
        because such a point tends to be an important sample anyway, and we don't have a lot of space between
        data points since we aim to show about one point per pixel.

        * browser-tests/index.html:
        (CanvasTest.canvasContainsColor): Extracted from one of the test cases and generalized. Returns true when
        the specified region of the canvas contains a specified color (alpha is optional).
        * browser-tests/time-series-chart-tests.js: Added a test case for sampling. It checks that sampling happens
        and that we always show some data point even when zoomed out to a large time interval.
        (createChartWithSampleCluster):

        * public/v3/components/interactive-time-series-chart.js:
        (InteractiveTimeSeriesChart.prototype._sampleTimeSeries):
        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart.prototype._ensureSampledTimeSeries): M, the number of data points we pick must be computed
        based on the width of data points we're about to draw constrained by the canvas size. e.g. when the canvas
        is only half filled, we shouldn't be showing two points per pixel in the filled region.
        (TimeSeriesChart.prototype._sampleTimeSeries): Refined the algorithm. First, compute the time difference or
        the rank for each N data points. Sort those ranks in descending order (in the order we prefer), and include
        all data points above the M-th rank in the sample.
        (TimeSeriesChart.prototype.computeTimeGrid): Revert the inadvertent change in r212935.

        * public/v3/models/time-series.js:
        (TimeSeriesView.prototype.filter): Fixed a bug that the indices passed onto the callback were shifted by the
        starting index.
        * unit-tests/time-series-tests.js: Added a test case to ensure callbacks are called with correct data points
        and indices.

2017-02-23  Ryosuke Niwa  <rniwa@webkit.org>

        REGRESSION(r212542): Make TimeSeriesChart.computeTimeGrid stops x-axis grid prematurely
        https://bugs.webkit.org/show_bug.cgi?id=168812

        Reviewed by Joseph Pecoraro.

        Add time iterator of two months, three months, and four months with some tests.

        Also for one-month time iterator, round the day of month to 1 or 15 whichever is closer.

        * browser-tests/time-series-chart-tests.js: Added more tests.
        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart._timeIterators.next):
        (TimeSeriesChart._timeIterators):

2017-02-22  Ryosuke Niwa  <rniwa@webkit.org>

        Add tests for InteractiveTimeSeriesChart and adopt actions
        https://bugs.webkit.org/show_bug.cgi?id=168750

        Reviewed by Chris Dumez.

        Added tests for InteractiveTimeSeriesChart.

        Also replaced selection.onchange, selection.onzoom, indicator.onchange, annotations.onclick callbacks
        by "selectionChange", "zoom", "indicatorChange", and "annotationClick" actions respectively.

        Also fixed various bugs and bad code I encountered while writing these tests.

        * browser-tests/index.html:
        (waitForComponentsToRender): Delay the call to enqueueToRender until the next run loop because there
        might be outstanding promises that just got resolved. e.g. for fetching measurement sets JSONs. Let
        all those promises get resolved first. Otherwise, some tests become racy.
        (canvasImageData): Extracted from time-series-chart-tests.js.
        (canvasRefTest): Ditto.
        (CanvasTest): Ditto.
        (CanvasTest.fillCanvasBeforeRedrawCheck): Ditto.
        (CanvasTest.hasCanvasBeenRedrawn): Ditto.
        (CanvasTest.canvasImageData): Ditto.
        (CanvasTest.expectCanvasesMatch): Ditto.
        (CanvasTest.expectCanvasesMismatch): Ditto.

        * browser-tests/time-series-chart-tests.js: Fixed some test cases where dpr multipler was not doing
        the right thing anymore in Safari under a high DPI screen. Also added a lot of test cases for interactive
        time series chart and one for rendering annotations.
        (scripts): Moved.
        (posixTime): Added. A helper function for sampleCluster.
        (dayInMilliseconds): Ditto.
        (sampleCluster): Moved here. Made the same cluster more artifical for an easier testing.
        (createChartWithSampleCluster): Moved out of one of the tests.
        (respondWithSampleCluster): Ditto.

        * public/v3/components/chart-pane-base.js:
        (ChartPaneBase.prototype.configure): Adopted new actions in InteractiveTimeSeriesChart.

        * public/v3/components/chart-status-view.js:
        (ChartStatusView.prototype.updateStatusIfNeeded): Removed a superflous console.log.

        * public/v3/components/chart-styles.js:
        (ChartStyles.mainChartOptions): Set zoomButton to true. InteractiveTimeSeriesChart used to determine
        whether to show the zoom button or not based on the precense of the zoom callback. We made it explicit.

        * public/v3/components/interactive-time-series-chart.js:
        (InteractiveTimeSeriesChart.prototype.setIndicator): Explicitly call _notifySelectionChanged with false
        instead of relying on undefined to be treated as falsey.
        (InteractiveTimeSeriesChart.prototype._createCanvas): Use id instead of selector to find elements.
        (InteractiveTimeSeriesChart.htmlTemplate):
        (InteractiveTimeSeriesChart.cssTemplate):
        (InteractiveTimeSeriesChart.prototype._mouseMove): Explicitly call _startOrContinueDragging with false
        instead of relying on undefined treated as falsey. Also added the missing call to enqueueToRender found
        by new tests. This was working fine on the dashboard due to other components invoking enqueueToRender
        but won't work in a standalone instance of InteractiveTimeSeriesChart.
        (InteractiveTimeSeriesChart.prototype._mouseLeave): Ditto, adding the missing call to enqueueToRender.
        (InteractiveTimeSeriesChart.prototype._click): Removed the assignment to _forceRender when calling
        _mouseMove in an early exist, which does set this flag and invokes enqueueToRender, and added the missing
        call to enqueueToRender in the other code path.
        (InteractiveTimeSeriesChart.prototype._startOrContinueDragging): Replaced annotations.onclick callback
        by the newly added "annotationClick" action, and added the missing call to enqueueToRender.
        (InteractiveTimeSeriesChart.prototype._endDragging): Use arrow function.
        (InteractiveTimeSeriesChart.prototype._notifyIndicatorChanged): Replaced indicator.onchange callback by
        the newly added "indicatorChange" action.
        (InteractiveTimeSeriesChart.prototype._notifySelectionChanged): Replaced selection.onchange callback by
        the newly added "selectionChange" action.
        (InteractiveTimeSeriesChart.prototype._renderChartContent): Show the zoom button when options.zoomButton
        is set instead of relying on the presence of selection.onzoom especially now that the callback has been
        replaced by the "zoom" action.

        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart.prototype.setAnnotations): Added the missing call to enqueueToRender.

        * public/v3/main.js:

2017-02-21  Ryosuke Niwa  <rniwa@webkit.org>

        Make sampling algorithm more stable and introduce an abstraction for sampled data
        https://bugs.webkit.org/show_bug.cgi?id=168693

        Reviewed by Chris Dumez.

        Before this patch, TimeSeriesChart's resampling resulted in some points poping up and disappearing as
        the width of a chart is changed. e.g. when resizing the browser window. The bug was by caused by
        the sample for a given width not always including all points for a smaller width so as the width is
        expanded, some point may be dropped.

        Fixed this by using a much simpler algorithm of always picking a point when the time interval between
        the preceding point and the succeeding point is larger than the minimum space we allow for a given width.

        Also introduced a new abstraction around the sample data: TimeSeriesView. A TimeSeriesView provides
        a similar API to TimeSeries for a subset of the time series filtered by a time range a custom function.
        This paves a way to adding the ability to select baseline, etc... on the chart status view.

        TimeSeriesView can be in two modes:
        Mode 1. The view represents a contiguous subrange of TimeSeries - In this mode, this._data references
                the underlying TimeSeries's _data directly, and we use _startingIndex to adjust index given to
                find the relative index. Finding the next point or the previous point of a given point is done
                via looking up the point's seriesIndex and doing a simple arithmetic. In general, an index is
                converted to the absolute index in the underlying TimeSeries's _data array.

        Mode 2. The view represents a filtered non-contiguous subset of TimeSeries -  In this mode, this._data is
                its own array. Finding the next point or the previous point of a given point requires finding
                a sibling point in the underlying TimeSeries which is in this view. Since this may result in O(n)
                traversal and a hash lookup, we lazily build a map of each point to its position in _data instead.

        * public/v3/components/chart-status-view.js:
        (ChartStatusView.prototype.updateStatusIfNeeded): Call selectedPoints instead of sampledDataBetween for
        clarity. This function now returns a TimeSeriesView instead of a raw array.

        * public/v3/components/interactive-time-series-chart.js:
        (InteractiveTimeSeriesChart.prototype.currentPoint): Updated now that _sampledTimeSeriesData contains
        an array of TimeSeriesView's. Note that diff is either 0, -1, or 1.
        (InteractiveTimeSeriesChart.prototype.selectedPoints): Ditto. sampledDataBetween no longer exists since
        we can simply call viewTimeRange on TimeSeriesView returned by sampledDataBetween.
        (InteractiveTimeSeriesChart.prototype.firstSelectedPoint): Ditto.
        (InteractiveTimeSeriesChart.prototype._sampleTimeSeries): Use add since excludedPoints is now a Set.

        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart.prototype.sampledDataBetween): Deleted.
        (TimeSeriesChart.prototype.firstSampledPointBetweenTime): Deleted.
        (TimeSeriesChart.prototype._ensureSampledTimeSeries): Modernized the code. Use the the time interval of
        the chart divided by the number of allowed points as the time interval used in the new sampling algorithm.
        (TimeSeriesChart.prototype._sampleTimeSeries): Rewritten. We also create TimeSeriesView here.
        (TimeSeriesChart.prototype._sampleTimeSeries.findMedian): Deleted.
        (TimeSeriesChart.prototype._updateCanvasSizeIfClientSizeChanged): Fixed a bug that the canvas size wasn't
        set to the correct value on Chrome when a high DPI screen is used.

        * public/v3/models/time-series.js:
        (TimeSeries.prototype.viewBetweenPoints): Renamed from dataBetweenPoints. Now returns a TimeSeriesView.
        (TimeSeriesView): Added. This constructor is to be called by viewBetweenPoints, viewTimeRange, and filter.
        (TimeSeriesView.prototype._buildPointIndexMap): Added. Used in mode (2).
        (TimeSeriesView.prototype.length): Added.
        (TimeSeriesView.prototype.firstPoint): Added.
        (TimeSeriesView.prototype.lastPoint): Added.
        (TimeSeriesView.prototype.nextPoint): Added. Note index is always a position in this._data. In mode (1),
        this is the position of the point in the underlying TimeSeries' _data. In mode (2), this is the position
        of the point in this._data which is dictinct from the underlying TimeSeries' _data.
        (TimeSeriesView.prototype.previousPoint): Ditto.
        (TimeSeriesView.prototype.findPointByIndex): Added. Finds the point using the positional index from the
        beginning of this view. findPointByIndex(0) on one view may not be same as findPointByIndex(0) of another.
        (TimeSeriesView.prototype.findById): Added. This is O(n).
        (TimeSeriesView.prototype.values): Added. Returns the value of each point in this view.
        (TimeSeriesView.prototype.filter): Added. Creates a new view with a subset of data points the predicate
        function returned true.
        (TimeSeriesView.prototype.viewTimeRange): Added. Creates a new view with a subset of data points for the
        given time ragne. When the resultant view would include all points of this view, it simply returns itself
        as an optimization.
        (TimeSeriesView.prototype.firstPointInTimeRange): Added. Returns the first point in the view which lies
        within the specified time range.
        (TimeSeriesView.prototype.Symbol.iterator): Added. Iterates over each point in the view.

        * public/v3/pages/analysis-task-page.js:
        (AnalysisTaskChartPane.prototype.selectedPoints): Use selectedPoints in lieu of getting selection and then
        calling sampledDataBetween with that range.

        * public/v3/pages/summary-page.js:
        (SummaryPageConfigurationGroup.set _medianForTimeRange): Modernized.

        * unit-tests/time-series-tests.js: Added tests for TimeSeries and TimeSeriesView. Already caught bugs!
        (addPointsToSeries):

2017-02-17  Ryosuke Niwa  <rniwa@webkit.org>

        Add tests for the time series chart and fix bugs I found along the way
        https://bugs.webkit.org/show_bug.cgi?id=168499

        Reviewed by Antti Koivisto.

        Add basic tests for the time series chart.

        Replaced the "ondata" callback set in the options by "dataChange" action now that ComponentBase provides
        a facility for defining event-like actions.

        Also fixed bugs I encountered while writing these tests see below for descriptions.

        * browser-tests/editable-text-tests.js:
        (waitToRender): Moved to index.html
        * browser-tests/index.html:
        (waitToRender): Moved from editable-text-tests.js.
        (wait): Added.
        * browser-tests/time-series-chart-tests.js: Added.
        * public/v3/components/chart-pane-base.js:
        (ChartPaneBase.prototype.configure):
        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart): Removed the code to set display and position inline properties. This is now done inside
        cssTemplate with :host pseudo class.
        (TimeSeriesChart.prototype._ensureCanvas): Don't strech the canvas to 100% of width and height. This was
        causing a flush of contents where the canvas is momentarily streched by the browser and the script later
        updates with the content with the correct aspect ratio.
        (TimeSeriesChart.cssTemplate): Added :host rule to set display: block and position: relative.
        (TimeSeriesChart._updateAllCharts): Deleted.
        (TimeSeriesChart.prototype.render): Only run the code for axis when options.axis is defined. Also, avoid
        setting the fill style because we never fill for axis drawing.
        (TimeSeriesChart.prototype._computeHorizontalRenderingMetrics): Ditto. Fallback to sensible values when
        options.axis is not defined.
        (TimeSeriesChart.prototype._renderYAxis): Now computeValueGrid generates a sequence of {time, label}.
        (TimeSeriesChart.prototype._renderTimeSeries): Don't draw the shades for confidence intervals unless its
        fill style is defined. Otherwise, we'd end up drawing black shade and mask the actual data points.
        (TimeSeriesChart.prototype._ensureSampledTimeSeries): Dispatch newly added "dataChange" action instead of
        calling "ondata" callback in options dictionary.
        (TimeSeriesChart.computeTimeGrid): Modernized to use const/let. Also fixed the bug that we were emitting
        the date even when the entire time range fit within a 24-hour interval.
        (TimeSeriesChart.computeValueGrid): Rewritten to make MB/GB use a nice round number instead of 0.98GB.
        We were using a power of 10 to round up the stepping value but the value formatter used a power of 1024
        to divide byte measurements (e.g. for memory). Use formatter.divisor to find the right scaling factor for
        each kind.
        * public/v3/models/metric.js:
        (Metric.prototype.makeFormatter):
        (Metric.makeFormatter): Extracted from the one on the prototype so that tests don't need a metric object
        just to test TimeSeriesChart. Added the second argument which specifies the maximum absolute value of the
        range we're formatting. This is needed to use the same number of decimal points when the most significant
        digit of some value is smaller than that of the biggest one. For example, we were emitting 0.50GB instead
        of 0.5G along with 2.0GB. The "adjustment" reduces the number of significant figures in these cases.
        * public/v3/pages/dashboard-page.js:
        (DashboardPage.prototype._createChartForCell):

2017-02-16  Ryosuke Niwa  <rniwa@webkit.org>

        Use expect.js instead of expect in browser tests
        https://bugs.webkit.org/show_bug.cgi?id=168492

        Reviewed by Joseph Pecoraro.

        Use expect.js (https://github.com/Automattic/expect.js) instead of expect (https://github.com/mjackson/expect).

        * browser-tests/close-button-tests.js:
        * browser-tests/component-base-tests.js:
        * browser-tests/editable-text-tests.js:
        * browser-tests/index.html:

2017-02-16  Ryosuke Niwa  <rniwa@webkit.org>

        Modernize and fix measurement-set tests
        https://bugs.webkit.org/show_bug.cgi?id=168484

        Reviewed by Joseph Pecoraro.

        Modernized and fixed the tests in measurement-set-tests.js.

        1. Return a promise instead of manually calling done in then/catch hanlders.
        2. Use arrow function everywhere.
        3. Explicitly assert the number of calls to callbacks instead of asserting never reached.

        The test case labled "should return false when the range ends after the fetched cluster"
        was incorrectly asserting that hasFetchedRange returns false when the end time is after
        the primary cluster's end time. Test an interval before the primary cluster instead.

        Added a test case for hasFetchedRange returning true when the end time appears after
        the end of the primary cluster and fixed hasFetchedRange to that end. Since there are
        no data points after the primary cluster which is chronologically the last cluster,
        there is nothing to fetch beyond its end time.

        * public/v3/models/measurement-set.js:
        (MeasurementSet.prototype.hasFetchedRange): Fixed the bug that this function returned
        false when the end time was after the primary cluster's end by truncating the range by
        the end of the primary cluster.
        * unit-tests/measurement-set-tests.js:
        * unit-tests/resources/mock-remote-api.js:
        (assert.notReached.assert.notReached): Deleted. It's no longer used by any tests.

2017-02-15  Ryosuke Niwa  <rniwa@webkit.org>

        Update ReadMe.md and merge it with Install.md
        https://bugs.webkit.org/show_bug.cgi?id=168405

        Reviewed by Michael Catanzaro.

        Merged Install.md and ReadMe.md into one file.

        * Install.md: Removed.
        * ReadMe.md: Merged Install.md at the top and updated the rest of the content.

2017-01-24  Ryosuke Niwa  <rniwa@webkit.org>

        Modernize editable-text component and add tests
        https://bugs.webkit.org/show_bug.cgi?id=167398

        Reviewed by Yusuke Suzuki.

        Modernized EditableText component to use the action feature added in r210938.

        * browser-tests/editable-text-tests.js: Added. Added tests for EditableText component.
        (.waitToRender):
        * browser-tests/index.html:
        * public/v3/components/base.js:
        (ComponentBase.prototype.dispatchAction): Return the result from the callback.
        * public/v3/components/editable-text.js:
        (EditableText): Removed a bunch of instance variables that are no longer needed.
        (EditableText.prototype.didConstructShadowTree): Added. Add event listeners on the Edit/Save button and the host.
        (EditableText.prototype.editedText): Return the text field's value directly.
        (EditableText.prototype.text): Added.
        (EditableText.prototype.setText): Call enqueueToRender automatically instead of relying on the parent component
        to do so in _startedEditingCallback, which has been removed.
        (EditableText.prototype.render): Modernized the code.
        (EditableText.prototype._didClick): No longer prevents the default action manually since that's automatically done
        in createEventHandler. Handle the case where the update action is not handled.
        (EditableText.prototype._endEditingMode): Renamed from _didUpdate.
        (EditableText.htmlTemplate): Added ids on various elements in the shadow tree.
        (EditableText.cssTemplate): Updated the CSS selectors per above change.
        * public/v3/main.js:
        (main): Fixed a typo.
        * public/v3/pages/analysis-task-page.js:
        (AnalysisTaskPage): Use the action listener instead of manually setting callbacks.
        (AnalysisTaskPage.prototype._createTestGroupListItem): Ditto.
        (AnalysisTaskPage.prototype._didStartEditingTaskName): Deleted.

2017-01-20  Ryosuke Niwa  <rniwa@webkit.org>

        Build fix after r210783. Didn't mean to require custom elements API.

        * public/v3/main.js:
        (main):

2017-01-20  Ryosuke Niwa  <rniwa@webkit.org>

        Make sync-commits.py robust against missing Subversion authors and missing parent Git commits
        https://bugs.webkit.org/show_bug.cgi?id=167231

        Reviewed by Antti Koivisto.

        Fixed a bug that a subversion commit that's missing author name (anonymous commit) results in an out of bound
        exception, and a bug that syncing a git repository starts failing once there was a merge commit which pulled
        in a commit data earlier than that of the last reported commit.

        For the latter fix, added --max-ancestor-fetch-count to specify the number of maximum commits to look back.

        * tools/sync-commits.py:
        (main): Added --max-ancestor-fetch-count.
        (Repository.fetch_commits_and_submit): If submit_commits fails with FailedToFindParentCommit, fetch the parent
        commit's information until we've resolved them all.
        (Repository.fetch_next_commit): Renamed from fetch_commit.
        (SVNRepository.fetch_next_commit): Renamed from fetch_commit. Don't try to get the author name if it's missing
        due to an anonymous commit. It's important to never include the "author" field in the JSON submitted to
        a dashboard since it rejects when "author" field is not an array (e.g. null). 
        (GitRepository.fetch_next_commit): Renamed from fetch_commit.
        (GitRepository.fetch_commit): Added. Fetches the commit information for a given git hash. Used to retrieve
        missing parent commits.
        (GitRepository._revision_from_tokens): Extracted from fetch_commit.

        * tools/util.py:
        (submit_commits): Optionally takes status_to_accept to avoid throwing in the case of FailedToFindParentCommit
        and returns the response JSON.

2017-01-20  Ryosuke Niwa  <rniwa@webkit.org>

        REGRESSION(r198234): /api/commits/%revision% always fails
        https://bugs.webkit.org/show_bug.cgi?id=167235

        Reviewed by Antti Koivisto.

        The bug was caused by a typo in CommitLogFetcher::fetch_revision, which was calling commit_for_revision on
        $this->db instead of $this. This had been monkey-patched in the internal dashboard so it was working there.

        Also fixed a bug that /latest wasn't doing what it claimed to do, and a bug that /oldest /latest,
        and /last-reported would return a commit with all values set to null instead of an empty list.

        Finally, added server API tests for /api/commits.

        * public/api/commits.php:
        (main): Add a comment for APIs that only exist for v2 UI.

        * public/include/commit-log-fetcher.php:
        (CommitLogFetcher::fetch_latest): Fixed the bug that this function was returning the oldest commit, not the
        the latest commit as desired.
        (CommitLogFetcher::fetch_revision): Fixed the bug that this function would always encounter an exception
        because commit_for_revision is defined on $this, not $this->db.
        (CommitLogFetcher::format_single_commit): Return an empty list instead of an array with a single commit with
        all values set to null.

        * server-tests/api-commits.js: Added. Added tests for the JSON API at /api/commits.
        (.assertCommitIsSameAsOneSubmitted): Added. A helper function to compare a commit returned by /api/commits
        to one sent to /api/report-commits.

2017-01-19  Ryosuke Niwa  <rniwa@webkit.org>

        measurement-sets API can incorrectly order points with OS version without commit time
        https://bugs.webkit.org/show_bug.cgi?id=167227

        Reviewed by Chris Dumez.

        Ignore revision_order for the purpose of ordering data points in /api/measurement-sets.

        These orderings are used in some UI (e.g A/B testing) to order OS build numbers which do not have a timestamp
        associated with each "revision".

        The baseline measurements made in our internal dashboard were using these ordering numbers before ordering
        results with build time. Because those data points don't have an associated Webkit revisions, all data points
        were ordered first by macOS's revision_order, then build time. Because v3 UI completely ignores revision_order
        for the purpose of plotting data points, this resulted in some data points being plotted in a wrong order
        with some lines going backwards in time.

        This patch addresses this discrepancy by stop ordering data points with revision_order in the JSON API.

        * public/api/measurement-set.php:
        (MeasurementSetFetcher::execute_query): Fixed the bug.
        * server-tests/api-measurement-set-tests.js: Added a test.

2017-01-19  Ryosuke Niwa  <rniwa@webkit.org>

        Build fix after r210626. We need to clear Triggerable's static map in each iteration.

        * tools/sync-buildbot.js:
        (syncLoop):

2017-01-18  Ryosuke Niwa  <rniwa@webkit.org>

        Add a mechanism to dispatch and listen to an action
        https://bugs.webkit.org/show_bug.cgi?id=167191

        Reviewed by Antti Koivisto.

        Added the notion of an action to components. Like DOM events, it can be dispatched or listen to.

        Also added ComponentBase.prototype.part which finds a sub-component inside a component's shadow tree,
        and made ComponentBase.prototype.content take an id to find an element that matches it.

        * browser-tests/close-button-tests.js: Added. Tests for CloseButton.
        * browser-tests/component-base-tests.js: Added tests for ComponentBase's part(~), content(id), dispatchEvent.
        * browser-tests/index.html:
        * public/v3/components/base.js:
        (ComponentBase): Added this._actionCallbacks, which is a map of an action name to a callback to be invoked.
        (ComponentBase.prototype.content): Return an element of the given id if one is specified.
        (ComponentBase.prototype.part): Find a component whose element has the matching id.
        (ComponentBase.prototype.dispatchAction): Added.
        (ComponentBase.prototype.listenToAction): Added.
        (ComponentBase.prototype._ensureShadowTree): Call didConstructShadowTree.
        (ComponentBase.prototype.didConstructShadowTree): Added.
        (ComponentBase.prototype._recursivelyReplaceUnknownElementsByComponents): Copy attributes when instantiating
        an element for a component when the browser doesn't support custom elements API.
        (ComponentBase.createLink):
        (ComponentBase.prototype.createEventHandler): Added.
        (ComponentBase.createEventHandler): Renamed from createActionHandler.
        * public/v3/components/button-base.js:
        (ButtonBase.prototype.didConstructShadowTree): Added. Dispatch "activate" action when the button is clicked.
        (ButtonBase.prototype.setCallback): Deleted.
        (ButtonBase.htmlTemplate): Use id instead of class so that this.content() can find it.
        (ButtonBase.cssTemplate): Updated style rules.
        * public/v3/pages/chart-pane.js:
        (ChartPane):
        (ChartPane.prototype.didConstructShadowTree): Added. Listen to "activate" action on the close button.
        (ChartPane.prototype.render): Fixed a bug that we were never calling enqueueToRender on the close button.
        (ChartPane.htmlTemplate): Add the id on the close button.

2017-01-18  Dewei Zhu  <dewei_zhu@apple.com>

        'buildbot-syncer.js' should be able to determine force build argument from a list of possible repositories.
        https://bugs.webkit.org/show_bug.cgi?id=167152

        Reviewed by Ryosuke Niwa.

        Add 'rootOptions' key which maps to a list of possible repositories.
        For a build request, only one of the repositories in the list is valid.

        * tools/js/buildbot-syncer.js:
        (BuildbotSyncer.prototype._propertiesForBuildRequest):
        (BuildbotSyncer._validateAndMergeProperties):
        (BuildbotSyncer):
        * unit-tests/buildbot-syncer-tests.js:
        (sampleiOSConfig):
        (sampleiOSConfigWithExpansions):
        (createSampleBuildRequest):
        (Promise.resolve.then):
        * unit-tests/resources/mock-v3-models.js:
        (MockModels.inject):

2017-01-17  Ryosuke Niwa  <rniwa@webkit.org>

        Make calls to render() functions async
        https://bugs.webkit.org/show_bug.cgi?id=167151

        Reviewed by Andreas Kling.

        Make calls to render() async by coalescing calls inside enqueueToRender(), which has been renamed from
        updateRendering(). We now queue up all the components and wait until the next animation frame to invoke
        render() on all those components.

        This reduces render() calls in the summary page of our internal dashboard by 15x from ~9400 to ~600 by
        eliminating pathological O(n^2) behavior between render() calls.

        Consolidated TimeSeriesChart's enqueueRender into this newly added feature of ComponentBase along with
        the support to call render() on a resize event. New implementation makes use of connectedCallback and
        disconnectedCallback to avoid the work when the component is not in a document tree.

        The rest of the patch concerns with renaming updateRendering to enqueueToRender and fixing a few minor bugs
        that I encountered while working on this patch.

        * browser-tests/component-base-tests.js: Added tests for ComponentBase.enqueueToRender().
        * browser-tests/index.html:
        (BrowserContext.prototype.importScripts): Renamed from importScript; Now supports loading multiple scripts.
        (BrowserContext.prototype.importScript): Added.
        (BrowserContext): Removed the unused createWithScripts.

        * public/v3/components/analysis-results-viewer.js:
        (AnalysisResultsViewer.prototype._expandBetween):
        * public/v3/components/bar-graph-group.js:
        (BarGraphGroup.prototype.updateGroupRendering):

        * public/v3/components/base.js:
        (ComponentBase): When the browser doesn't support custom elements and 
        (ComponentBase.prototype.enqueueToRender): Renamed from updateRendering. Queues up the component to call
        render() instead of immediately invoking it.
        (ComponentBase.renderingTimerDidFire): Call render(). Since render() function often calls enqueueToRender
        on child components, go ahead and invoke render() on any components enqueued during render() calls.
        (ComponentBase._connectedComponentToRenderOnResize): Added.
        (ComponentBase._disconnectedComponentToRenderOnResize): Added.
        (ComponentBase.defineElement.elementClass.prototype.connectedCallback): Added. This is an optimization to
        avoid the work when the component is not in the document; e.g. because the entire page component has been
        detached from the document. The old implementation in TimeSeriesChart was not doing this.
        (ComponentBase.defineElement.elementClass.prototype.disconnectedCallback): Added.
        (ComponentBase): Replaced unused static variables with _componentsToRender and _componentsToRenderOnResize.

        * public/v3/components/chart-pane-base.js:
        (ChartPaneBase.prototype.fetchAnalysisTasks):
        (ChartPaneBase.prototype.didUpdateAnnotations): Added. Addresses the bug that the annotation bars in the
        charts shown an an analysis task doesn't update its color when the state is updated in the UI. 
        (ChartPaneBase.prototype._mainSelectionDidZoom):
        (ChartPaneBase.prototype._updateStatus):
        (ChartPaneBase.prototype._requestOpeningCommitViewer):
        (ChartPaneBase.prototype._keyup):
        (ChartPaneBase.prototype.render):
        * public/v3/components/commit-log-viewer.js:
        * public/v3/components/customizable-test-group-form.js:
        (CustomizableTestGroupForm):
        (CustomizableTestGroupForm.prototype._customize):
        * public/v3/components/editable-text.js:
        (EditableText.prototype._didUpdate):
        * public/v3/components/interactive-time-series-chart.js:
        * public/v3/components/pane-selector.js:
        (PaneSelector.prototype._selectedItem):
        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart): Removed the logic to update upon resize. See _connectedComponentToRenderOnResize above.
        (TimeSeriesChart.prototype.get enqueueToRenderOnResize): Added. Returns true.
        (TimeSeriesChart.prototype.enqueueToRender): Deleted.
        (TimeSeriesChart._renderEnqueuedCharts): Deleted.
        (TimeSeriesChart): Call ComponentBase.defineElement to make this a proper component so that the logic in
        connectedCallback to update upon resize event would work.
        * public/v3/instrumentation.js:
        (Instrumentation.dumpStatistics): Sort results by the key names.
        * public/v3/models/time-series.js:
        (TimeSeries.prototype.values): Added. This method was never ported to v3 in r198462, and broke the feature
        to show moving averages, etc... on the charts page.
        * public/v3/pages/analysis-category-page.js:
        (AnalysisCategoryPage.prototype.open):
        (AnalysisCategoryPage.prototype.updateFromSerializedState):
        (AnalysisCategoryPage.prototype.filterDidChange):
        (AnalysisCategoryPage.prototype.render):
        * public/v3/pages/analysis-task-page.js:
        (AnalysisTaskChartPane.prototype._updateStatus):
        (AnalysisTaskPage.prototype.updateFromSerializedState):
        (AnalysisTaskPage.prototype._didFetchTask):
        (AnalysisTaskPage.prototype._didFetchRelatedAnalysisTasks):
        (AnalysisTaskPage.prototype._didFetchMeasurement):
        (AnalysisTaskPage.prototype._didFetchTestGroups):
        (AnalysisTaskPage.prototype._showAllTestGroups):
        (AnalysisTaskPage.prototype._didFetchAnalysisResults):
        (AnalysisTaskPage.prototype.render):
        (AnalysisTaskPage.prototype._renderTestGroupList.):
        (AnalysisTaskPage.prototype._renderTestGroupList):
        (AnalysisTaskPage.prototype._createTestGroupListItem):
        (AnalysisTaskPage.prototype._showTestGroup):
        (AnalysisTaskPage.prototype._didStartEditingTaskName):
        (AnalysisTaskPage.prototype._updateTaskName):
        (AnalysisTaskPage.prototype._updateTestGroupName):
        (AnalysisTaskPage.prototype._hideCurrentTestGroup):
        (AnalysisTaskPage.prototype._updateChangeType): Fixed the bug that we were never updating annotation bars
        in the main chart by calling didUpdateAnnotations.
        (AnalysisTaskPage.prototype._associateBug):
        (AnalysisTaskPage.prototype._dissociateBug):
        (AnalysisTaskPage.prototype._associateCommit):
        (AnalysisTaskPage.prototype._dissociateCommit):
        (AnalysisTaskPage.prototype._chartSelectionDidChange):
        (AnalysisTaskPage.prototype._selectedRowInAnalysisResultsViewer):
        * public/v3/pages/build-request-queue-page.js:
        (BuildRequestQueuePage.prototype.open.):
        (BuildRequestQueuePage.prototype.open):
        * public/v3/pages/chart-pane.js:
        (ChartPane.prototype.setOpenRepository):
        (ChartPane.prototype._renderTrendLinePopover): Fixed a race condition. Insert a select element as needed
        before trying to assign the current value on it.
        (ChartPane.prototype._trendLineTypeDidChange):
        (ChartPane.prototype._updateTrendLine):
        * public/v3/pages/charts-page.js:
        (ChartsPage.prototype.updateFromSerializedState):
        (ChartsPage.prototype._updateDomainsFromSerializedState):
        (ChartsPage.prototype.setNumberOfDaysFromToolbar):
        (ChartsPage.prototype._didMutatePaneList):
        (ChartsPage.prototype.render):
        * public/v3/pages/charts-toolbar.js:
        (ChartsToolbar.prototype.render):
        * public/v3/pages/create-analysis-task-page.js:
        (CreateAnalysisTaskPage.prototype.updateFromSerializedState):
        * public/v3/pages/dashboard-page.js:
        (DashboardPage.prototype.updateFromSerializedState):
        (DashboardPage.prototype._fetchedData):
        * public/v3/pages/heading.js:
        (Heading.prototype.render):
        * public/v3/pages/page-with-heading.js:
        (PageWithHeading.prototype.render):
        * public/v3/pages/page.js:
        (Page.prototype.open):
        * public/v3/pages/summary-page.js:
        (SummaryPage.prototype.open):
        (SummaryPage.prototype.this._renderQueue.push):
        (SummaryPage):
        (SummaryPage.prototype._renderCell):

2017-01-15  Ryosuke Niwa  <rniwa@webkit.org>

        Add the build fix for browsers that don't yet support custom elements SPI.
        It was supposedly to be a part of the previous commit.

        * public/v3/components/base.js:
        (ComponentBase.defineElement):

2017-01-12  Ryosuke Niwa  <rniwa@webkit.org>

        Adopt custom elements API in perf dashboard
        https://bugs.webkit.org/show_bug.cgi?id=167045

        Reviewed by Darin Adler.

        Adopt custom elements API in ComponentBase, and create the shadow tree lazily in content() and render()
        instead of eagerly creating it inside the constructor.

        For now, create a separate element class for each component in ComponentBase.defineElement instead of
        making ComponentBase inherit from HTMLElement to preserve the semantics we have as well as to test
        the boundaries of what custom elements API allows for framework authors.

        In order to ensure one-to-one correspondence between elements and their components, we use a static map,
        ComponentBase._currentlyConstructedByInterface, to remember which element or component is being created
        and use that in custom element's constructor to update element.component() and this._element.

        Also dropped the support for not having attachShadow as we've shipped this feature in Safari 10.

        Finally, added tests to be ran inside a browser to test the front end code in browser-tests.

        * browser-tests/component-base-tests.js: Added. Basic tests for ComponentBase.
        * browser-tests/index.html: Added.

        * public/v3/components/base.js:
        (ComponentBase): Don't create the shadow tree. Use the currently constructed element as this._element if
        there is one (the custom element's constructor is getting called). Otherwise create a new element but
        store this component in the map to avoid creating a new component in the custom element's constructor.
        (ComponentBase.prototype.content): Lazily create the shadow tree now.
        (ComponentBase.prototype.render): Ditto.
        (ComponentBase.prototype._ensureShadowTree): Renamed from _constructShadowTree. Dropped the support for
        not having shadow DOM API. This is now required. Also use importNode instead of cloneNode in cloning
        the template content since the latter would not get upgraded.
        (ComponentBase.prototype._recursivelyReplaceUnknownElementsByComponents): Modernized the code. Don't
        re-create a component if its element had already been upgraded by its custom element constructor.
        (ComponentBase.defineElement): Add this component to the static maps. _componentByName is used by
        _recursivelyReplaceUnknownElementsByComponents to instantiate new components in the browsers that don't
        support custom elements API and _componentByClass is used by ComponentBase's constructor to lookup the
        element name. The latter should go away once all components fully adopt ComponentBase.defineElement.
        (ComponentBase.defineElement.elementClass): A class to define a custom element for the component.
        We need to reconfigure the property since class's name is not writable but configurable.

        * public/v3/components/button-base.js:
        (ButtonBase.htmlTemplate): Added. Extracted the common code from CloseButton and WarningIcon.
        (ButtonBase.buttonContent): Added. An abstract method overridden by CloseButton and WarningIcon.
        (ButtonBase.sizeFactor): Added. Overridden by WarningIcon.
        (ButtonBase.cssTemplate): Updated to use :host.
        * public/v3/components/close-button.js:
        (CloseButton.buttonContent): Renamed from htmlTemplate.
        * public/v3/components/spinner-icon.js:
        (SpinnerIcon.cssTemplate): Removed webkit prefixed properties, and updated it to animate stroke instead
        of opacity to reduce the power usage.
        (SpinnerIcon.htmlTemplate): Factored stroke, stroke-width, and stroke-linecap into cssTemplate.
        * public/v3/components/warning-icon.js:
        (WarningIcon.cssTemplate): Deleted.
        (WarningIcon.sizeFactor): Added.
        (WarningIcon.buttonContent): Renamed from htmlTemplate.
        * public/v3/pages/summary-page.js:
        (SummaryPage._constructRatioGraph): Fixed a bug that we were not never calling spinner.updateRendering().
        (SummaryPage.prototype._renderCell):

2017-01-13  Ryosuke Niwa  <rniwa@webkit.org>

        Instrument calls to render()
        https://bugs.webkit.org/show_bug.cgi?id=167037

        Reviewed by Sam Weinig.

        Wrap every call to render() by newly added ComponentBase.updateRendering() to instrument it.
        Also, use arrow functions instead of this.render.bind or regular closures for simplicity.

        Currently, we're making 5100 calls to render() while opening the summary page, and that's way too high.

        * public/v3/components/analysis-results-viewer.js:
        (AnalysisResultsViewer.prototype._expandBetween):
        * public/v3/components/bar-graph-group.js:
        (BarGraphGroup.prototype.updateGroupRendering): Renamed form render() as BarGraphGroup is not a component.
        * public/v3/components/base.js:
        (ComponentBase.prototype.updateRendering): Added. Instruments render() call.
        * public/v3/components/chart-pane-base.js:
        (ChartPaneBase.prototype.fetchAnalysisTasks):
        (ChartPaneBase.prototype._mainSelectionDidZoom):
        (ChartPaneBase.prototype._updateStatus):
        (ChartPaneBase.prototype._requestOpeningCommitViewer):
        (ChartPaneBase.prototype._keyup):
        (ChartPaneBase.prototype.render):
        * public/v3/components/commit-log-viewer.js:
        * public/v3/components/customizable-test-group-form.js:
        (CustomizableTestGroupForm):
        (CustomizableTestGroupForm.prototype._customize):
        * public/v3/components/editable-text.js:
        (EditableText.prototype._didUpdate):
        * public/v3/components/pane-selector.js:
        (PaneSelector.prototype._selectedItem):
        * public/v3/components/results-table.js:
        (ResultsTable.prototype.render):
        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart._renderEnqueuedCharts):
        * public/v3/pages/analysis-category-page.js:
        (AnalysisCategoryPage.prototype.open):
        (AnalysisCategoryPage.prototype.updateFromSerializedState):
        (AnalysisCategoryPage.prototype.filterDidChange):
        (AnalysisCategoryPage.prototype.render):
        * public/v3/pages/analysis-task-page.js:
        (AnalysisTaskChartPane.prototype._updateStatus):
        (AnalysisTaskPage.prototype.updateFromSerializedState):
        (AnalysisTaskPage.prototype._didFetchTask):
        (AnalysisTaskPage.prototype._didFetchRelatedAnalysisTasks):
        (AnalysisTaskPage.prototype._didFetchMeasurement):
        (AnalysisTaskPage.prototype._didFetchTestGroups):
        (AnalysisTaskPage.prototype._showAllTestGroups):
        (AnalysisTaskPage.prototype._didFetchAnalysisResults):
        (AnalysisTaskPage.prototype.render):
        (AnalysisTaskPage.prototype._renderTestGroupList.):
        (AnalysisTaskPage.prototype._renderTestGroupList):
        (AnalysisTaskPage.prototype._createTestGroupListItem):
        (AnalysisTaskPage.prototype._showTestGroup):
        (AnalysisTaskPage.prototype._didStartEditingTaskName):
        (AnalysisTaskPage.prototype._updateTaskName):
        (AnalysisTaskPage.prototype._updateTestGroupName):
        (AnalysisTaskPage.prototype._hideCurrentTestGroup):
        (AnalysisTaskPage.prototype._updateChangeType):
        (AnalysisTaskPage.prototype._associateBug):
        (AnalysisTaskPage.prototype._dissociateBug):
        (AnalysisTaskPage.prototype._associateCommit):
        (AnalysisTaskPage.prototype._dissociateCommit):
        (AnalysisTaskPage.prototype._chartSelectionDidChange):
        (AnalysisTaskPage.prototype._selectedRowInAnalysisResultsViewer):
        * public/v3/pages/build-request-queue-page.js:
        (BuildRequestQueuePage.prototype.open.):
        (BuildRequestQueuePage.prototype.open):
        * public/v3/pages/chart-pane.js:
        (ChartPane.prototype.setOpenRepository):
        (ChartPane.prototype._trendLineTypeDidChange):
        (ChartPane.prototype._updateTrendLine):
        * public/v3/pages/charts-page.js:
        (ChartsPage.prototype.updateFromSerializedState):
        (ChartsPage.prototype._updateDomainsFromSerializedState):
        (ChartsPage.prototype.setNumberOfDaysFromToolbar):
        (ChartsPage.prototype._didMutatePaneList):
        (ChartsPage.prototype.render):
        * public/v3/pages/charts-toolbar.js:
        (ChartsToolbar.prototype.render):
        * public/v3/pages/create-analysis-task-page.js:
        (CreateAnalysisTaskPage.prototype.updateFromSerializedState):
        * public/v3/pages/dashboard-page.js:
        (DashboardPage.prototype.updateFromSerializedState):
        (DashboardPage.prototype._fetchedData):
        * public/v3/pages/heading.js:
        (Heading.prototype.render):
        * public/v3/pages/page-with-heading.js:
        (PageWithHeading.prototype.render):
        * public/v3/pages/page.js:
        (Page.prototype.open):
        * public/v3/pages/summary-page.js:
        (SummaryPage.prototype.open):
        (SummaryPage.prototype.this._renderQueue.push):
        (SummaryPage):
        (SummaryPage.prototype._renderCell):

2017-01-12  Ryosuke Niwa  <rniwa@webkit.org>

        Outliers are not hidden in v3 UI
        https://bugs.webkit.org/show_bug.cgi?id=166966

        Reviewed by Andreas Kling.

        Fixed the typo in addToSeries. An outlier has markedOutlier set to true, not isOutlier.

        Also fixed a bug unveiled by new tests in MeasurementRootSet.ensureSingleton. It was was creating
        a new MeasurementRootSet each time it was called instead of finding an existing instance. Fixed the bug
        by merging the static maps of MeasurementRootSet and RootSet.

        * public/v3/models/measurement-cluster.js:
        (MeasurementCluster.prototype.addToSeries): Fixed the bug.
        * public/v3/models/root-set.js:
        (MeasurementRootSet.prototype.namedStaticMap): Added.
        (MeasurementRootSet.prototype.ensureNamedStaticMap): Added.
        (MeasurementRootSet.namedStaticMap): Added.
        (MeasurementRootSet.ensureNamedStaticMap): Added.
        * unit-tests/measurement-set-tests.js: Added tests for adopting time series data from a cluster.

2017-01-12  Ryosuke Niwa  <rniwa@webkit.org>

        Hide the UI to trigger an A/B testing when there are no triggerables
        https://bugs.webkit.org/show_bug.cgi?id=166964

        Reviewed by Yusuke Suzuki.

        Hide the "Start A/B Testing" button on analysis task pages instead of showing it and failing later
        when the user tries to create one it with a TriggerableNotFound error.

        Added the list of triggerables to the manifest JSON so that we can determine this condition without
        having to fetch /api/triggerable for each analysis task as done in v2 UI.

        * public/admin/reprocess-report.php:
        * public/api/manifest.php:
        * public/api/report.php:
        * public/include/admin-header.php:
        * public/include/manifest-generator.php: Moved from public/include/manifest.php.
        (ManifestGenerator::generate):
        (ManifestGenerator::triggerables): Added. Include the list of repositories this triggerable accepts
        as well as the list of (test, platform) pairs on which this triggerable is available.
        Use [testId, platformId] instead of a dictionary to reduce the file size.
        * public/v3/components/customizable-test-group-form.js:
        (CustomizableTestGroupForm): Removed this._disabled. This variable was used in TestGroupFrom to
        disable the "Start A/B Testing" button when no range is selected but this ended up racy. Compute
        the visibility of the button in render() function instead.
        (CustomizableTestGroupForm.prototype.setRootSetMap):
        (CustomizableTestGroupForm.prototype._submitted):
        (CustomizableTestGroupForm.prototype.render): Hide the customize link and the button as needed.
        The "Start A/B Testing" button must be hidden when either no range is selected or no title is typed.
        "Customize" button must be hidden when no range is selected.
        * public/v3/components/test-group-form.js:
        (TestGroupForm): Removed _disabled since it's no longer used.
        (TestGroupForm.prototype.setDisabled): Ditto.
        (TestGroupForm.prototype.render): Ditto.
        * public/v3/index.html: Include triggerable.js.
        * public/v3/models/manifest.js:
        (Manifest._didFetchManifest): Modernized. Create Triggerable objects from the manifest JSON.
        * public/v3/models/triggerable.js: Added.
        (Triggerable): Add this triggerable object to the static map of (test id, platform id) pair.
        (Triggerable.prototype.acceptedRepositories): Added.
        (Triggerable.findByTestConfiguration): Added. Finds a triggerable in the aforementioned static map.
        * public/v3/pages/analysis-task-page.js:
        (AnalysisTaskChartPane.prototype._updateStatus): Added. Re-render the page since time series data
        points that were previously not available may have become available. The lack of this update was
        causing a race condition in which the "Start A/B Testing" button for the charts is disabled even
        after a group name had been specified because setRootSetMap was never called with a valid set.
        (AnalysisTaskPage): Added this._triggerable.
        (AnalysisTaskPage.prototype._didFetchTask): Find the triggerable now that we've fetched the task.
        (AnalysisTaskPage.prototype.render): Hide the group view (the table of A/B testing results) entirely
        when there are no groups to show. Also hide the forms to start A/B testing when there are no matching
        triggerable, which is the main feature of this patch.
        * server-tests/api-manifest.js: Added a test for including a list of triggerables in the manifest JSON.
        * server-tests/resources/mock-data.js:
        (MockData.resetV3Models): Reset Triggerable's static map.
        * server-tests/tools-buildbot-triggerable-tests.js: Assert that Triggerable objects are  constructed
        with appropriate list of repositories and (test, platform) associations.
        * tools/js/database.js:
        (tableToPrefixMap): Added triggerable_repositories's prefix.
        * tools/js/remote.js:
        (RemoteAPI.prototype.getJSON): Log the entire response to stderr when JSON.parse fails to aid debugging.
        * tools/js/v3-models.js: Import triggerable.js.

2017-01-11  Ryosuke Niwa  <rniwa@webkit.org>

        fetch-from-remote doesn’t work with some websites
        https://bugs.webkit.org/show_bug.cgi?id=166963

        Reviewed by Yusuke Suzuki.

        Apparently file_get_contents is not compatible with some SSL/TLS connections.
        Use curl_* functions to access remote servers instead.

        * public/admin/fetch-from-remote.php:
        (fetch_remote):

2017-01-10  Ryosuke Niwa  <rniwa@webkit.org>

        Another build fix. Always use UTC when expressing commit times in UNIX-epoch timestamps.

        * public/api/measurement-set.php:

2017-01-10  Ryosuke Niwa  <rniwa@webkit.org>

        Fix a typo in the previous commit.

        * public/api/measurement-set.php:

2017-01-10  Ryosuke Niwa  <rniwa@webkit.org>

        Build fixes for older versions of Postgres.

        Also redirect / and /# to /v3/ as intended in r200820.

        * public/api/measurement-set.php:
        * public/api/runs.php:
        * public/index.html:

2016-10-18  Dewei Zhu  <dewei_zhu@apple.com>

        Update test cases for change r206465.
        https://bugs.webkit.org/show_bug.cgi?id=163618

        Reviewed by Ryosuke Niwa.

        Update test case for change r206465 which added support for multiple summary pages.
        Use deepStrictEqual instead of deepEqual as deepEqual will not complain in the case like 'deepEqual([],{})'.
        Fix a test failure in tools-buildbot-triggerable-tests.js.
        Fix a bug in generating manifest.

        * config.json:
        * public/include/manifest.php:
        * server-tests/api-manifest.js:
        (TestServer.remoteAPI.getJSON.string_appeared_here.then):
        * server-tests/tools-buildbot-triggerable-tests.js:
        (then):

2016-09-27  Dewei Zhu  <dewei_zhu@apple.com>

        Extend perf dashboard to support multiple summary pages.
        https://bugs.webkit.org/show_bug.cgi?id=162594

        Reviewed by Ryosuke Niwa.

        Start support multiple summary pages instead of one.
        Specify 'summaryPages' as key that map to a list of summaries which follows
        current 'summary' format in 'config.json' but with 2 more properties:
           'name': specifying the name shows on perf dashboard,
           'route': specifying the path to this page.

        * public/include/manifest.php:
        * public/v3/main.js:
        (main):
        (main.): Deleted.
        * public/v3/models/manifest.js:
        (Manifest._didFetchManifest):
        (Manifest):
        * public/v3/pages/summary-page.js:
        (SummaryPage):
        (SummaryPage.prototype.routeName):

2016-08-09  Ryosuke Niwa  <rniwa@webkit.org>

        Don't filter out the latest data point in chart data sampling
        https://bugs.webkit.org/show_bug.cgi?id=160714

        Reviewed by Chris Dumez.

        Exclude the last data point from sampling so that it's always included in the "sampled" charts data.
        Without this, the last data point can change as we zoom out the time domain.

        Luckily, we already had a mechanism to exclude the user selected point from sampling. Extend this
        feature by supporting an array of point IDs instead of a single ID to exclude from filering.

        * public/v3/components/interactive-time-series-chart.js:
        (InteractiveTimeSeriesChart.prototype._sampleTimeSeries): Replaced exclusionPointID by excludedPoints. 

        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart.prototype._ensureSampledTimeSeries): Put the last data point in excludedPoints.
        (TimeSeriesChart.prototype._sampleTimeSeries): Check point's id against the list of IDs.

2016-08-09  Ryosuke Niwa  <rniwa@webkit.org>

        Build fix after r204187. interval has to be a getter, not a method.

        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart.prototype._renderTimeSeries):
        * public/v3/models/measurement-adaptor.js:
        (MeasurementAdaptor.prototype.applyTo):

2016-08-09  Ryosuke Niwa  <rniwa@webkit.org>

        Fix a typo that was supposed to be fixed in r204296 before landing.

        * public/v3/pages/chart-pane.js:
        (ChartPane):
        (ChartPane.prototype.updateFromSerializedState):
        (ChartPane.prototype._analyzeRange):
        (ChartPane.prototype._renderTrendLinePopover):
        (ChartPane.prototype._trendLineTypeDidChange):

2016-08-08  Ryosuke Niwa  <rniwa@webkit.org>

        Always show segmentation on v3 charts page
        https://bugs.webkit.org/show_bug.cgi?id=160576

        Rubber-stamped by Chris Dumez.

        Added "Trend Lines" popover to select and customize a moving average or a segmentation to show on charts page
        and made Schwarz criterion segmentation the default trend line for all charts.

        Because computing the segmentation is expensive, we use WebWorker to parallelize the computation via AsyncTask.
        We also compute and cache the segmentation for each cluster separately to avoid processing the entire measurement
        set as that could take 10-20s total, which was a huge problem in v2 UI. v3 UI's approach is more incremental and
        even opens up an opportunity to cache the results in the server side.

        Also brought back "shading" for the confidence interval drawing as done in v1 and v2 UI.

        * public/shared/statistics.js:
        (Statistics.segmentTimeSeriesByMaximizingSchwarzCriterion): Added segmentCountWeight and gridSize as arguments
        to customize the algorithm.
        (Statistics.splitIntoSegmentsUntilGoodEnough): Takes segmentCountWeight as BirgeAndMassartC.

        * public/v3/async-task.js: Added.
        (AsyncTask): Added. This class represents a task such as computing segmentation to be executed in a worker.
        (AsyncTask.prototype.execute): Added. Returns a promise that gets resolved when the specified task completes.
        (AsyncTaskWorker.waitForAvailableWorker): Added. Calls the given callback with the first available worker. When
        all workers are processing some tasks, it waits until one becomes available by putting the callback into a queue.
        _didRecieveMessage pops an item out of this queue when a worker completes a task. We don't use a promise here
        because calling this function multiple times synchronously could result in all the returned promises getting
        resolved with the same worker as none of the callers get to lock away the first available worker until the end
        of the current micro-task.
        (AsyncTaskWorker._makeWorkerEventuallyAvailable): Added. A helper function for waitForAvailableWorker. Start
        a new worker if the number of workers we've started is less than the number of extra cores (e.g. 7 if there are
        8 cores on the machine). Avoid starting a new worker if we've started a new worker within the last 50 ms since
        starting a new worker takes some time.
        (AsyncTaskWorker._findAvailableWorker): Added. Finds a worker that's available right now if there is any.
        (AsyncTaskWorker): Added. An instance of AsyncTaskWorker represents a Web worker.
        (AsyncTaskWorker.prototype.id): Added.
        (AsyncTaskWorker.prototype.sendTask): Added. Sends a task represented by AsyncTask to the worker.
        (AsyncTaskWorker.prototype._didRecieveMessage): Added. This function gets called when the current task completes
        in the worker. Pop the next callback if some caller of waitForAvailableWorker is still waiting. Otherwise stop
        the worker after one second of waiting to avoid worker churning.
        (AsyncTaskWorker.workerDidRecieveMessage): Added. Called by onmessage on the worker. Executes the specified task
        and sends back a message upon completion with the appropriate timing data.

        * public/v3/components/chart-pane-base.js:
        (ChartPaneBase.prototype.configure): Uses _createSourceList.
        (ChartPaneBase.prototype._createSourceList): Added. Extracted from configure to customize the source list for
        the main chart and the overview chart.
        (ChartPaneBase.prototype._updateSourceList): Uses _createSourceList.

        * public/v3/components/chart-styles.js:
        (ChartStyles.createSourceList): Added a boolean showPoint as an extra argument. This specifies whether circles
        are drawn for each data point.
        (ChartStyles.baselineStyle): Added styles for foreground lines and background lines. They're used for trend lines
        and underlying raw data respectively when trend lines are shown.
        (ChartStyles.targetStyle): Ditto.
        (ChartStyles.currentStyle): Ditto.

        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart): Added _trendLines, _renderedTrendLines, and _fetchedTimeSeries as instance variables.
        (TimeSeriesChart.prototype.setSourceList): Clear _fetchedTimeSeries before calling setSourceList for consistency.
        (TimeSeriesChart.prototype.sourceList): Added.
        (TimeSeriesChart.prototype.clearTrendLines): Added.
        (TimeSeriesChart.prototype.setTrendLine): Added. Preserves the existing trend lines for other sources. This is
        necessary because segmentation for "current" and "baseline" lines may become available at different times, and we
        don't want to clear one or the other when setting one.
        (TimeSeriesChart.prototype._layout): Added a call to _ensureTrendLines.
        (TimeSeriesChart.prototype._renderChartContent): Call _renderTimeSeries for trend lines. Trend lines are always
        foreground lines and "regular" raw data points are drawn as background if there are trend lines.
        (TimeSeriesChart.prototype._renderTimeSeries): Added layerName as an argument. It could be an empty string,
        "foreground", or "background". Draw a "shade" just like v1 and v2 UI instead of vertical lines for the confidence
        intervals. Pick "foreground", "background", or "regular" chart style based on layerName. Also avoid drawing data
        points when *PointRadius is set to zero to reduce the runtime of this function.
        (TimeSeriesChart.prototype._sourceOptionWithFallback): Added.
        (TimeSeriesChart.prototype._ensureSampledTimeSeries): When *PointRadius is 0, show as many points as there are x
        coordinates as a fallback instead of showing every point.
        (TimeSeriesChart.prototype._ensureTrendLines): Added. Returns true if the chart contents haven't been re-rendered
        since the last update to trend lines. This flag is unset by setTrendLine.

        * public/v3/index.html:

        * public/v3/models/measurement-cluster.js:
        (MeasurementCluster.prototype.addToSeries): Store the data points' index to idMap to help aid MeasurementSet's
        _cachedClusterSegmentation efficiently re-create the segmentation from the cache.

        * public/v3/models/measurement-set.js:
        (MeasurementSet): Added _segmentationCache as an instance variable.
        (MeasurementSet.prototype.fetchSegmentation): Added. Calls _cachedClusterSegmentation on each cluster, and
        constructs the time series representation of the segmentation from the results.
        (MeasurementSet.prototype._cachedClusterSegmentation): Computes and caches the segmentation for each cluster.
        The cache of segmentation stores ID of each measurement set at which segment changes instead of its index since
        the latter could change in any moment when a new test result is reported, or an existing test result is removed
        from the time series; e.g. when it's marked as an outlier.
        (MeasurementSet.prototype._validateSegmentationCache): Added. Checks whether the cached segmentation's name and
        its parameters match that of the requested one.
        (MeasurementSet.prototype._invokeSegmentationAlgorithm): Added. Invokes the segmentation algorithm either in the
        main thread or in a Web worker via AsyncTask API based on the size of the time series. While parallelizing the
        work is beneficial when the data set is large, the overhead can add up if we keep processing a very small data
        set in a worker.

        * public/v3/models/time-series.js: Made the file compatible with Node.
        (TimeSeries.prototype.length): Added.
        (TimeSeries.prototype.valuesBetweenRange): Added.

        * public/v3/pages/chart-pane.js:
        (createTrendLineExecutableFromAveragingFunction): Added.
        (ChartTrendLineTypes): Added. Similar to StatisticsStrategies (statistics-strategies.js) in v2 UI.
        (ChartPane): Added _trendLineType, _trendLineParameters, _trendLineVersion, and _renderedTrendLineOptions as
        instance variables.
        (ChartPane.prototype.serializeState): Serialize the trend line option. This format is compatible with v2 UI.
        (ChartPane.prototype.updateFromSerializedState): Ditto. Parsing is compatible with v2 UI except that we now have
        the default trend line set when the specified ID doesn't match an existing type ID.
        (ChartPane.prototype._renderActionToolbar): Added a call to _renderTrendLinePopover. This is the popover that
        specifies the type of a trend line to show as well as its parameters.
        (ChartPane.prototype._renderTrendLinePopover): Added. A popover for specifying and customizing a trend line.
        (ChartPane.prototype._trendLineTypeDidChange): Added. Called when a new trend line is selected.
        (ChartPane.prototype._defaultParametersForTrendLine): Added.
        (ChartPane.prototype._trendLineParameterDidChange): Added. Called when the trend lines' parameters are changed.
        (ChartPane.prototype._didFetchData): Added. Overrides the one in ChartPaneBase to trigger a trend line update.
        (ChartPane.prototype._updateTrendLine): Added. Update the trend line. Since segmentation can take an arbitrary
        long time, avoid updating trend lines if this function had been called again (possibly for a different trend line
        type or with different parameters) before the results become available; hence the versioning.
        (ChartPane.paneHeaderTemplate): Added the trend line popover.
        (ChartPane.cssTemplate): Added styles for the trend line popover. Also use a more opaque background color behind
        popovers when the -webkit-backdrop-filter property is not supported.

        * public/v3/pages/dashboard-page.js:
        (DashboardPage.prototype._createChartForCell): Call createSourceList with showPoints set to true to preserve the
        existing behavior.

        * tools/js/v3-models.js: Include TimeSeries object.

        * unit-tests/measurement-set-tests.js: Added two test cases for MeasurementSet's fetchSegmentation.

        * unit-tests/resources/almost-equal.js: Added.
        (almostEqual): Extracted out of statistics-tests.js.

        * unit-tests/statistics-tests.js:

2016-08-05  Ryosuke Niwa  <rniwa@webkit.org>

        segmentTimeSeriesByMaximizingSchwarzCriterion returns a bogus result on empty charts
        https://bugs.webkit.org/show_bug.cgi?id=160575

        Rubber-stamped by Chris Dumez.

        The bug was caused by an early return in segmentTimeSeriesByMaximizingSchwarzCriterion.
        Removed this early return as the one in splitIntoSegmentsUntilGoodEnough was sufficient.

        Also factored out a few functions in findOptimalSegmentation so that they can be better
        optimized in JSC's DFG and FTL tiers, and removed unused debuggingTestingRangeNomination.

        * public/shared/statistics.js:
        (Statistics.segmentTimeSeriesByMaximizingSchwarzCriterion): Removed an early return.
        (Statistics.splitIntoSegmentsUntilGoodEnough):
        (.allocateCostUpperTriangularForSegmentation): Extracted from findOptimalSegmentation.
        (.allocatePreviousNodeForSegmentation): Ditto.
        (.findOptimalSegmentationInternal): Ditto.
        (.findOptimalSegmentation):

        * unit-tests/statistics-tests.js: Added a test case.

2016-08-05  Ryosuke Niwa  <rniwa@webkit.org>

        Perf dashboard sometimes tries to fetch a non-existent measurement-set JSON
        https://bugs.webkit.org/show_bug.cgi?id=160577

        Rubber-stamped by Chris Dumez.

        The bug was caused by findClusters computing the first cluster's endTime incorrectly. Namely, we were
        multiplying the number of clusters by clusterStart instead of clusterSize with an off-by-one error.

        * public/v3/models/measurement-set.js:
        (MeasurementSet.prototype.findClusters): Folded computeClusterStart into where clusterEnd is computed
        for clarity. Also fixed a bug that we were not computing the first cluster to fetch correctly when
        the fetched time range started before clusterStart (i.e. when startTime - clusterStart is negative).
        Finally, fixed the main bug by multiplying the number of clusters by clusterSize instead of
        clusterStart to compute the end time of the very first cluster in this measurement set. Because what
        we're computing here is the end time of the first cluster, not the start time, we also need to subtract
        one from the number of clusters. e.g. if there was exactly one cluster, then firstClusterEndTime is
        identically equal to lastClusterEndTime.
        (MeasurementSet.prototype.fetchedTimeSeries): Removed the unused argument to TimeSeries's constructor.

        * unit-tests/analysis-task-tests.js: Fixed the tests for the latest version of Mocha which complains if
        we returned a promise in unit tests when "done" function is used.
        * unit-tests/checkconfig.js: Ditto.
        * unit-tests/measurement-set-tests.js: Added a test case for findClusters and a test to make sure
        fetchBetween doesn't try to fetch a cluster before the first cluster in the set. Also fixed other test
        cases which were relying on the bug this patch fixed.

2016-08-04  Ryosuke Niwa  <rniwa@webkit.org>

        MeasurementCluster's addToSeries is slow
        https://bugs.webkit.org/show_bug.cgi?id=160581

        Rubber-stamped by Chris Dumez.

        The bulk of time was spent in MeasurementAdaptor.prototype.applyTo where we computed the interval.

        Since some of data points are filtered out by TimeSeriesChart component before intervals are used,
        we can significantly reduce the CPU time by lazily compute them. This patch reduces the runtime of
        applyTo from ~60ms to ~30ms on my machine.

        * public/v3/models/measurement-adaptor.js:
        (MeasurementAdaptor.prototype.applyTo): Lazily compute and cache the interval. Also cache the build
        object instead of always creating a new object.
        * public/v3/models/measurement-cluster.js:
        (MeasurementCluster.prototype.addToSeries): Call applyTo first before checking whether the point is
        an outlier or its id to avoid extracting those values twice since they show up in the profiler. Also
        use "of" instead "forEach" since "of" seems to be faster here.

2016-08-04  Ryosuke Niwa  <rniwa@webkit.org>

        Syncing script's configuration duplicates a lot of boilerplate
        https://bugs.webkit.org/show_bug.cgi?id=160574

        Rubber-stamped by Chris Dumez.

        This patch makes each configuration accept an array of platforms and types so that we can write:

        {"type": "speedometer", "builder": "mba", "platform": "Trunk El Capitan MacBookAir"},
        {"type": "speedometer", "builder": "mbp", "platform": "Trunk El Capitan MacBookPro"},
        {"type": "speedometer", "builder": "mba", "platform": "Trunk Sierra MacBookAir"},
        {"type": "speedometer", "builder": "mbp", "platform": "Trunk Sierra MacBookPro"},
        {"type": "jetstream", "builder": "mba", "platform": "Trunk El Capitan MacBookAir"},
        {"type": "jetstream", "builder": "mbp", "platform": "Trunk El Capitan MacBookPro"},
        {"type": "jetstream", "builder": "mba", "platform": "Trunk Sierra MacBookAir"},
        {"type": "jetstream", "builder": "mbp", "platform": "Trunk Sierra MacBookPro"},

        more concisely as:

        {"builder": "mba", "types": ["speedometer", "jetstream"],
            "platforms": ["Trunk El Capitan MacBookAir", "Trunk Sierra MacBookAir"]},
        {"builder": "mbp", "types": ["speedometer", "jetstream"],
            "platforms": ["Trunk El Capitan MacBookPro", "Trunk Sierra MacBookPro"]},

        * tools/js/buildbot-syncer.js:
        (BuildbotSyncer._loadConfig):
        (BuildbotSyncer._expandTypesAndPlatforms): Added. Clones a new configuration entry for each type
        and platform.
        (BuildbotSyncer._createTestConfiguration): Extracted from _loadConfig.
        (BuildbotSyncer._validateAndMergeConfig): Added a new argument that specifies a property that
        shouldn't be merged into the configuration. Also added the support for 'types' and 'platforms',
        and merged the code for verify an array of strings. Finally, allow the appearance of 'properties'
        since this function can now be called on a cloned configuration in which 'arguments' had already
        been renamed to 'properties'.

        * unit-tests/buildbot-syncer-tests.js: Added a test case to parse a consolidated configuration.
        (sampleiOSConfigWithExpansions): Added.

        * unit-tests/resources/mock-v3-models.js:
        (MockModels.inject): Added a few more mock models for the newly added test.

2016-07-26  Ryosuke Niwa  <rniwa@webkit.org>

        REGRESSION: Tooltip for analysis tasks doesn't show up on charts
        https://bugs.webkit.org/show_bug.cgi?id=160221

        Rubber-stamped by Chris Dumez.

        The bug was caused by ChartPaneBase resetting annotation bars every time the current point has moved.
        Avoid doing this in ChartPaneBase's _renderAnnotations().

        * public/v3/components/chart-pane-base.js:
        (ChartPaneBase):
        (ChartPaneBase.prototype.fetchAnalysisTasks):
        (ChartPaneBase.prototype._renderAnnotations):

2016-07-26  Ryosuke Niwa  <rniwa@webkit.org>

        REGRESSION: The arrow indicating the current page doesn't get updated
        https://bugs.webkit.org/show_bug.cgi?id=160185

        Reviewed by Chris Dumez.

        The bug was caused by Heading's render() function not updating the DOM more than once. I don't understand
        how this has ever worked. Fixed the bug by rendering DOM whenever the current page has changed.

        * public/v3/pages/heading.js:
        (Heading):
        (Heading.prototype.render):

2016-07-25  Ryosuke Niwa  <rniwa@webkit.org>

        Build fix for Safari 9.

        * public/v3/models/build-request.js:
        (BuildRequest.prototype.waitingTime): Don't use "const" in strict mode.

2016-07-23  Ryosuke Niwa  <rniwa@webkit.org>

        Perf dashboard should show the list of a pending A/B testing jobs
        https://bugs.webkit.org/show_bug.cgi?id=160138

        Rubber-stamped by Chris Dumez.

        Add a page to show the list of A/B testing build requests per triggerable. Ideally, we would like to
        see a queue per builder but that would require changes to database tables and syncing scripts.

        Because this page is most useful when the analysis task with which each build request is associated,
        JSON API at /api/build-requests/ has been modified to return the analysis task ID for each request.

        Also streamlined the page that shows the list of analysis tasks per Chris' feedback by consolidating
        "Bisecting" and "Identified" into "Investigated" and moving the toolbar from the upper left corner
        inside the heading to right beneath the heading above the table. Also made the category page serialize
        the filter an user had typed in so that reloading the page doesn't clear it.

        * public/api/analysis-tasks.php:
        (fetch_associated_data_for_tasks): Removed 'category' from the list of columns returned as the notion
        of 'category' is only relevant in UI, and it's better computed in the front-end.
        (format_task): Ditto.
        (determine_category): Deleted.

        * public/api/test-groups.php:
        (main):

        * public/include/build-requests-fetcher.php:
        (BuildRequestsFetcher::fetch_for_task): Include the analysis task ID in the rows.
        (BuildRequestsFetcher::fetch_for_group): Ditto. Ditto.
        (BuildRequestsFetcher::fetch_incomplete_requests_for_triggerable): Ditto.
        (BuildRequestsFetcher::results_internal): Ditto.

        * public/v3/index.html:

        * public/v3/main.js:
        (main): Create a newly introduced BuildRequestQueuePage as a subpage of AnalysisCategoryPage.

        * public/v3/components/ratio-bar-graph.js:
        (RatioBarGraph.prototype.update): Fixed a bogus assertion here. ratio can be any number. The coercion
        into [-1, 1] is done inside RatioBarGraph's render() function.

        * public/v3/models/analysis-task.js:
        (AnalysisTask.prototype.category): Moved the code to compute the analysis task's category from
        determine_category in analysis-tasks.php. Also merged "bisecting" and "identified" into "investigated".
        (AnalysisTask.categories): Merged "bisecting" and "identified" into "investigated".

        * public/v3/models/build-request.js:
        (BuildRequest): Remember the triggerable and the analysis task associated with this request as well as
        the time at when this request was created.        
        (BuildRequest.prototype.analysisTaskId): Added.
        (BuildRequest.prototype.statusLabel): Use a shorter label: "Waiting" for "pending" status.
        (BuildRequest.prototype.createdAt): Added.
        (BuildRequest.prototype.waitingTime): Added. Returns a human readable time duration since the creation
        of this build request such as "2 hours 21 minutes" against a reference time.
        (BuildRequest.fetchTriggerables): Added.
        (BuildRequest.cachedRequestsForTriggerableID): Added. Used when navigating back to 

        * public/v3/pages/analysis-category-page.js:
        (AnalysisCategoryPage): Construct AnalysisCategoryToolbar and store it in this._categoryToolbar since it
        no longer inherits from Toolbar class, which PageWithHeading recognizes and stores.
        (AnalysisCategoryPage.prototype.title):
        (AnalysisCategoryPage.prototype.serializeState): Added.
        (AnalysisCategoryPage.prototype.stateForCategory): Added. Include the filter in the serialization.
        (AnalysisCategoryPage.prototype.updateFromSerializedState): Restore the filter from the URL state.
        (AnalysisCategoryPage.prototype.filterDidChange): Added. Called by AnalysisCategoryToolbar to update
        the URL state in addition to calling render() as done previously via setFilterCallback.
        (AnalysisCategoryPage.prototype.render): Always call _categoryToolbar.render() since the hyperlinks for
        the category pages now include the filter, which can be updated in each call.
        (AnalysisCategoryPage.cssTemplate):

        * public/v3/pages/analysis-category-toolbar.js:
        (AnalysisCategoryToolbar): Inherits from ComponentBase instead of Toolbar since AnalysisCategoryToolbar
        no longer works with Heading class unlike other subclasses of Toolbar class.
        (AnalysisCategoryToolbar.prototype.setCategoryPage): Added.
        (AnalysisCategoryToolbar.prototype.setFilterCallback): Deleted.
        (AnalysisCategoryToolbar.prototype.setFilter): Added. Used to restore from a serialized URL state.
        (AnalysisCategoryToolbar.prototype.render): Don't recreate the input element as it clears the value as
        well as the selection of the element. Also use AnalysisCategoryPage's stateForCategory to serialize the
        category name and the current filter for each hyperlink.
        (AnalysisCategoryToolbar.prototype._filterMayHaveChanged): Now takes an boolean argument specifying
        whether the URL state should be updated or not. We update the URL only when a change event is fired to
        avoid constantly updating it while an user is still typing.
        (AnalysisCategoryToolbar.cssTemplate): Added.
        (AnalysisCategoryToolbar.htmlTemplate): Added a button to open the newly added queue page.

        * public/v3/pages/build-request-queue-page.js:
        (BuildRequestQueuePage): Added.
        (BuildRequestQueuePage.prototype.routeName): Added.
        (BuildRequestQueuePage.prototype.pageTitle): Added.
        (BuildRequestQueuePage.prototype.open): Added. Fetch open build requests for every triggerables using
        the same API as the syncing scripts.
        (BuildRequestQueuePage.prototype.render): Added.
        (BuildRequestQueuePage.prototype._constructBuildRequestTable): Added. Construct a table for the list of
        pending, scheduled or running build requests in the order syncing scripts would see. Note that the list
        of build requests returned by /api/build-requests/* can contain completed, canceled, or failed requests
        since the JSON returns all build requests associated with each test group if one of the requests of the
        group have not finished. This helps syncing scripts picking the right builder for A/B testing when it
        had previously been unloaded or crashed in the middle of processing a test group. This characteristics
        of the API actually helps us here because we can reliably compute the total number of build requests in
        the group. The first half of this function does this counting as well as collapses all but the first
        unfinished build requests into a "contraction" row, which just shows the number of build requests that
        are remaining in the group.
        (BuildRequestQueuePage.cssTemplate): Added.
        (BuildRequestQueuePage.htmlTemplate): Added.

        * public/v3/pages/summary-page.js:
        (SummaryPage.prototype.open): Use one-day median instead of seven-day median to compute the status.
        (SummaryPageConfigurationGroup): Initialize _ratio to NaN. This was causing assertion failures in
        RatioBarGraph's update() while measurement sets are being fetched.

        * server-tests/api-build-requests-tests.js: Updated the tests per change in BuildRequest's statusLabel.
        * unit-tests/analysis-task-tests.js: Ditto.
        * unit-tests/test-groups-tests.js: Ditto.
        * unit-tests/build-request-tests.js: Added tests for BuildRequest's waitingTime.

2016-07-22  Ryosuke Niwa  <rniwa@webkit.org>

        REGRESSION(r203035): Marking points as an outlier no longer updates charts
        https://bugs.webkit.org/show_bug.cgi?id=160106

        Reviewed by Darin Adler.

        The bug was caused by MeasurementSet's fetchBetween clearing previously registered callbacks when noCache
        option is specified.

        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart.prototype.setSourceList): Clear this._fetchedTimeSeries when changing chart options.
        e.g. need to start including or excluding outliers.
        (TimeSeriesChart.prototype.fetchMeasurementSets): Don't skip the fetching when noCache is true.

        * public/v3/models/measurement-set.js:
        (MeasurementSet): Added this._callbackMap as an instance variable to keep track of all callbacks on every
        cluster since we may need to call each callback multiple times per cluster when noCache option is used.
        (MeasurementSet.prototype.fetchBetween): Moved the code to add _primaryClusterPromise to _allFetches here
        so that now this function and _ensureClusterPromise are only functions that touch _allFetches.
        (MeasurementSet.prototype._ensureClusterPromise): Extracted out of fetchBetween. Queue up all callbacks
        for each cluster when creating a new promise.
        (MeasurementSet.prototype._fetchPrimaryCluster): Removed the code to add _primaryClusterPromise now that
        it's done in fetchBetween.

        * public/v3/remote.js:
        (RemoteAPI.postJSONWithStatus): Removed superfluous call to console.log.

        * unit-tests/measurement-set-tests.js: Updated the test case for noCache. The callback registered before
        fetchBetween is called with noCache=true is now invoked so callCount must be 3 instead of 2.

2016-07-19  Ryosuke Niwa  <rniwa@webkit.org>

        Perf dashboard always re-generate measurement set JSON
        https://bugs.webkit.org/show_bug.cgi?id=159951

        Reviewed by Chris Dumez.

        The bug was caused by manifest.json reporting the last modified date of a measurement set in floating point,
        and a measurement set JSON reporting it as an integer. Fixed the bug by always using an integer.

        * public/api/measurement-set.php:
        (main): Return 404 when the results is empty.
        (MeasurementSetFetcher::execute_query): Use "extract(epoch from commit_time)" like ManifestGenerator to improve
        the generation speed. This is ~10% runtime improvement.
        (MeasurementSetFetcher::format_map): Updated to reflect the above change.
        (MeasurementSetFetcher::parse_revisions_array): Ditto.

        * public/include/manifest.php:
        (ManifestGenerator::platforms): Fixed the bug by coercing lastModified to integer (instead of float).

        * server-tests/api-measurement-set-tests.js: Added a test case for returning empty results, and a test case for
        making sure lastModified dates in manifest.json and measurement sets match.

        * tools/js/remote.js:
        (RemoteAPI.prototype.sendHttpRequest): Reject the promise when HTTP status code is not 200.

2016-07-09  Ryosuke Niwa  <rniwa@webkit.org>

        Perf dashboard can consume 50-70% of CPU on MacBook even if user is not interacting at all
        https://bugs.webkit.org/show_bug.cgi?id=159597

        Reviewed by Chris Dumez.

        TimeSeriesChart and InteractiveTimeSeriesChart had been relying on continually polling on requestAnimationFrame
        to update itself in response to its canvas resizing. Even though there as an early exit in the case there was
        nothing to update, this is still causing a significant power drain when the user is not interacting at all.

        Let TimeSeriesChart use the regular top-down render path like other components with exceptions of listening to
        window's resize eventas well as when new JSONs are fetched from the server. The render() call to the latter case
        will be coerced into a single callback on requestAnimationFrame to avoid DOM-mutation-layout churn.

        * public/v3/components/base.js:
        (ComponentBase.isElementInViewport): Deleted.
        * public/v3/components/chart-pane-base.js:
        (ChartPaneBase.prototype.render): Enqueue charts to render.
        * public/v3/components/chart-styles.js:
        (ChartStyles.dashboardOptions): Removed updateOnRequestAnimationFrame which is no longer an available option.
        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart): Replaced the code to register itself for rAF by the code to listen to resize events on window.
        (TimeSeriesChart._updateOnRAF): Deleted.
        (TimeSeriesChart._updateAllCharts): Added.
        (TimeSeriesChart.prototype._enqueueToRender): Added.
        (TimeSeriesChart._renderEnqueuedCharts): Added.
        (TimeSeriesChart.prototype.fetchMeasurementSets): Avoid calling fetchBetween when the range had been fetched.
        Without this change, we can incur a significant number of redundant calls to render() when adjusting the domain
        in charts page by the slider. When no new JSON is fetched, simply enqueue this chart to render on rAF.
        (TimeSeriesChart.prototype._didFetchMeasurementSet): Enqueue this chart to render on rAF.
        (TimeSeriesChart.prototype.render): Removed the check for isElementInViewport since we no longer get render() call
        when this chart moves into the viewport (as we no longer listen to every rAF or scroll event).
        * public/v3/models/measurement-set.js:
        (MeasurementSet.prototype.hasFetchedRange): Fixed various bugs revealed by the new use in fetchMeasurementSets.
        * public/v3/pages/chart-pane-status-view.js:
        (ChartPaneStatusView.prototype._updateRevisionListForNewCurrentRepository): Removed the dead code. It was probably
        copied from when this code was in InteractiveTimeSeries chart. There is no this._forceRender in this component.
        * public/v3/pages/dashboard-page.js:
        (DashboardPage.prototype.render): Enqueue charts to render. 
        * public/v3/pages/heading.js:
        (SummaryPage.prototype.open): Removed the call to isElementInViewport. This wasn't really doing anything useful in
        this component.
        * unit-tests/measurement-set-tests.js: Added tests for hasFetchedRange.
        (.waitForMeasurementSet): Moved to be used in a newly added test case.

2016-07-09  Ryosuke Niwa  <rniwa@webkit.org>

        REGRESSION: manifest.json generation takes multiple seconds on perf dashboard
        https://bugs.webkit.org/show_bug.cgi?id=159596

        Reviewed by Chris Dumez.

        The most of CPU time was spent looking for a duplicate entry in an array of metrics by array_search.
        This patch moves to postgres by using aggregate functions in the query. Also moved the code to convert
        datetime to UNIX epoch timestamp from PHP to within postgres query.

        These improvements reduce total runtime of Manifest::generate from ~4s to ~350ms on my machine.

        * public/include/manifest.php:
        (Manifest::generate): No longer fetches test_configurations table as this is done in Manifest::platforms now.
        Also moved calls to each method in the class to separate lines for easier instrumentation.
        (Manifest::platforms): Group test configurations (current, baseline, target) by platform and metric.
        Use the max of the last modified dates in UNIX epoch timestamps (ms to be compatible with JS's representation).
        A given platform, metric pair is considered to be in the v1 dashboard if any test configuration is in.

2016-07-08  Ryosuke Niwa  <rniwa@webkit.org>

        bundle-v3-scripts.py should compress HTML/CSS templates
        https://bugs.webkit.org/show_bug.cgi?id=159582

        Reviewed by Joseph Pecoraro.

        Strip leading and trailing whitespaces from HTML and CSS templates. This is a 8% progression on the file size.

        * Install.md: Updated the list of MIME types to apply deflate for newer versions of Apache.
        * tools/bundle-v3-scripts.py:
        (main):
        (compress_template): Added.

2016-06-13  Ryosuke Niwa  <rniwa@webkit.org>

        Build fix. Strip out "use strict" everywhere so that the perf dashboard works on the shipping Safari.

        * tools/bundle-v3-scripts.py:
        (main):

2016-06-13  Ryosuke Niwa  <rniwa@webkit.org>

        Invalid token error when trying to create an A/B analysis for a range
        https://bugs.webkit.org/show_bug.cgi?id=158679

        Reviewed by Chris Dumez.

        The problem in this particular case was due to another website overriding cookies for our subdomain.
        Make PrivilegedAPI robust against its token becoming invalid in general to fix the bug since the cookie
        is only available under /privileged-api/ and the v3 UI can't access it for security reasons.

        This patch factors out PrivilegedAPI out of remote.js so that it can be tested separately in server tests
        as well as unit tests even though RemoteAPI itself is implemented differently in each case.

        * init-database.sql: Added a forgotten default value "false" to run_marked_outlier.
        * public/v3/index.html:
        * public/v3/privileged-api.js: Added. Extracted out of public/v3/remote.js.
        (PrivilegedAPI.sendRequest): Fixed the bug. When the initial request fails with "InvalidToken" error,
        re-generate the token and re-issue the request.
        (PrivilegedAPI.requestCSRFToken):
        * public/v3/remote.js:
        (RemoteAPI.postJSON): Added to match tools/js/remote.js.
        (RemoteAPI.postJSONWithStatus): Ditto.
        (PrivilegedAPI): Moved to privileged-api.js.
        * server-tests/api-measurement-set-tests.js: Removed the unused require for crypto.
        * server-tests/privileged-api-upate-run-status.js: Added tests for /privileged-api/update-run-status.
        * server-tests/resources/test-server.js:
        (TestServer.prototype.inject): Clear the cookies as well as tokens in PrivilegedAPI.
        * tools/js/remote.js:
        (RemoteAPI): Added the support for PrivilegedAPI by making cookie set by the server persist.
        (RemoteAPI.prototype.clearCookies): Added for tests.
        (RemoteAPI.prototype.postJSON): Make sure sendHttpRequest always sends a valid JSON.
        (RemoteAPI.prototype.postJSONWithStatus): Added since this API is used PrivilegedAPI.
        (RemoteAPI.prototype.sendHttpRequest): Retain the cookie set by the server and send it back in each request.
        * tools/js/v3-models.js:
        * unit-tests/privileged-api-tests.js: Added unit tests for PrivilegedAPI.
        * unit-tests/resources/mock-remote-api.js:
        (MockRemoteAPI.postJSON): Added for unit testing.
        (MockRemoteAPI.postJSONWithStatus): Ditto.

2016-06-13  Ryosuke Niwa  <rniwa@webkit.org>

        /admin/tests is very slow
        https://bugs.webkit.org/show_bug.cgi?id=158682

        Reviewed by Chris Dumez.

        The slowness came from TestNameResolver::__construct, which was fetching the entire table of test_configurations,
        which at this point contains more than 32,000 rows. Don't fetch the entire table in the constructor. Instead,
        fetch a subset of rows as needed in configurations_for_metric_and_platform. Even though this results in many SQL
        queries being issued, that's a lot more efficient in practice because we only fetch a few dozen rows in practice.

        Also removed a whole bunch of features from /admin/tests to simplify the page. In particular, the ability to update
        the list of triggerables has been removed now that sync-buildbot.js automatically updates that for us. This removed
        the last use of test_exists_on_platform, which was also dependent on fetching test_configurations upfront.

        * public/admin/tests.php:
        * public/include/test-name-resolver.php:
        (TestNameResolver::__construct): Don't fetch the entire table of test_configurations.
        (TestNameResolver::configurations_for_metric_and_platform): Just issue a SQL query for the specified platform and metric.
        (TestNameResolver::test_exists_on_platform): Removed.

2016-06-08  Ryosuke Niwa  <rniwa@webkit.org>

        sync-buildbot.js should update the list of tests and platforms associated with a triggerable
        https://bugs.webkit.org/show_bug.cgi?id=158406

        Reviewed by Chris Dumez.

        Add /api/update-triggerable and a test for it, which were supposed to be added in r201718
        but for which I forgot to run svn add.

        * public/api/update-triggerable.php: Added.
        (main):
        * server-tests/api-update-triggerable.js: Added.

2016-06-07  Ryosuke Niwa  <rniwa@webkit.org>

        Build fix after r201739. attachShadow was throwing an exception on this component.

        * public/v3/pages/analysis-category-toolbar.js:
        (AnalysisCategoryToolbar):

2016-06-06  Ryosuke Niwa  <rniwa@webkit.org>

        sync-buildbot.js should update the list of tests and platforms associated with a triggerable
        https://bugs.webkit.org/show_bug.cgi?id=158406
        <rdar://problem/26185737>

        Reviewed by Darin Adler.

        Added /api/update-triggerable to update the list of configurations (platform and test pairs)
        associated with a given triggerable, and make sync-buildbot.js use this JSON API before each
        syncing cycle so that the association gets updated automatically by simply updating the JSON.

        * server-tests/api-manifest.js: Use const for imported modules.
        * server-tests/api-report-commits-tests.js: Removed unnecessary importing of crypto.
        * server-tests/resources/mock-data.js:
        (MockData.someTestId): Added.
        (MockData.somePlatformId): Added.
        (MockData.addMockData):
        * server-tests/tools-buildbot-triggerable-tests.js: Use const for imported modules. Also added
        a test for BuildbotTriggerable's updateTriggerable.
        * tools/js/buildbot-triggerable.js:
        (BuildbotTriggerable.prototype.updateTriggerable): Added. Find the list of all configurations
        associated with this triggeerable and post it to /api/update-triggerable.
        * tools/js/database.js: Added triggerable_configurations to the list of tables.
        * tools/js/remote.js:
        (RemoteAPI.prototype.postJSON): Print the whole response when JSON parsing fails for debugging.
        * tools/sync-buildbot.js:
        (syncLoop): Call BuildbotTriggerable's updateTriggerable before syncing.

2016-06-02  Ryosuke Niwa  <rniwa@webkit.org>

        Build fix after r201564.

        * public/v3/pages/analysis-category-page.js:
        (AnalysisCategoryPage.prototype.updateFromSerializedState):
        * public/v3/pages/create-analysis-task-page.js:
        (CreateAnalysisTaskPage.prototype.updateFromSerializedState):
        (CreateAnalysisTaskPage.prototype.render):

2016-05-31  Ryosuke Niwa  <rniwa@webkit.org>

        v3 UI should support marking and unmarking outliers as well as hiding them
        https://bugs.webkit.org/show_bug.cgi?id=158248

        Rubber-stamped by Chris Dumez.

        Added the support for marking and unmarking a sequence of points as outliers. Unlike v2, we now support marking
        multiple points as outliers in a single click. Also fixed a bug that outliers are never explicitly hidden in v3 UI.

        This patch splits ChartStyles.createChartSourceList into two functions: resolveConfiguration and createSourceList
        to separate the work of resolving platform and metric IDs to their respective model objects, and creating a source
        list used by TimeSeriesChart to fetch measurement sets. createSourceList is called again when filtering options are
        changed.

        It also adds noCache option to TimeSeriesChart's fetchMeasurementSets, MeasurementSet's fetchBetween and
        _fetchPrimaryCluster to update the measurement sets after marking or unmarking points as outliers. In addition, it
        fixes a bug that the annotation bars for analysis tasks are not updated in charts page after creating an analysis
        task by adding noCache option to ChartPaneBase's fetchAnalysisTasks, AnalysisTask's fetchByPlatformAndMetric and
        _fetchSubset.

        Finally, this patch splits ChartPane._makeAnchorToOpenPane into _makePopoverActionItem, _makePopoverOpenOnHover and
        _setPopoverVisibility for clarity.

        * public/v3/components/chart-pane-base.js:
        (ChartPaneBase): Added _disableSampling and _showOutliers as instance variables.
        (ChartPaneBase.prototype.configure):
        (ChartPaneBase.prototype.isSamplingEnabled): Added.
        (ChartPaneBase.prototype.setSamplingEnabled): Added. When a filtering option is updated, recreate the source list
        so that TimeSeriesChart.setSourceList can re-fetch the measurement set JSONs.
        (ChartPaneBase.prototype.isShowingOutliers): Added.
        (ChartPaneBase.prototype.setShowOutliers): Added. Ditto for calling _updateSourceList.
        (ChartPaneBase.prototype._updateSourceList): Added.
        (ChartPaneBase.prototype.fetchAnalysisTasks): Renamed from _fetchAnalysisTasks. Now takes noCache as an argument
        instead of platform and metric IDs since they're on instance variables.

        * public/v3/components/chart-styles.js:
        (ChartStyles.resolveConfiguration): Renamed from createChartSourceList. Just resolves platform and metric IDs.
        (ChartStyles.createSourceList): Extracted from createChartSourceList since it needs to be called when a filtering
        option is changed as well as when ChartPaneBase.prototype.configure is called.
        (ChartStyles.baselineStyle): Now takes filtering options.
        (ChartStyles.targetStyle): Ditto.
        (ChartStyles.currentStyle): Ditto.

        * public/v3/components/interactive-time-series-chart.js:
        (InteractiveTimeSeriesChart.prototype.currentPoint): Find the point in _fetchedTimeSeries when
        _sampledTimeSeriesData hasn't been computed yet as a fallback (e.g. when the chart hasn't been rendered yet).
        (InteractiveTimeSeriesChart.prototype.selectedPoints): Added.
        (InteractiveTimeSeriesChart.prototype.firstSelectedPoint): Added.
        (InteractiveTimeSeriesChart.prototype.lockedIndicator): Added. Returns the current point if it's locked.

        * public/v3/components/time-series-chart.js:
        (TimeSeriesChart.prototype.setDomain):
        (TimeSeriesChart.prototype.setSourceList): Added. Re-create _fetchedTimeSeries when filtering options have changed.
        Don't re-fetch measurement set JSONs here since showing outliers can be done entirely in the front end.
        (TimeSeriesChart.prototype.fetchMeasurementSets): Extracted out of setDomain. Now takes noCache as an argument.
        ChartPane._markAsOutlier
        (TimeSeriesChart.prototype.firstSampledPointBetweenTime): Added.

        * public/v3/models/analysis-task.js:
        (AnalysisTask.fetchByPlatformAndMetric): Added noCache as an argument.
        (AnalysisTask._fetchSubset): Ditto.

        * public/v3/models/measurement-adaptor.js:
        (MeasurementAdaptor.prototype.isOutlier): Added.
        (MeasurementAdaptor.prototype.applyToAnalysisResults): Add markedOutlier as a property on each point.

        * public/v3/models/measurement-cluster.js:
        (MeasurementCluster.prototype.addToSeries): Fixed the bug that filtering outliers was broken as _markedOutlierIndex
        is undefined here. Use MeasurementAdaptor's isOutlier instead.

        * public/v3/models/measurement-set.js:
        (MeasurementSet.prototype.fetchBetween): Added noCache as an argument. Reset _primaryClusterPromise and _allFetches
        when noCache is true since we need to re-fetch the primary cluster as well as all secondary clusters now.
        (MeasurementSet.prototype._fetchPrimaryCluster): Added noCache as an argument. Directly invoke the JSON API at
        /api/measurement-set to re-generate all clusters' JSON files instead of first fetching the cached version.
        (MeasurementSet.prototype._fetchSecondaryCluster):
        (MeasurementSet.prototype._didFetchJSON): Removed a bogus assertion since this function is called on secondary
        clusters as well as primary clusters.
        (MeasurementSet.prototype._addFetchedCluster): Reimplemented this function using an insertion sort. Also remove the
        existing entry if the fetch cluster should replace it.

        * public/v3/models/time-series.js:
        (TimeSeries.prototype.dataBetweenPoints): Removed the dead code to filter out outliers. This is done in addToSeries
        of MeasurementCluster instead.

        * public/v3/pages/chart-pane.js:
        (ChartPane): Renamed pane to popover since it was confusing to have a pane inside a pane class. As such, renamed
        _paneOpenedByClick to _lockedPopover.
        (ChartPane.prototype.serializeState): Added the code to serialize filtering options in the serialized state URL.
        (ChartPane.prototype.updateFromSerializedState): Ditto for parsing.
        (ChartPane.prototype._analyzeRange): Extracted out of render(). Also fixed a bug that the charts page don't show
        the newly created analysis task by invoking fetchAnalysisTasks with noCache set to true.
        (ChartPane.prototype._markAsOutlier): Added.
        (ChartPane.prototype._renderActionToolbar): A bunch of changes due to pane -> popover rename. Also added a popover
        for filtering options.
        (ChartPane.prototype._makePopoverActionItem): Extracted from _makeAnchorToOpenPane.
        (ChartPane.prototype._makePopoverOpenOnHover): Ditto.
        (ChartPane.prototype._setPopoverVisibility): Ditto.
        (ChartPane.prototype._renderFilteringPopover): Added.
        (ChartPane.htmlTemplate): Added a popover for specifying filtering options. Also added .popover on each popover.
        (ChartPane.cssTemplate): Updated the style to make use of .popover.

        * public/v3/pages/charts-page.js:
        (ChartsPage.prototype.graphOptionsDidChange): Added. Updates the URL state when a filtering option is modified.

        * public/v3/pages/dashboard-page.js:
        (DashboardPage.prototype._createChartForCell):

        * public/v3/pages/page-router.js:
        (PageRouter.prototype._serializeHashQueryValue): Serialize a set of strings as | separated tokens.
        (PageRouter.prototype._deserializeHashQueryValue): Rewrote the function as the serialized URL can no longer be
        parsed as a JSON as | separated tokens can't be converted into a valid JSON construct with a simple regex.

        * unit-tests/measurement-set-tests.js: Added a test case for fetchBetween with noCache=true.

2016-05-24  Ryosuke Niwa  <rniwa@webkit.org>

        Another build fix after r201307.

        * public/v3/pages/page-router.js:
        (PageRouter.prototype._deserializeHashQueryValue):
        (PageRouter.prototype._countOccurrences): Moved from _deserializeHashQueryValue.

2016-05-23  Ryosuke Niwa  <rniwa@webkit.org>

        Build fix after r201307.

        * public/v3/pages/chart-pane.js:
        (ChartPane.prototype.serializeState):

2016-05-23  Ryosuke Niwa  <rniwa@webkit.org>

        Some applications truncates the last closing parenthesis in perf dashboard URL
        https://bugs.webkit.org/show_bug.cgi?id=157976

        Reviewed by Tim Horton.

        Unfortunately, different applications use different heuristics to determine the end of each URL. Two trailing
        parentheses, for example, is just fine in Radar as well as Apple Mail if the URL is short enough. Using other
        characters such as ] and } wouldn't work either because they would be %-escaped. At that point, we might as well
        as %-escape everything.

        Work around the bug by parsing the URL as if it had one extra ')' if the parsing had failed. Also shorten the charts
        page's URL by avoid emitting "-null" for each pane when the pane doesn't have a currently selected point or selection.
        This improves the odds of the entire URL being recognized by various applications.

        We could, in theory, implement some sort of a URL shorter but that can wait until when we support real user accounts.

        * public/v3/pages/chart-pane.js:
        (ChartPane.prototype.serializeState): Don't serialize the selection or the current point if nothing is selected.
        * public/v3/pages/page-router.js:
        (PageRouter.prototype._deserializeHashQueryValue): Try parsing the value again with one extra ] at the end if
        the JSON parsing had failed.

2016-05-18  Ryosuke Niwa  <rniwa@webkit.org>

        Perf dashboard "Add pane" should list first by test, then by machine
        https://bugs.webkit.org/show_bug.cgi?id=157880

        Reviewed by Stephanie Lewis.

        Reversed the order which tests and platforms are selected. Also split .pane-selector-container into #tests and
        #platform for the ease of DOM node manipulations.

        * public/v3/components/pane-selector.js:
        (PaneSelector):
        (PaneSelector.prototype._renderPlatformList): Renamed from _renderPlatformLists since there is a single list
        for platforms. This list now disappears while a non-metric item is selected in the collection of test lists.
        e.g. "Speedometer" instead of its "Score" metric. Remember the last metric we rendered to avoid churning.
        (PaneSelector.prototype._renderTestLists): Render the top level tests once. The index of lists have been
        decreased by one since test lists are now inside #tests instead of appearing after the platform list.
        (PaneSelector.prototype._buildTestList): Don't filter tests since platform is chosen after tests now.
        (PaneSelector.prototype._replaceList):
        (PaneSelector.prototype._selectedItem): Don't reset the test path (specifies which subtest or metric is picked)
        when a platform is selected since it happens after a test metric is chosen now.
        (PaneSelector.prototype._clickedItem): Add a pane when a platform is clicked, not when a metric is clicked.
        (PaneSelector.cssTemplate):

2016-05-18  Ryosuke Niwa  <rniwa@webkit.org>

        Analysis task should look for a git commit based on abridged hashes
        https://bugs.webkit.org/show_bug.cgi?id=157877
        <rdar://problem/26254374>

        Reviewed by Chris Dumez.

        Made /privileged-api/associate-commit look for commits using LIKE instead of an exact match.
        Associate the commit when there is exactly one match.

        * public/include/commit-log-fetcher.php:
        (CommitLogFetcher::fetch_between):
        * public/include/db.php:
        (Database::escape_for_like): Extracted from CommitLogFetcher::fetch_between.
        * public/privileged-api/associate-commit.php:
        (main): Look for the commits using LIKE. Reject whenever there are multiple commits. We limit the number of
        matches to two for performance when the user specifies something that almost thousands of commits: e.g. "1".
        * public/v3/pages/analysis-task-page.js:
        (AnalysisTaskPage.prototype._associateCommit): Added human friendly error messages for mismatching commits.

2016-05-18  Ryosuke Niwa  <rniwa@webkit.org>

        Unreviewed build fix. Use --date-order so that every child commit appears after its parent.
        Otherwise we'll hit a FailedToFindParentCommit error while submitting a commit that appears before its parent.

        * tools/sync-commits.py:
        (GitRepository._fetch_all_hashes):

2016-05-18  Ryosuke Niwa  <rniwa@webkit.org>

        Removed the erroneously committed debug code.

        * tools/sync-commits.py:
        (GitRepository.fetch_commit):

2016-05-18  Ryosuke Niwa  <rniwa@webkit.org>

        Perf dashboard should have a script to sync git commits
        https://bugs.webkit.org/show_bug.cgi?id=157867

        Reviewed by Chris Dumez.

        Added the support to pull from a Git repo to pull-svn.py and renamed it to sync-commits.py.

        Added two classes SVNRepository and GitRepository which inherits from an abstract class, Repository.
        The code that fetches commit and format revision number / git hash is specialized in each.

        * Install.md:
        * tools/pull-svn.py: Removed.
        * tools/sync-commits.py: Renamed from Websites/perf.webkit.org/tools/pull-svn.py.
        (main): Renamed --svn-config-json to --repository-config-json. Also made it robust against exceptions
        inside fetch_commits_and_submit of each Repository class.
        (load_repository): A factory function for SVNRepository and GitRepository.
        (Repository): Added.
        (Repository.__init__): Added.
        (Repository.fetch_commits_and_submit): Extracted from standalone fetch_commits_and_submit.
        (Repository.fetch_commit): Added. Implemented by a subclass.
        (Repository.format_revision): Ditto.
        (Repository.determine_last_reported_revision): Extracted from alonealone
        determine_first_revision_to_fetch. The fallback to use "oldest" has been moved to SVNRepository's
        fetch_commit since it doesn't work in Git.
        (Repository.fetch_revision_from_dasbhoard): Extracted from fetch_revision_from_dasbhoard.
        (SVNRepository): Added.
        (SVNRepository.__init__): Added.
        (SVNRepository.fetch_commit): Extracted from standalone fetch_commit_and_resolve_author and fetch_commit.
        (SVNRepository._resolve_author_name): Renamed from resolve_author_name_from_account.
        (SVNRepository.format_revision): Added.
        (GitRepository): Added.
        (GitRepository.__init__):
        (GitRepository.fetch_commit): Added. Fetches the list of all git hashes if needed, and finds the next hash.
        (GitRepository._find_next_hash): Added. Finds the first commit that appears after the specified hash.
        (GitRepository._fetch_all_hashes): Added. Gets the list of all git hashs chronologically (old to new).
        (GitRepository._run_git_command): Added.
        (GitRepository.format_revision): Added. Use the first 8 characters of the hash.

2016-05-17  Ryosuke Niwa  <rniwa@webkit.org>

        Add a subtitle under platform name in the summary page
        https://bugs.webkit.org/show_bug.cgi?id=157809

        Reviewed by Chris Dumez.

        Add a description beneath the platform names.

        * public/v3/pages/summary-page.js:
        (SummaryPage.prototype._constructTable): Add a br and a span if subtitle is present.
        (SummaryPage.cssTemplate): Added CSS rules for .subtitle.

2016-05-13  Ryosuke Niwa  <rniwa@webkit.org>

        v3 UI shows full git hash instead of the first 8 characters for a blame range
        https://bugs.webkit.org/show_bug.cgi?id=157691

        Reviewed by Stephanie Lewis.

        Fixed the bug that v3 UI shows the full 40 character git hash instead of the first 8 character as done in v2 UI.

        * public/v3/models/commit-log.js:
        (CommitLog.prototype.diff): Fixed the bug.
        * tools/run-tests.py:
        (main): Add the support for running a subset of tests as mocha does.
        * unit-tests/commit-log-tests.js: Added.

2016-05-13  Ryosuke Niwa  <rniwa@webkit.org>

        Unreviewed. Added the missing executable bits.

        * tools/bundle-v3-scripts.py: Added property svn:executable.
        * tools/detect-changes.js: Added property svn:executable.
        * tools/process-maintenance-backlog.py: Added property svn:executable.

2016-05-13  Ryosuke Niwa  <rniwa@webkit.org>

        Summary page doesn't report some missing platforms
        https://bugs.webkit.org/show_bug.cgi?id=157670

        Reviewed by Darin Adler.

        This patch improves the warning text for missing platforms and fixes the bug that platforms that don't have
        any data reported for a given test would not be reported as missing.

        * public/v3/pages/summary-page.js:
        (SummaryPage.prototype.render): Added instrumentations.
        (SummaryPage.prototype._constructRatioGraph): Always create both the ratio bar graph and the spinner icon.
        (SummaryPage.prototype._renderCell): Extracted from _constructRatioGraph. Toggle the displayed-ness of the
        spinner and the ratio bar graph in the cell by CSS for better performance.
        (SummaryPage.prototype._warningTextForGroup): Extracted from _constructRatioGraph. Rephrased warning text
        for clarity and adopted new API of SummaryPageConfigurationGroup.
        (SummaryPage.prototype._warningTextForGroup.mapAndSortByName): Added.
        (SummaryPage.prototype._warningTextForGroup.pluralizeIfNeeded): Added.
        (SummaryPage.cssTemplate): Added rules to toggle the visibility of spinner icons and bar graphs.
        (SummaryPageConfigurationGroup): Replaced this._warnings by more explicitly named this._missingPlatforms
        and this._platformsWithoutBaseline. Also add a platform to this._missingPlatforms if it didn't appear in
        any metrics. Note that adding a platform whenever it doesn't in any one metric would be incorrect since
        some tests uses a different test name on different platforms: e.g. PLT-Mac and PLT-iPhone.
        (SummaryPageConfigurationGroup.prototype.missingPlatforms): Added.
        (SummaryPageConfigurationGroup.prototype.platformsWithoutBaseline): Added.
        (SummaryPageConfigurationGroup.prototype._fetchAndComputeRatio):

2016-05-13  Ryosuke Niwa  <rniwa@webkit.org>

        Always use v3 UI for dashboards and analysis task pages
        https://bugs.webkit.org/show_bug.cgi?id=157647

        Reviewed by Darin Adler.

        Redirect dashboard pages from v1 and v2 to v3's summary page. Also redirect v1 UI's charts page and v2 UI's
        analysis task pages to the corresponding v3 pages.

        Keep v2's charts page accessible since some features such as segmentation is still only available on v2 UI.

        * public/index.html:
        (init.showCharts): Redirect to v3 UI once the chart list has been parsed.
        (init.redirectChartsToV3): Added.
        * public/v2/index.html:

2016-05-13  Ryosuke Niwa  <rniwa@webkit.org>

        Show a spinner while fetching data on summary page
        https://bugs.webkit.org/show_bug.cgi?id=157658

        Reviewed by Darin Adler.

        Show a spinner while fetching JSON files on the summary page.

        * public/v3/components/base.js:
        (ComponentBase.prototype.renderReplace): Added a new implementation that simply calls the static version.
        (ComponentBase.renderReplace): Made this static.

        * public/v3/pages/summary-page.js:
        (SummaryPage.prototype._constructRatioGraph): Show a spinner icon when SummaryPageConfigurationGroup's
        isFetching returns true.
        (SummaryPage.cssTemplate): Force the height of each cell to be 2.5rem so that the height of cell doesn't
        change when a spinner is replaced by a ratio bar graph.

        (SummaryPageConfigurationGroup): Added this._isFetching as an instance variable.
        (SummaryPageConfigurationGroup.prototype.isFetching): Added.
        (SummaryPageConfigurationGroup.prototype.fetchAndComputeSummary): Set this._isFetching while waiting for
        the promises to resolve after 50ms. We don't immediately set this._isFetching to avoid FOC when all JSON
        files have been cached.

2016-05-07  Ryosuke Niwa  <rniwa@webkit.org>

        Add horizontal between categories of tests
        https://bugs.webkit.org/show_bug.cgi?id=157386

        Reviewed by Darin Adler.

        Wrap tests in each category by tbody and add a horizontal bar between each category.

        * public/v3/pages/summary-page.js:
        (SummaryPage.prototype._constructTable):
        (SummaryPage.cssTemplate):

2016-05-04  Dewei Zhu  <dewei_zhu@apple.com>

        Summary page should show warnings when current or baseline data is missing.
        https://bugs.webkit.org/show_bug.cgi?id=157339

        Reviewed by Ryosuke Niwa.

        Set summary page to be the default page of v3 UI.
        Show warning icon when either baseline or current data is missing.
        Make fetchBetween returns a promise.
        Update unit tests for MeasurementSet.fetchBetween since it returns a promise now.
        Add a workaround to skip some platform and metric configurations.

        * public/v3/components/ratio-bar-graph.js:
        (RatioBarGraph):
        (RatioBarGraph.prototype.update): Add showWarningIcon flag to indicate whether we should show warning icon.
        (RatioBarGraph.prototype.render): Show warning icon when showWarningIcon is true.
        (RatioBarGraph.cssTemplate): Add style for warning icon.
        * public/v3/components/warning-icon.js: Add warning icon.
        (WarningIcon):
        (WarningIcon.cssTemplate):
        * public/v3/index.html:
        * public/v3/main.js:
        (main): Set summary page to be the default page of v3 UI.
        * public/v3/models/measurement-set.js:
        (MeasurementSet):
        (MeasurementSet.prototype.fetchBetween): Returns a promise. Fix the bug in previous implementation that we miss
        some callbacks sometimes. Basically, we will fetch primary cluster first, then secondary clusters. For each
        secondary cluster fetch, we will always invoke callback even when it fails.
        (MeasurementSet.prototype._fetchSecondaryClusters): Deleted.
        (MeasurementSet.prototype._fetch.else.url.api.measurement.set platform): Deleted.
        * public/v3/pages/summary-page.js:
        (SummaryPage): Add a variable for excluded configurations.
        (SummaryPage.prototype._createConfigurationGroup): Pass excluded configurations while building config groups.
        (SummaryPage.prototype._constructTable): Remove the logic for unified header since it breaks consistency of the table appearance.
        (SummaryPage.prototype.this._renderQueue.push): Show warning message when baseline/current data is missing.
        (SummaryPageConfigurationGroup): Add a variable to keep track of the warnings while computing summary.
        (SummaryPageConfigurationGroup.prototype.warnings): A getter for warnings.
        (SummaryPageConfigurationGroup._computeSummary): Fix a bug in calculating ratios. We should always use
        current/baseline for ratio and present the difference between ratio and 1 in the summary page.
        (SummaryPageConfigurationGroup.set then): Deleted.
        (SummaryPageConfigurationGroup.set var): Deleted.
        * unit-tests/measurement-set-tests.js: Add a helper function to wait for fetchBetween. Update unit tests since fetchBetween returns a promise now.
        (promise.set fetchBetween):
        (set MeasurementSet):
        (set fetchBetween): Deleted.

2016-04-26  Ryosuke Niwa  <rniwa@webkit.org>

        Chart status should always be computed against prior values
        https://bugs.webkit.org/show_bug.cgi?id=157014

        Reviewed by Darin Adler.

        Compare the current value against the last baseline or target value that appear before the current value in time
        so that the comparison stay the same even when new baseline and target values are reported. Also include the compared
        baseline or target value in the label for clarity.

        * public/v3/components/chart-status-view.js:
        (ChartStatusView.prototype._computeChartStatus):
        (ChartStatusView.prototype._computeChartStatus.labelForDiff):
        (ChartStatusView.prototype._findLastPointPriorToTime): Extracted from _relativeDifferenceToLaterPointInTimeSeries.
        Now finds the last point before the current point's time if there is any, or the last point in baseline / target.
        (ChartStatusView.prototype._relativeDifferenceToLaterPointInTimeSeries): Deleted.
        * public/v3/models/metric.js:
        (Metric.prototype.makeFormatter): Don't use SI units for unit-less metrics.

2016-04-13  Ryosuke Niwa  <rniwa@webkit.org>

        REGRESSION(r199444): Perf dashboard always fetches all measurement sets
        https://bugs.webkit.org/show_bug.cgi?id=156534

        Reviewed by Darin Adler.

        The bug was cased by SummaryPage's constructor fetching all measurement sets. Since each page is always
        constructed in main(), this resulted in all measurement sets being fetched on all pages.

        * public/v3/pages/summary-page.js:
        (SummaryPage):
        (SummaryPage.prototype.open): Fetch measurement set JSONs here.
        (SummaryPage.prototype._createConfigurationGroup): Renamed from _createConfigurationGroupAndStartFetchingData.

2016-04-12  Ryosuke Niwa  <rniwa@webkit.org>

        Add a summary page to v3 UI
        https://bugs.webkit.org/show_bug.cgi?id=156531

        Reviewed by Stephanie Lewis.

        Add new "Summary" page, which shows the average difference (better or worse) from the baseline across
        multiple platforms and tests by a single number.

        * public/include/manifest.php:
        (ManifestGenerator::generate): Include "summary" in manifest.json.
        * public/shared/statistics.js:
        (Statistics.mean): Added.
        (Statistics.median): Added.
        * public/v3/components/ratio-bar-graph.js: Added.
        (RatioBarGraph): Shows a horizontal bar graph that visualizes the relative difference (e.g. 3% better).
        (RatioBarGraph.prototype.update):
        (RatioBarGraph.prototype.render):
        (RatioBarGraph.cssTemplate):
        (RatioBarGraph.htmlTemplate):
        * public/v3/index.html:
        * public/v3/main.js:
        (main): Instantiate SummaryPage and add it to the navigation bar and the router.
        * public/v3/models/manifest.js:
        (Manifest._didFetchManifest): Let "summary" pass through from manifest.json to main().
        * public/v3/models/measurement-set.js:
        (MeasurementSet.prototype._failedToFetchJSON): Invoke the callback with an error or true in order for
        the callback can detect a failure.
        (MeasurementSet.prototype._invokeCallbacks): Ditto.
        * public/v3/pages/charts-page.js:
        (ChartsPage.createStateForConfigurationList): Added to add a hyperlink from summary page to charts page.
        * public/v3/pages/summary-page.js: Added.
        (SummaryPage): Added.
        (SummaryPage.prototype.routeName): Added.
        (SummaryPage.prototype.open): Added.
        (SummaryPage.prototype.render): Added.
        (SummaryPage.prototype._createConfigurationGroupAndStartFetchingData): Added.
        (SummaryPage.prototype._constructTable): Added.
        (SummaryPage.prototype._constructRatioGraph): Added.
        (SummaryPage.htmlTemplate): Added.
        (SummaryPage.cssTemplate): Added.
        (SummaryPageConfigurationGroup): Added. Represents a set of platforms and tests shown in a single cell.
        (SummaryPageConfigurationGroup.prototype.ratio): Added.
        (SummaryPageConfigurationGroup.prototype.label): Added.
        (SummaryPageConfigurationGroup.prototype.changeType): Added.
        (SummaryPageConfigurationGroup.prototype.configurationList): Added.
        (SummaryPageConfigurationGroup.prototype.fetchAndComputeSummary): Added.
        (SummaryPageConfigurationGroup.prototype._computeSummary): Added.
        (SummaryPageConfigurationGroup.prototype._fetchAndComputeRatio): Added. Invoked for each time series in
        the set, and stores the computed ratio of the current values to the baseline in this._setToRatio.
        The results are aggregated by _computeSummary as a single number later.
        (SummaryPageConfigurationGroup._medianForTimeRange): Added.
        (SummaryPageConfigurationGroup._fetchData): A thin wrapper to make MeasurementSet.fetchBetween promise
        friendly since MeasurementSet doesn't support Promise at the moment (but it should!).
        * server-tests/api-manifest.js: Updated a test case.

2016-04-12  Ryosuke Niwa  <rniwa@webkit.org>

        Make sync-buildbot.js fault safe
        https://bugs.webkit.org/show_bug.cgi?id=156498

        Reviewed by Chris Dumez.

        Fixed a bug that sync-buildbot.js will continue to schedule build requests from multiple test groups
        if multiple test groups are simultaneously in-progress on the same builder. Also fixed a bug that if
        a build request had failed without leaving a trace (i.e. no entry on any of the builders we know of),
        sync-buildbot.js throws an exception.

        * server-tests/tools-buildbot-triggerable-tests.js: Added test cases.
        * tools/js/buildbot-syncer.js:
        (BuildbotSyncer.prototype.scheduleRequestInGroupIfAvailable): Renamed. Optionally takes the slave name.
        When this parameter is specified, schedule the request only if the specified slave is available.
        * tools/js/buildbot-triggerable.js:
        (BuildbotTriggerable.prototype._scheduleNextRequestInGroupIfSlaveIsAvailable): Always use
        scheduleRequestInGroupIfAvailable to schedule a new build request. Using scheduleRequest for non-first
        build requests was problematic when there were multiple test groups with pending requests because then
        we would schedule those pending requests without checking whether there is already a pending job or if
        we have previously scheduled a job. Also fallback to use any syncer / builder when groupInfo.syncer is
        not set even if the next request was not the first one in the test group since we can't determine on
        which builder preceding requests are processed in such cases.
        * unit-tests/buildbot-syncer-tests.js:

2016-04-11  Ryosuke Niwa  <rniwa@webkit.org>

        Replace script runner to use mocha.js tests
        https://bugs.webkit.org/show_bug.cgi?id=156490

        Reviewed by Chris Dumez.

        Replaced run-tests.js, which was a whole test harness for running legacy tests by tools/run-tests.py
        which is a thin wrapper around mocha.js.

        * run-tests.js: Removed.
        * tests: Removed.
        * tools/run-tests.py: Added.
        (main):

2016-04-11  Ryosuke Niwa  <rniwa@webkit.org>

        New syncing script sometimes schedules a build request on a wrong builder
        https://bugs.webkit.org/show_bug.cgi?id=156489

        Reviewed by Stephanie Lewis.

        The bug was caused by _scheduleNextRequestInGroupIfSlaveIsAvailable scheduling the next build request on
        any available syncer regardless of whether the request is the first one in the test group or not because
        BuildRequest.order was returning a string instead of a number.

        Also fixed a bug that BuildbotTriggerable.syncOnce was re-ordering test groups by their id's instead of
        respecting the order in which the perf dashboard returned.

        * public/v3/models/build-request.js:
        (BuildRequest.prototype.order): Force the order to be a number.
        * server-tests/api-build-requests-tests.js: Assert the order as numbers.
        * server-tests/resources/mock-data.js:
        (MockData.addAnotherMockTestGroup): Changed the test group id to 601, which is after the first mock data.
        The old number was masking a bug in BuildbotTriggerable that it was re-ordering test groups by their id's
        instead of using the order set forth by the perf dashboard.
        (MockData.mockTestSyncConfigWithSingleBuilder):
        * server-tests/tools-buildbot-triggerable-tests.js: Added a test case for scheduling two build requests in
        a single call to syncOnce. Each build request should be scheduled on the same builder as the previous build
        requests in the same test group.
        * tools/js/buildbot-triggerable.js:
        (BuildbotTriggerable.prototype.syncOnce): Order test groups by groupOrder, which is the index at which first
        build request in the group appeared.
        (BuildbotTriggerable.prototype._scheduleNextRequestInGroupIfSlaveIsAvailable): Don't re-order build requests
        as they're already sorted on the server side.
        (BuildbotTriggerable._testGroupMapForBuildRequests): Added groupOrder to test group info 

2016-04-09  Ryosuke Niwa  <rniwa@webkit.org>

        Build fix. Don't treat a build number 0 as a pending build.

        * tools/js/buildbot-syncer.js:
        (BuildbotBuildEntry.prototype.isPending):

2016-04-08  Ryosuke Niwa  <rniwa@webkit.org>

        Escape builder names in url* and pathFor* methods of BuildbotSyncer
        https://bugs.webkit.org/show_bug.cgi?id=156427

        Reviewed by Darin Adler.

        The build fix in r199251 breaks other usage of RemoteAPI. Fix it properly by escaping builder names in
        various methods of BuildbotSyncer.

        Also fixed a typo in the logging and a bug that the new syncing script never updated "scheduled" to "running".

        * server-tests/resources/mock-data.js:
        (MockData.mockTestSyncConfigWithTwoBuilders): Renamed "some-builder-2" to "some builder 2" to test the
        new escaping behavior in tools-buildbot-triggerable-tests.js and buildbot-syncer-tests.js.

        * server-tests/tools-buildbot-triggerable-tests.js: Added tests for status url, and added a new test case
        for updating "scheduled" to "running".

        * tools/js/buildbot-syncer.js:
        (BuildbotBuildEntry.buildRequestStatusIfUpdateIsNeeded): Update the status to "running" when the request's
        status is "scheduled" and the buildbot's build is currently in progress.
        (BuildbotSyncer.prototype.pathForPendingBuildsJSON): Escape the builder name.
        (BuildbotSyncer.prototype.pathForBuildJSON): Ditto.
        (BuildbotSyncer.prototype.pathForForceBuild): Ditto.
        (BuildbotSyncer.prototype.url): Ditto.
        (BuildbotSyncer.prototype.urlForBuildNumber): Ditto.

        * tools/js/buildbot-triggerable.js:
        (BuildbotTriggerable.prototype._pullBuildbotOnAllSyncers):
        (BuildbotTriggerable.prototype._scheduleNextRequestInGroupIfSlaveIsAvailable): Fixed a typo. We are
        scheduling new build requests, not syncing them.
        * tools/js/remote.js:
        (RemoteAPI.sendHttpRequest): Reverted r199251.
        * unit-tests/buildbot-syncer-tests.js:

2016-04-08  Ryosuke Niwa  <rniwa@webkit.org>

        Build fix. We need to escape the path or http.request would fail.

        * tools/js/remote.js:

2016-04-08  Ryosuke Niwa  <rniwa@webkit.org>

        Fix various bugs in the new syncing script
        https://bugs.webkit.org/show_bug.cgi?id=156393

        Reviewed by Darin Adler.

        * server-tests/resources/common-operations.js: Added. This file was supposed to be added in r199191.
        (addBuilderForReport):
        (addSlaveForReport):
        (connectToDatabaseInEveryTest):
        (submitReport):
        * tools/js/buildbot-triggerable.js:
        (BuildbotTriggerable.prototype._pullBuildbotOnAllSyncers): Don't log every time we pull from buildbot
        builder as this dramatically increases the amount of log we generate.
        * tools/js/parse-arguments.js:
        (parseArguments): Fixed a typo. This should be parseArgument*s*, not parseArgument.
        * tools/js/remote.js:
        (RemoteAPI.prototype.url): Fixed a bug that portSuffix wasn't being expanded in the template literal.
        (RemoteAPI.prototype.configure): Added more validations with nice error messages.
        (RemoteAPI.prototype.sendHttpRequest): Falling back to port 80 isn't right when scheme is https. Compute
        the right port in configure instead based on the scheme.
        * tools/sync-buildbot.js:
        (syncLoop): Fixed the bug that syncing multiple times fail because Manifest.fetch() create new Platform
        and Test objects. This results in various references in BuildRequest objects to get outdated. Fixing this
        properly in Manifest.fetch() because we do need to "forget" about some tests and platforms in some cases.
        For now, delete all v3 model objects and start over in each syncing cycle.
        * unit-tests/tools-js-remote-tests.js: Added. Unit tests for the aforementioned changes to RemoteAPI.

2016-04-07  Ryosuke Niwa  <rniwa@webkit.org>

        sync-buildbot.js doesn't mark disappeared builds as failed
        https://bugs.webkit.org/show_bug.cgi?id=156386

        Reviewed by Chris Dumez.

        Fix a bug that new syncing script doesn't mark builds that it scheduled but doesn't appear when queried
        by buildbot's JSON API. These are builds that got canceled by humans (e.g. buildbot was restarted, data
        loss, pending build was canceled, etc...)

        * server-tests/tools-buildbot-triggerable-tests.js: Added a test case.
        * tools/js/buildbot-triggerable.js:
        (BuildbotTriggerable.prototype._pullBuildbotOnAllSyncers): Added a set of build requests we've matched
        against BuildbotBuildEntry's. Mark build requests that didn't have any entry but supposed to be in either
        'scheduled' or 'running' status as failed.

2016-04-07  Ryosuke Niwa  <rniwa@webkit.org>

        A/B testing bots should prioritize user created test groups
        https://bugs.webkit.org/show_bug.cgi?id=156375

        Reviewed by Chris Dumez.

        Order build requests preferring user created ones over ones automatically created by detect-changes.js.

        Also fixed a bug in BuildbotSyncer.scheduleFirstRequestInGroupIfAvailable that it was scheduling a new
        build request on a builder/slave even when we had previously scheduled another build request.

        * public/include/build-requests-fetcher.php:
        (BuildRequestsFetcher::fetch_incomplete_requests_for_triggerable): Order build requested based on
        author_order which is 0 when it's created by an user and 1 when it's created by detect-changes.js.
        Since we're using ascending order, this would put user created test groups first.
        * server-tests/api-build-requests-tests.js: Updated an existing test case and added a new test case
        for testing that build requests for an user created test group shows up first.
        * server-tests/resources/mock-data.js:
        (MockData.addAnotherMockTestGroup): Takes an extra argument to specify the author name.
        * server-tests/tools-buildbot-triggerable-tests.js: Added a test case for testing that build requests
        for an user created test group shows up first.
        * tools/js/buildbot-syncer.js:
        (BuildbotSyncer): Added _slavesWithNewRequests to keep track of build slaves on which we have already
        scheduled new build requests. Don't schedule more requests on these slaves.
        (BuildbotSyncer.prototype.scheduleRequest):
        (BuildbotSyncer.prototype.scheduleFirstRequestInGroupIfAvailable): Add the specified slave name (or null
        when slaveList is not specified) to _slavesWithNewRequests.
        (BuildbotSyncer.prototype.pullBuildbot): Clear the set after pulling buildbot since any build request
        we have previously scheduled should be included in one of the entires now.
        * unit-tests/buildbot-syncer-tests.js: Added test cases for the aforementioned bug.
        (sampleiOSConfig): Added a second slave for new test cases.

2016-04-07  Ryosuke Niwa  <rniwa@webkit.org>

        Migrate legacy perf dashboard tests to mocha.js based tests
        https://bugs.webkit.org/show_bug.cgi?id=156335

        Reviewed by Chris Dumez.

        Migrated all legacy run-tests.js tests to mocha.js based tests. Since the new harness uses Promise
        for most of asynchronous operations, refactored the tests to use Promises as well, and added more
        assertions where appropriate.

        Also consolidated common helper functions into server-tests/resources/common-operations.js.
        Unfortunately there were multiple inconsistent implementations of addBuilder/addSlave. Some were
        taking an array of reports while others were taking a single report. New shared implementation in
        common-operations.js now takes a single report.

        Also decreased the timeout in most tests from 10s to 1s so that tests fail early when they timeout.
        Most of tests are passing under 100ms on my computer so 1s should be plenty still.

        * run-tests.js: Removed.
        * server-tests/admin-platforms-tests.js: Moved from tests/admin-platforms.js.
        (reportsForDifferentPlatforms):
        * server-tests/admin-reprocess-report-tests.js: Moved from tests/admin-reprocess-report.js.
        (.addBuilder): Moved to common-operations.js.
        * server-tests/api-build-requests-tests.js:
        * server-tests/api-manifest.js: Use MockData.resetV3Models() instead of manually clearing maps.
        * server-tests/api-measurement-set-tests.js: Moved from tests/api-measurement-set.js.
        (.queryPlatformAndMetric):
        (.format):
        * server-tests/api-report-commits-tests.js: Moved from tests/api-report-commits.js.
        * server-tests/api-report-tests.js: Moved from tests/api-report.js.
        (.emptyReport):
        (.emptySlaveReport):
        (.reportWithSameSubtestName):
        * server-tests/resources/common-operations.js: Added.
        (addBuilderForReport): Extracted from tests.
        (addSlaveForReport): Ditto.
        (connectToDatabaseInEveryTest): Added.
        (submitReport): Extracted from admin-platforms-tests.js.
        * server-tests/resources/test-server.js:
        (TestServer): Make TestServer a singleton since it doesn't make any sense for each module to start
        its own Apache instance (that would certainly will fail).
        * server-tests/tools-buildbot-triggerable-tests.js:
        * tests: Removed.
        * tools/js/database.js:
        (Database.prototype.selectAll): Added.
        (Database.prototype.selectFirstRow): Added.
        (Database.prototype.selectRows): Added. Dynamically construct a query string based on arguments.

2016-04-05  Ryosuke Niwa  <rniwa@webkit.org>

        New buildbot syncing scripts that supports multiple builders and slaves
        https://bugs.webkit.org/show_bug.cgi?id=156269

        Reviewed by Chris Dumez.

        Add sync-buildbot.js that supports scheduling A/B testing jobs on multiple builders and slaves.
        The old python script (sync-with-buildbot.py) could only support a single builder and slave
        for each platform, test pair.

        The main logic is implemented in BuildbotTriggerable.syncOnce. Various helper methods are added
        throughout the codebase and tests have been refactored.

        BuildbotSyncer has been updated to support multiple platform, test pairs. It's now responsible
        for syncing everything on each builder (on a buildbot).

        Added more unit tests for BuildbotSyncer and server tests for BuildbotTriggerable, and refactored
        test helpers and mocks as needed.

        * public/v3/models/build-request.js:
        (BuildRequest.prototype.status): Added.
        (BuildRequest.prototype.isScheduled): Added.
        * public/v3/models/metric.js:
        (Metric.prototype.fullName): Added.
        * public/v3/models/platform.js:
        (Platform): Added the map based on platform name.
        (Platform.findByName): Added.
        * public/v3/models/test.js:
        (Test.topLevelTests):
        (Test.findByPath): Added. Finds a test based on an array of test names; e.g. ['A', 'B'] would
        find the test whose name is "B" which has a parent test named "A".
        (Test.prototype.fullName): Added.
        * server-tests/api-build-requests-tests.js:
        (addMockData): Moved to resources/mock-data.js.
        (addAnotherMockTestGroup): Ditto.
        * server-tests/resources/mock-data.js: Added.
        (MockData.resetV3Models): Added.
        (MockData.addMockData): Moved from api-build-requests-tests.js.
        (MockData.addAnotherMockTestGroup): Ditto.
        (MockData.mockTestSyncConfigWithSingleBuilder): Added.
        (MockData.mockTestSyncConfigWithTwoBuilders): Added.
        (MockData.pendingBuild): Added.
        (MockData.runningBuild): Added.
        (MockData.finishedBuild): Added.
        * server-tests/resources/test-server.js:
        (TestServer):
        (TestServer.prototype.remoteAPI):
        (TestServer.prototype._ensureTestDatabase): Don't fail even if the test database doesn't exit.
        (TestServer.prototype._startApache): Create a RemoteAPI instance to access the test sever.
        (TestServer.prototype._waitForPid): Increase the timeout.
        (TestServer.prototype.inject): Replace global.RemoteAPI during the test and restore it afterwards.
        * server-tests/tools-buildbot-triggerable-tests.js: Added. Tests BuildbotTriggerable.syncOnce.
        (MockLogger): Added.
        (MockLogger.prototype.log): Added.
        (MockLogger.prototype.error): Added.
        * tools/detect-changes.js:
        (parseArgument): Moved to js/parse-arguments.js.
        * tools/js/buildbot-syncer.js:
        (BuildbotBuildEntry):
        (BuildbotBuildEntry.prototype.syncer): Added.
        (BuildbotBuildEntry.prototype.buildRequestStatusIfUpdateIsNeeded): Added. Returns a new status
        for a build request (of the matching build request ID) if it needs to be updated in the server.
        (BuildbotSyncer): This class 
        (BuildbotSyncer.prototype.addTestConfiguration): Added.
        (BuildbotSyncer.prototype.testConfigurations): Returns the list of test configurations.
        (BuildbotSyncer.prototype.matchesConfiguration): Returns true iff the request can be scheduled on
        this builder.
        (BuildbotSyncer.prototype.scheduleRequest): Added. Schedules a new job on buildbot for a request.
        (BuildbotSyncer.prototype.scheduleFirstRequestInGroupIfAvailable): Added. Schedules a new job for
        the specified build request on the first slave that's available.
        (BuildbotSyncer.prototype.pullBuildbot): Return a list of BuildbotBuildEntry instead of an object.
        Also store it on an instance variable so that scheduleFirstRequestInGroupIfAvailable could use it.
        (BuildbotSyncer.prototype._pullRecentBuilds):
        (BuildbotSyncer.prototype.pathForPendingBuildsJSON): Renamed from urlForPendingBuildsJSON and now
        only returns the path instead of the full URL since RemoteAPI takes a path, not full URL.
        (BuildbotSyncer.prototype.pathForBuildJSON): Ditto from pathForBuildJSON.
        (BuildbotSyncer.prototype.pathForForceBuild): Added.
        (BuildbotSyncer.prototype.url): Use RemoteAPI's url method instead of manually constructing URL.
        (BuildbotSyncer.prototype.urlForBuildNumber): Ditto.
        (BuildbotSyncer.prototype._propertiesForBuildRequest): Now that each syncer can have multiple test
        configurations associated with it, find the one matching for this request.
        (BuildbotSyncer._loadConfig): Create a syncer per builder and add all test configurations to it.
        (BuildbotSyncer._validateAndMergeConfig): Added the support for 'SlaveList', which is a list of
        slave names present on this builder.
        * tools/js/buildbot-triggerable.js: Added.
        (BuildbotTriggerable): Added.
        (BuildbotTriggerable.prototype.name): Added.
        (BuildbotTriggerable.prototype.syncOnce): Added. The main logic for the syncing script. It pulls
        existing build requests from the perf dashboard, pulls buildbot for pending and running/completed
        builds on each builder (represented by each syncer), schedules build requests on buildbot if there
        is any builder/slave available, and updates the status of build requests in the database.
        (BuildbotTriggerable.prototype._validateRequests): Added.
        (BuildbotTriggerable.prototype._pullBuildbotOnAllSyncers): Added.
        (BuildbotTriggerable.prototype._scheduleNextRequestInGroupIfSlaveIsAvailable): Added.
        (BuildbotTriggerable._testGroupMapForBuildRequests): Added.
        * tools/js/database.js:
        * tools/js/parse-arguments.js: Added. Extracted out of tools/detect-changes.js.
        (parseArguments):
        * tools/js/remote.js:
        (RemoteAPI): Now optionally takes the server configuration.
        (RemoteAPI.prototype.url): Added.
        (RemoteAPI.prototype.getJSON): Removed the code for specifying request content.
        (RemoteAPI.prototype.getJSONWithStatus): Ditto.
        (RemoteAPI.prototype.postJSON): Added.
        (RemoteAPI.prototype.postFormUrlencodedData): Added.
        (RemoteAPI.prototype.sendHttpRequest): Fixed the code to specify auth.
        * tools/js/v3-models.js: Don't include RemoteAPI here as they require a configuration for each host.
        * tools/sync-buildbot.js: Added.
        (main): Added. Parse the arguments and start the loop.
        (syncLoop): Added.
        * unit-tests/buildbot-syncer-tests.js: Added tests for pullBuildbot, scheduleRequest, as well as
        scheduleFirstRequestInGroupIfAvailable. Refactored helper functions as needed.
        (sampleiOSConfig):
        (smallConfiguration): Added.
        (smallPendingBuild): Added.
        (smallInProgressBuild): Added.
        (smallFinishedBuild): Added.
        (createSampleBuildRequest): Create a unique build request for each platform.
        (samplePendingBuild): Optionally specify build time and slave name.
        (sampleInProgressBuild): Optionally specify slave name.
        (sampleFinishedBuild): Ditto.
        * unit-tests/resources/mock-remote-api.js:
        (assert.notReached.assert.notReached):
        (MockRemoteAPI.url): Added.
        (MockRemoteAPI.postFormUrlencodedData): Added.
        (MockRemoteAPI._addRequest): Extracted from getJSONWithStatus.
        (MockRemoteAPI.waitForRequest): Extracted from inject. For tools-buildbot-triggerable-tests.js, we
        need to instantiate a RemoteAPI for buildbot without replacing global.RemoteAPI.
        (MockRemoteAPI.inject):
        (MockRemoteAPI.reset): Added.

2016-03-30  Ryosuke Niwa  <rniwa@webkit.org>

        Simplify API of Test model by removing Test.setParentTest
        https://bugs.webkit.org/show_bug.cgi?id=156055

        Reviewed by Joseph Pecoraro.

        Removed Test.setParentTest. Keep track of the child-parent relationship using the static map instead.

        Now each test only stores parent's id and uses the ID static map in Test.parentTest().

        * public/v3/models/manifest.js:
        (Manifest._didFetchManifest.buildObjectsFromIdMap): Removed the code to create the map of child-parent
        relationship and call setParentTest.
        * public/v3/models/test.js:
        (Test): Updated a static map by the name of "childTestMap" to store itself. We should probably sort
        child tests using some fixed criteria in the future instead of relying on the creation order but
        preserve the old code's ordering for now.
        (Test.prototype.parentTest): Look up the static map by the parent test's id.
        (Test.prototype.onlyContainsSingleMetric):
        (Test.prototype.setParentTest): Deleted.
        (Test.prototype.childTests): Look up the child test map.

2016-03-30  Ryosuke Niwa  <rniwa@webkit.org>

        BuildRequest should have associated platform and test
        https://bugs.webkit.org/show_bug.cgi?id=156054

        Reviewed by Joseph Pecoraro.

        Added methods to retrieve the platform and the test associated with a build request with tests.

        * public/v3/models/build-request.js:
        (BuildRequest):
        (BuildRequest.prototype.platform): Added.
        (BuildRequest.prototype.test): Added.
        * server-tests/api-build-requests-tests.js:
        * server-tests/api-manifest.js: Fixed a typo. This tests /api/manifest, not /api/build-requests.
        * unit-tests/buildbot-syncer-tests.js:
        (.createSampleBuildRequest): Now takes Platform and Test objects to avoid hitting assertions in
        BuildRequest's constructor.

2016-03-30  Ryosuke Niwa  <rniwa@webkit.org>

        BuildRequest should have a method to fetch all in-progress and pending requests for a triggerable
        https://bugs.webkit.org/show_bug.cgi?id=156008

        Reviewed by Darin Adler.

        Add a method to BuildRequest that fetches all pending and in-progress requests for a triggerable.

        Now, new syncing scripts must be able to figure out the build slave the first build requests in
        a given test group had used in order to schedule subsequent build requests in the test group.

        For this purpose, /api/build-requests has been modified to return all build requests whose test
        group had not finished yet. A test group is finished if all build requests in the test group had
        finished (completed, failed, or canceled).

        * public/include/build-requests-fetcher.php:
        (BuildRequestFetcher::fetch_incomplete_requests_for_triggerable): Return all build requests in test
        groups that have not been finished.
        * public/v3/models/build-request.js:
        (BuildRequest):
        (BuildRequest.prototype.testGroupId): Added.
        (BuildRequest.prototype.isPending): Renamed from hasPending to fix a bad grammar.
        (BuildRequest.fetchForTriggerable): Added.
        (BuildRequest.constructBuildRequestsFromData): Extracted from _createModelsFromFetchedTestGroups in
        TestGroup.
        * public/v3/models/manifest.js:
        (Manifest.fetch): Use the full path from root so that it works in server tests.
        * public/v3/models/test-group.js:
        (TestGroup.hasPending):
        (TestGroup._createModelsFromFetchedTestGroups):
        * server-tests/api-build-requests-tests.js: Added tests to ensure all build requests for a test group
        is present in the response returned by /api/build-requests iff any build request in the group had not
        finished yet.
        (.addMockData):
        (.addAnotherMockTestGroup): Added.
        * unit-tests/test-groups-tests.js:

2016-03-29  Ryosuke Niwa  <rniwa@webkit.org>

        Make dependency injection in unit tests more explicit
        https://bugs.webkit.org/show_bug.cgi?id=156006

        Reviewed by Joseph Pecoraro.

        Make the dependency injection of model objects in unit tests explicit so that server tests that create
        "real" model objects won't create these mock objects. Now each test that uses mock model objects would call
        MockModels.inject() to inject before / beforeEach and access each object using a property on MockModels
        instead of them being implicitly defined on the global object.

        Similarly, MockRemoteAPI now only replaces global.RemoteAPI during each test so that server tests can use
        real RemoteAPI to access the test Apache server.

        * unit-tests/analysis-task-tests.js:
        * unit-tests/buildbot-syncer-tests.js:
        (createSampleBuildRequest):
        * unit-tests/measurement-adaptor-tests.js:
        * unit-tests/measurement-set-tests.js:
        * unit-tests/resources/mock-remote-api.js:
        (MockRemoteAPI.getJSONWithStatus):
        (MockRemoteAPI.inject): Added. Override RemoteAPI on the global object during each test.
        * unit-tests/resources/mock-v3-models.js:
        (MockModels.inject): Added. Create mock model objects before each test, and clear all static maps of
        various v3 model classes (to remove all singleton objects for those model classes).
        * unit-tests/test-groups-tests.js:

2016-03-29  Ryosuke Niwa  <rniwa@webkit.org>

        BuildbotSyncer should be able to fetch JSON from buildbot
        https://bugs.webkit.org/show_bug.cgi?id=155921

        Reviewed by Joseph Pecoraro.

        Added BuildbotSyncer.pullBuildbot which fetches pending, in-progress, and finished builds from buildbot
        with lots of unit tests as this has historically been a source of subtle bugs in the old script.

        New implementation fixes a subtle bug in the old pythons script which overlooked the possibility that
        the state of some builds may change between each HTTP request. In the old script, we fetched the list
        of the pending builds, and requested -1, -2, etc... builds for N times. But between each request,
        a pending build may start running or an in-progress build finish and shift the offset by one. The new
        script avoids this problem by first requesting all pending builds, then all in-progress and finished
        builds in a single HTTP request. The results are then merged so that entries for in-progress and
        finished builds would override the entries for pending builds if they overlap.

        Also renamed RemoteAPI.fetchJSON to RemoteAPI.getJSON to match v3 UI's RemoteAPI. This change makes
        the class interchangeable between frontend (public/v3/remote.js) and backend (tools/js/remote.js).

        * server-tests/api-build-requests-tests.js:
        * server-tests/api-manifest.js:
        * tools/js/buildbot-syncer.js:
        (BuildbotBuildEntry): Removed the unused argument "type". Store the syncer as an instance variable as
        we'd need to query for the buildbot URL. Also fixed a bug that _isInProgress was true for finished
        builds as 'currentStep' is always defined but null in those builds.
        (BuildbotBuildEntry.prototype.buildNumber): Added.
        (BuildbotBuildEntry.prototype.isPending): Added.
        (BuildbotBuildEntry.prototype.hasFinished): Added.
        (BuildbotSyncer.prototype.pullBuildbot): Added. Fetches pending builds first and then finished builds.
        (BuildbotSyncer.prototype._pullRecentBuilds): Added. Fetches in-progress and finished builds.
        (BuildbotSyncer.prototype.urlForPendingBuildsJSON): Added.
        (BuildbotSyncer.prototype.urlForBuildJSON): Added.
        (BuildbotSyncer.prototype.url): Added.
        (BuildbotSyncer.prototype.urlForBuildNumber): Added.
        * tools/js/remote.js:
        (RemoteAPI.prototype.getJSON): Renamed from fetchJSON.
        (RemoteAPI.prototype.getJSONWithStatus): Renamed from fetchJSONWithStatus.
        * tools/js/v3-models.js: Load tools/js/remote.js instead of public/v3/remote.js inside node.
        * unit-tests/buildbot-syncer-tests.js: Added a lot of unit tests for BuildbotSyncer.pullBuildbot
        (samplePendingBuild):
        (sampleInProgressBuild): Added.
        (sampleFinishedBuild): Added.
        * unit-tests/resources/mock-remote-api.js:
        (global.RemoteAPI.getJSON): Use the same mock as getJSONWithStatus.

2016-03-24  Ryosuke Niwa  <rniwa@webkit.org>

        Migrate admin-regenerate-manifest.js to mocha.js and test v3 UI code
        https://bugs.webkit.org/show_bug.cgi?id=155863

        Reviewed by Joseph Pecoraro.

        Replaced admin-regenerate-manifest.js by a new mocha.js tests using the new server testing capability
        added in r198642 and tested v3 UI code (parsing manifest.json and creating models). Also removed
        /admin/regenerate-manifest since it has been superseded by /api/manifest.

        This patch also extracts manifest.js out of main.js so that it could be used and tested without the
        DOM support in node.

        * public/admin/regenerate-manifest.php: Deleted.
        * public/include/db.php: Fixed a regression from r198642 since CONFIG_DIR now doesn't end with
        a trailing backslash.
        * public/include/manifest.php:
        (ManifestGenerator::bug_trackers): Avoid a warning message when there are no repositories.
        * public/v3/index.html:
        * public/v3/main.js:
        (main):
        * public/v3/models/bug-tracker.js:
        (BugTracker.prototype.newBugUrl): Added.
        (BugTracker.prototype.repositories): Added.
        * public/v3/models/manifest.js: Added. Extracted from main.js.
        (Manifest.fetch): Moved from main.js' fetchManifest.
        (Manifest._didFetchManifest): Moved from main.js' didFetchManifest.
        * public/v3/models/platform.js:
        (Platform.prototype.hasTest): Fixed the bug that "test" here was shadowing the function parameter of
        the same name. This is tested by the newly added test cases.
        * server-tests/api-build-requests-tests.js:
        * server-tests/api-manifest.js: Added. Migrated test cases from tests/admin-regenerate-manifest.js
        with additional assertions for v3 UI model objects.
        * server-tests/resources/test-server.js:
        (TestServer.prototype.start):
        (TestServer.prototype.testConfig): Renamed from _constructTestConfig now that this is a public API.
        Also no longer takes dataDirectory as an argument since it's always the same.
        (TestServer.prototype._ensureDataDirectory): Fixed a bug that we weren't making public/data.
        (TestServer.prototype.cleanDataDirectory): Added. Remove all files inside public/data between tests.
        (TestServer.prototype.inject): Added. Calls before, etc... because always calling before had an
        unintended side effect of slowing down unit tests even through they don't need Postgres or Apache.
        * tests/admin-regenerate-manifest.js: Removed.
        * tools/js/database.js:
        * tools/js/v3-models.js:

2016-03-23  Ryosuke Niwa  <rniwa@webkit.org>

        Add mocha server tests for /api/build-requests
        https://bugs.webkit.org/show_bug.cgi?id=155831

        Reviewed by Chris Dumez.

        Added the new mocha.js based server-tests for /api/build-requests. The new harness automatically:
         - starts a new Apache instance
         - switches the database during testing via setting an environmental variable 
         - backups and restores public/data directory during testing

        As a result, developer no longer has to manually setup Apache, edit config.json manually to use
        a testing database, or run /api/manifest.php to re-generate the manifest file after testing.

        This patch also makes ID resolution optional on /api/build-requests so that v3 model based syncing
        scripts can re-use the same code as the v3 UI to process the JSON. tools/sync-with-buildbot.py has
        been modified to use this option (useLegacyIdResolution).

        * config.json: Added configurations for the test httpd server.
        * init-database.sql: Don't error when tables and types don't exist (when database is empty).
        * public/api/build-requests.php:
        (main): Made the ID resolution optional with useLegacyIdResolution. Also removed "updates" from the
        results JSON since it's never used.
        * public/include/build-requests-fetcher.php:
        (BuildRequestsFetcher::__construct):
        (BuildRequestsFetcher::fetch_roots_for_set_if_needed): Fixed the bug that we would include the same
        commit multiple times for each root set.
        * public/include/db.php:
        (config): If present, use ORG_WEBKIT_PERF_CONFIG_PATH instead of Websites/perf.webkit.org/config.json.
        * server-tests: Added.
        * server-tests/api-build-requests-tests.js: Added. Tests for /api/build-requests.
        (.addMockData):
        * server-tests/resources: Added.
        * server-tests/resources/test-server.conf: Added. Apache configuration file for testing.
        * server-tests/resources/test-server.js: Added.
        (TestSever): Added.
        (TestSever.prototype.start): Added.
        (TestSever.prototype.stop): Added.
        (TestSever.prototype.remoteAPI): Added. Configures RemoteAPI to be used with the test sever.
        (TestSever.prototype.database): Added. Returns Database configured to use the test database.
        (TestSever.prototype._constructTestConfig): Creates config.json for testing. The file is generated by
        _start and db.php's config() reads it from the environmental variable: ORG_WEBKIT_PERF_CONFIG_PATH.
        (TestSever.prototype._ensureDataDirectory): Renames public/data to public/original-data if exists,
        and creates a new empty public/data.
        (TestSever.prototype._restoreDataDirectory): Deletes public/data and renames public/original-data
        back to public/data.
        (TestSever.prototype._ensureTestDatabase): Drops the test database if exists and creates a new one.
        (TestSever.prototype.initDatabase): Run init-database.sql to start each test with a consistent state.
        (TestSever.prototype._executePgsqlCommand): Executes a postgres command line tool such as psql.
        (TestSever.prototype._determinePgsqlDirectory): Finds the directory that contains psql.
        (TestSever.prototype._startApache): Starts an Apache instance for testing.
        (TestSever.prototype._stopApache): Stops the Apache instance for testing.
        (TestSever.prototype._waitForPid): Waits for the Apache pid file to appear or disappear.
        (before): Start the test server at the beginning.
        (beforeEach): Re-initialize all tables before each test.
        (after): Stop the test server at the end.
        * tools/js/config.js:
        (Config.prototype.path):
        (Config.prototype.serverRoot): Added. The path to Websites/perf.webkit.org/public/.
        (Config.prototype.pathFromRoot): Added. Resolves a path from Websites/perf.webkit.org.
        * tools/js/database.js:
        (Database): Now optionally takes the database name to use a different database during testing.
        (Database.prototype.connect):
        (Database.prototype.query): Added.
        (Database.prototype.insert): Added.
        (tableToPrefixMap): Maps table name to its prefix. Used by Database.insert.
        * tools/js/remote.js: Added.
        (RemoteAPI): Added. This is node.js equivalent of RemoteAPI in public/v3/remote.js.
        (RemoteAPI.prototype.configure): Added.
        (RemoteAPI.prototype.fetchJSON): Added.
        (RemoteAPI.prototype.fetchJSONWithStatus): Added.
        (RemoteAPI.prototype.sendHttpRequest): Added.
        * tools/sync-with-buildbot.py:
        (main): Use useLegacyIdResolution as this script relies on the legacy behavior.
        * unit-tests/checkconfig.js: pg was never directly used in this test.

2016-03-23  Ryosuke Niwa  <rniwa@webkit.org>

        Delete a file that was supposed to be removed in r198614 for real.

        * unit-tests/resources/v3-models.js: Removed.

2016-03-23  Ryosuke Niwa  <rniwa@webkit.org>

        Add a model for parsing buildbot JSON with unit tests
        https://bugs.webkit.org/show_bug.cgi?id=155814

        Reviewed by Joseph Pecoraro.

        Added BuildbotSyncer and BuildbotBuildEntry classes to parse buildbot JSON files with unit tests.
        They will be used in the new syncing scripts to improve A/B testing.

        * public/v3/models/build-request.js:
        (BuildRequest):
        * tools/js/buildbot-syncer.js: Added.
        (BuildbotBuildEntry): Added.
        (BuildbotBuildEntry.prototype.slaveName): Added.
        (BuildbotBuildEntry.prototype.buildRequestId): Added.
        (BuildbotBuildEntry.prototype.isInProgress): Added.
        (BuildbotSyncer): Added.
        (BuildbotSyncer.prototype.testPath): Added.
        (BuildbotSyncer.prototype.builderName): Added.
        (BuildbotSyncer.prototype.platformName): Added.
        (BuildbotSyncer.prototype.fetchPendingRequests): Added.
        (BuildbotSyncer.prototype._propertiesForBuildRequest): Added.
        (BuildbotSyncer.prototype._revisionSetFromRootSetWithExclusionList): Added.
        (BuildbotSyncer._loadConfig): Added.
        (BuildbotSyncer._validateAndMergeConfig): Added.
        (BuildbotSyncer._validateAndMergeProperties): Added.
        * tools/js/v3-models.js: Copied from unit-tests/resources/v3-models.js.
        (beforeEach): Deleted since this only defined inside mocha.
        * unit-tests/analysis-task-tests.js:
        * unit-tests/buildbot-syncer-tests.js: Added.
        (sampleiOSConfig):
        (createSampleBuildRequest):
        (.smallConfiguration):
        * unit-tests/measurement-adaptor-tests.js:
        * unit-tests/measurement-set-tests.js:
        * unit-tests/resources/mock-v3-models.js: Renamed from unit-tests/resources/v3-models.js.
        (beforeEach):
        * unit-tests/test-groups-tests.js:
        (sampleTestGroup):

2016-03-22  Ryosuke Niwa  <rniwa@webkit.org>

        Add unit tests for test-group.js
        https://bugs.webkit.org/show_bug.cgi?id=155781

        Reviewed by Joseph Pecoraro.

        Added unit tests for test-group.js that would have caught regressions fixed in r198503.

        * public/v3/components/chart-pane-base.js:
        (ChartPaneBase.prototype._renderAnnotations): Added a forgotten break statement.
        * public/v3/models/build-request.js:
        (BuildRequest.prototype.setResult):
        (BuildRequest):
        * public/v3/models/test-group.js:
        * unit-tests/measurement-set-tests.js: Use ./resources/v3-models.js to reduce the code duplication.
        * unit-tests/resources/v3-models.js: Import more stuff from v3 models.
        (beforeEach):
        * unit-tests/test-groups-tests.js: Added. Added some unit tests for TestGroup.
        (sampleTestGroup):
        (.testGroupWithStatusList):

2016-03-22  Ryosuke Niwa  <rniwa@webkit.org>

        Fix a typo.

        * config.json:

2016-03-21  Ryosuke Niwa  <rniwa@webkit.org>

        Commit log viewer repaints too frequently after r198499
        https://bugs.webkit.org/show_bug.cgi?id=155732

        Reviewed by Joseph Pecoraro.

        The bug was caused by InteractiveTimeSeriesChart invoking onchange callback whenever mouse moved even
        if the current point didn't change. Fixed the bug by avoiding the work if the indicator hadn't changed
        and avoiding work in the commit log viewer when the requested repository and the revision range were
        the same as those of the last request.

        * public/v3/components/commit-log-viewer.js:
        (CommitLogViewer):
        (CommitLogViewer.prototype.currentRepository): Exit early when repository and the revision range are
        identical to the one we already have to avoid repaints and issuing multiple network requests.
        * public/v3/components/interactive-time-series-chart.js:
        (InteractiveTimeSeriesChart.prototype._mouseMove): Don't invoke _notifyIndicatorChanged if the current
        indicator hadn't changed.
        * public/v3/pages/chart-pane.js:
        (ChartPane.prototype._indicatorDidChange): Fixed the bug that unlocking the indicator wouldn't update
        the URL. We need to check whether the lock state had changed. The old condition was also redundant
        since _mainChartIndicatorWasLocked is always identically equal to isLocked per the prior assignment.

2016-03-21  Ryosuke Niwa  <rniwa@webkit.org>

        Fix A/B testing after r198503.

        * public/include/build-requests-fetcher.php:

2016-03-21  Ryosuke Niwa  <rniwa@webkit.org>

        Analysis task page is broken after r198479
        https://bugs.webkit.org/show_bug.cgi?id=155735

        Rubber-stamped by Chris Dumez.

        * public/api/measurement-set.php:
        (AnalysisResultsFetcher::fetch_commits): We need to emit the commit ID as done for regular data.
        * public/include/build-requests-fetcher.php:
        (BuildRequestsFetcher::fetch_roots_for_set_if_needed): Ditto. Don't use a fake ID after r198479.
        * public/v3/models/commit-log.js:
        (CommitLog): Assert that all commit log IDs are integers to catch regressions like this in future.
        * public/v3/models/root-set.js:
        (RootSet): Don't resolve Repository here as doing so would modify the shared "root" entry in the JSON
        we fetched, and subsequent construction of RootSet would fail since this line would blow up trying to
        find the repository with "[object]" as the ID.
        * public/v3/models/test-group.js:
        (TestGroup._createModelsFromFetchedTestGroups): Resolve Repository here.

2016-03-21  Ryosuke Niwa  <rniwa@webkit.org>

        v3 UI sometimes don't update the list of revisions on the commit log viewer
        https://bugs.webkit.org/show_bug.cgi?id=155729

        Rubber-stamped by Chris Dumez.

        Fixed multiple bugs that were affecting the list of blame range and commit logs for the range weren't
        updated in some cases on v3 UI. Also, the commit log viewer state is now a part of the URL state so
        opening and closing the commit log viewer will persist across page loads.

        Also fixed a regression from r198479 that Test object can't be created for a top level test.

        * public/v3/components/chart-pane-base.js:
        (ChartPaneBase.prototype.configure):
        (ChartPaneBase.prototype._mainSelectionDidChange): Fixed the bug that the list of blame range nor the
        commit log viewer don't get updated when the selected range changes.
        (ChartPaneBas