third_party/WebKit/LayoutTests/fast/hidpi/scrollbar-appearance-decrease-device-scale-factor.html - Issue 2377453004: Scrollbar width is set using thickness to compute correct rect size

Side by Side Diff: third_party/WebKit/LayoutTests/fast/hidpi/scrollbar-appearance-decrease-device-scale-factor.html

Issue 2377453004: Scrollbar width is set using thickness to compute correct rect size (Closed)

Patch Set: Added test expectations Created 4 years, 2 months ago

Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.

Jump to:

« no previous file with comments | « third_party/WebKit/LayoutTests/TestExpectations ('k') | third_party/WebKit/LayoutTests/fast/hidpi/scrollbar-appearance-decrease-device-scale-factor-expected.png » ('j') | no next file with comments »
Toggle Intra-line Diffs ('i') | Expand Comments ('e') | Collapse Comments ('c') | Hide Comments ('s')

OLD	NEW
(Empty)
	1 <!DOCTYPE html>

	2 <html>

	3 <head>
	Xianzhu 2016/09/30 03:36:14 Normally we omit <html>, <head>, <body> in layout Normally we omit <html>, <head>, <body> in layout tests. malaykeshav 2016/09/30 18:23:39 Done Show quoted text On 2016/09/30 at 03:36:14, Xianzhu wrote: > Normally we omit <html>, <head>, <body> in layout tests. Done
	4 <script>

	5 window.enablePixelTesting = true;
	Xianzhu 2016/09/30 03:36:14 Please remove the above line. It's no use here. Y Please remove the above line. It's no use here. You may have seen some tests using it. These tests also includes some script files which will read the variable. malaykeshav 2016/09/30 18:23:39 Done Show quoted text On 2016/09/30 at 03:36:14, Xianzhu wrote: > Please remove the above line. It's no use here. > > You may have seen some tests using it. These tests also includes some script files which will read the variable. Done
	6 function finishTest() {

	7 setTimeout(function() {testRunner.notifyDone();}, 0);
	Xianzhu 2016/09/30 03:36:14 Is the setTimeout necessary? Is the setTimeout necessary? malaykeshav 2016/09/30 18:23:39 Removed. Show quoted text On 2016/09/30 at 03:36:14, Xianzhu wrote: > Is the setTimeout necessary? Removed.
	8 }

	9

	10 function scaleDown() {

	11 if (window.testRunner) {

	12 testRunner.setBackingScaleFactor(1, finishTest);

	13 }

	14 }

	15 function startTest() {

	16 testRunner.waitUntilDone();

	17 if (window.testRunner) {
	Xianzhu 2016/09/30 03:36:14 Line 16 should be also protected by if (window.tes Line 16 should be also protected by if (window.testRunner). Or you can just add 'if (window.testRunner)' before line 22 and remove all other guards. malaykeshav 2016/09/30 18:23:39 Done Show quoted text On 2016/09/30 at 03:36:14, Xianzhu wrote: > Line 16 should be also protected by if (window.testRunner). > Or you can just add 'if (window.testRunner)' before line 22 and remove all other guards. Done
	18 testRunner.setBackingScaleFactor(3, scaleDown);

	19 }

	20 }

	21

	22 window.addEventListener("load", startTest, false);
	Xianzhu 2016/09/30 03:36:14 You can use shorter form: onload = startTest; or You can use shorter form: onload = startTest; or replace line 15 with onload = function() { malaykeshav 2016/09/30 18:23:39 Done Show quoted text On 2016/09/30 at 03:36:14, Xianzhu wrote: > You can use shorter form: > onload = startTest; > or replace line 15 with > onload = function() { Done
	23 </script>

	24 </head>

	25 <body>

	26 <div id="div" style="width:200px; height:120px; overflow:auto">

	27 <br>This should be a scrollable text box with horizontal and vertical scroll.

	28 <div style="width:400px; height: 200px; background-color:red"></div>
	Xianzhu 2016/09/30 03:36:14 Remove "background-color: red". Normally we use re Remove "background-color: red". Normally we use red to indicate failure. malaykeshav 2016/09/30 18:23:39 Done Show quoted text On 2016/09/30 at 03:36:14, Xianzhu wrote: > Remove "background-color: red". Normally we use red to indicate failure. Done
	29 </div>

	30 </body>

	31 </html>
	Xianzhu 2016/09/30 03:36:14 We prefer reference tests to pixel tests because t We prefer reference tests to pixel tests because the former can avoid pixel and text expectations and platform baselines. Just create a file named scrollbar-appearance-decrease-device-scale-factor-expected.html beside this file containing the following contents: <!DOCTYPE html> <div id="div" style="width:200px; height:120px; overflow:auto"> <br>This should be a scrollable text box with horizontal and vertical scroll. <div style="width:400px; height: 200px; background-color:red"></div> </div> The layout test runner script will run both html file and compare their pixel outputs. The test passes if the two files generate the same output. You can keep the other test as a pixel test. malaykeshav 2016/09/30 18:23:39 The two files will not be _exactly_ the same due t Show quoted text On 2016/09/30 at 03:36:14, Xianzhu wrote: > We prefer reference tests to pixel tests because the former can avoid pixel and text expectations and platform baselines. > > Just create a file named scrollbar-appearance-decrease-device-scale-factor-expected.html beside this file containing the following contents: > <!DOCTYPE html> > <div id="div" style="width:200px; height:120px; overflow:auto"> > <br>This should be a scrollable text box with horizontal and vertical scroll. > <div style="width:400px; height: 200px; background-color:red"></div> > </div> > > The layout test runner script will run both html file and compare their pixel outputs. The test passes if the two files generate the same output. > > You can keep the other test as a pixel test. The two files will not be _exactly_ the same due to the scaling. There is always a difference of a pixel or two due to loss from non-integral values during scaling. Xianzhu 2016/09/30 18:36:20 This looks weird to me. The final scale is 1 which Show quoted text On 2016/09/30 18:23:39, malaykeshav wrote: > On 2016/09/30 at 03:36:14, Xianzhu wrote: > > We prefer reference tests to pixel tests because the former can avoid pixel > and text expectations and platform baselines. > > > > Just create a file named > scrollbar-appearance-decrease-device-scale-factor-expected.html beside this file > containing the following contents: > > <!DOCTYPE html> > > <div id="div" style="width:200px; height:120px; overflow:auto"> > > <br>This should be a scrollable text box with horizontal and vertical > scroll. > > <div style="width:400px; height: 200px; background-color:red"></div> > > </div> > > > > The layout test runner script will run both html file and compare their pixel > outputs. The test passes if the two files generate the same output. > > > > You can keep the other test as a pixel test. > > The two files will not be _exactly_ the same due to the scaling. There is always > a difference of a pixel or two due to loss from non-integral values during > scaling. This looks weird to me. The final scale is 1 which is the same as the initial scale and the scale of reference html. We should do full repaint and full re-rasterization after scale change, so the final pixels should be exactly the same as the original and the reference. Can you share the pixel results?
OLD	NEW