Our team is on the process of triaging. 

Meanwhile could you please update to the latest chrome version- 54.0.2840.100 from chrome://chrome and provide us the latest update.

It would be great if you could also provide a sample test case of the ease finding regression.

I am in a enterprise controlled environment, so I am unable to update the version of Chrome, but will test on my local machine when I get some free time.

Could you also please provide a test case, so that we can try to identify the breaking change?

Worked up a test case and attached.  Tested in Chrome 54.  In this test case, the issue seems to be overflow and setting role='presentation' on the container fixes it.  The issue we have in our app in Chrome 52 is similar, but we fixed it by using display:table-row-group;. I'll try and get a test case for that next week. 

Deleted: test2.html 1.7 KB

test2.html 1.7 KB View Download

Tested this issue on Win-10, Ubuntu 14.04 and Mac 10.11.6 using chrome reported version #52.0.2743.116, latest stable #54.0.2840.99, latest canary #57.0.2926.0 and chrome version #43.0.2334.0.

Attaching a screencast tested in the chrome version 43 for reference.

Following are the steps followed to reproduce the issue.
------------
1. Launched chrome.
2. Opened the test2.html in the browser.
3. Observed that all rows in the grid are represented as expected.

Note: The behavior is same in all the 3 OS.

Reporter@ - Could you please let us know if anything missed from our side and if possible please provide another sample test case to test this issue.

Thanks...!!

Deleted: 666341_43version.mp4 317 KB

	666341_43version.mp4 317 KB View Download

Its not an issue of screen redrawing, rather an accessibility issue.  If you go to chrome://accessibility, you'll notice that the rows are missing from the Accessibility Tree.

Thank you for providing more feedback. Adding requester "schenney@chromium.org" for another review and adding "Needs-Review" label for tracking.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Confirmed.

To reproduce:
1. Open the test file in one tab
2. Open chrome://accessibility in another tab and locate the line for test2.html, the test file
3. Enable accessibility and "show accessibility tree"
4. Note that the second set of R1C1, R1C2 etc are missing.

Back to testers for a bisect.

Here's the output I get. The missing lines should appear before IDs 60, 61, 62:

[document web] enabled focusable focused id=1

++name='Works as expected' enabled id=4

++++[text] name='Works as expected' enabled id=10

++[table] enabled id=5

++++enabled id=14

++++++[column header] enabled id=15

++++++++[text] name='C1' enabled id=35

++++++[column header] enabled id=18

++++++++[text] name='C2' enabled id=37

++++enabled id=20

++++++enabled id=22

++++++++[text] name='R1C1' enabled id=39

++++++enabled id=25

++++++++[text] name='R1C2' enabled id=41

++++enabled id=21

++++++enabled id=27

++++++++[text] name='R2C1' enabled id=43

++++++enabled id=30

++++++++[text] name='R2C2' enabled id=45

++++enabled id=32

++++enabled id=33

++++enabled id=34

++name='No rows show due to the overflow' enabled id=6

++++[text] name='No rows show due to the overflow' enabled id=47

++[table] enabled id=7

++++enabled id=51

++++++[column header] enabled id=52

++++++++[text] name='C1' enabled id=63

++++++[column header] enabled id=55

++++++++[text] name='C2' enabled id=65

++++enabled id=60

++++enabled id=61

++++enabled id=62

++name='Works, but requires role presentation on overflow container' enabled id=8

++++[text] name='Works, but requires role presentation on overflow container' enabled id=67

++[table] enabled id=9

++++enabled id=71

++++++[column header] enabled id=72

++++++++[text] name='C1' enabled id=93

++++++[column header] enabled id=75

++++++++[text] name='C2' enabled id=95

++++enabled id=78

++++++enabled id=80

++++++++[text] name='R1C1' enabled id=97

++++++enabled id=83

++++++++[text] name='R1C2' enabled id=99

++++enabled id=79

++++++enabled id=85

++++++++[text] name='R2C1' enabled id=101

++++++enabled id=88

++++++++[text] name='R2C2' enabled id=103

++++enabled id=90

++++enabled id=91

++++enabled id=92

Issue is reproducible on Mac 10.11.6,Win 10 and Ubuntu 14.04 using 55.0.2883.75 and canary 57.0.2944.0.

Bisect info:
============
Good:44.0.2401.0
Bad :44.0.2402.0

CHANGELOG URL:  https://chromium.googlesource.com/chromium/src/+log/171777f7e0ef8be56f2c60fe7e62dd2480bc7fd3..738487a69f150975376c36e214c64a77b1ff0cc6

BLINK CHANGELOG URL:
http://build.chromium.org/f/chromium/perf/dashboard/ui/changelog_blink.html?url=/trunk&range=195338%3A195336

Suspect from Blink CL:
https://codereview.chromium.org/1121473004
dmazzoni@: Could you please take a look into this if its related to your change.

I think this was likely caused by the linked change - it would make the <div style="overflow: auto"> no longer ignored (effectively role=presentation) meaning that the table rows are no longer considered direct children of the table. I think for tree we loosened the restrictions around what it meant for something to be a "child of" the parent which requires the owned children - seems like we should do something similar here. CCing aleventhal for thoughts.

@aboxhall, SGTM

This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Works as expected now.