Issue metadata
Sign in to add a comment
|
Regression: Tutorial overlay is seen chopped on ‘Reditr’ app.
Reported by
dmascare...@etouch.net,
May 19 2016
|
||||||||||||||||||||||
Issue descriptionChrome Version:52.0.2741.0 (Official Build) 19d105a2a1ec66924ff415f27ff170db7a67ba36-refs/heads/master@{#394609} 32/64 bit OS:Windows (7, 8, 10), Mac (10.10.5)(10.11.4),Linux (ubuntu 14.04 LTS) Test url: https://chrome.google.com/webstore/detail/reditr-the-best-reddit-cl/pmfcbbijgnhoebddbjpmlikabnbnddgb/related?hl=en What steps will reproduce the problem? 1. Launch chrome and navigate to above url. 2. Click on 'Add to chrome' and then Launch the app. 3. Observe tutorial overlay (floating) Actual: Tutorial overlay is seen chopped. Expected: Tutorial overlay should be seen properly. This is regression issue,broken in ‘M 47’ and below is narrow bisect: https://chromium.googlesource.com/chromium/src/+log/e2f0db1966773fa42be903330ad6497821272dec..92b949321abee66eeda20027ecfdf0973c4756ee?pretty=fuller&n=100 Blink bisect: https://chromium.googlesource.com/chromium/blink/+/684d928aea2fa5ad1563bc05cf04e85ae8744735?pretty=fuller&n=10 Suspecting: 202578 from blink bisect Good build:47.0.2515.0 Bad build:47.0.2516.0
,
May 24 2016
They have a style rule that attempts to set z-index to 999999999999999999999999999999999999999999999999999. CSS folk WDYT? Should we clamp instead of reject at parse time?
,
May 24 2016
Made a quick test case, Firefox clamps instead of rejecting. http://jsbin.com/pafehehaki/edit?html,css,js,output
,
May 24 2016
I mean this is more of a spec question, but Tab wanted this (https://codereview.chromium.org/1914693002) so I think we should? Currently the parser rejects sufficiently large numbers, but otherwise we clamp during style computation.
,
May 24 2016
Assigning to Tab given existing patch, please reassign if you don't intend to work on this anymore.
,
May 24 2016
Sorry about that, I'll finish up my patch this week. ^_^ (As a spec question, we intend to mandate clamping. There's a commented-out section of V&U that specifies this, but we couldn't get sufficient agreement to actually put it in the spec. Knowledge of browser-engine internals shouldn't be required to write valid stylesheets.)
,
Jul 25 2016
Tab, any update?
,
Aug 15 2016
Supposedly fixed by: https://codereview.chromium.org/2115923002/ At least, the jsbin case now passes. |
|||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||
Comment 1 by alancutter@chromium.org
, May 24 2016