if an inner element is positioned relative, and has both min-width and max-width specified where min-width > max-width, an outer element with position absolute won't receive the proper width value. min-width / max-width algorithm: 10.4 @ https://www.w3.org/TR/CSS2/visudet.html#min-max-widths JSFiddle demonstrating this behavior: https://jsfiddle.net/booc0mtaco/Lr9bzh07/ (chrome and firefox appear to resolve the width as expected) Expected: outer wrapper (blue) should be as wide as the inner element due to the width algorithm. Actual result: the outer element appears to simply take the max-width value, while the sized element does the proper calculation.
<rdar://problem/50967601>
Created attachment 370942 [details] Patch
Created attachment 371034 [details] Add a WPT test
Created attachment 371035 [details] Complete the WPT test's export process
Created attachment 371036 [details] Complete the WPT test's export process
Could you please review this change? And I wonder the included wpt test case will be merged automatically to the wpt github repository. At the command prompt below, Would you like to export the web-platform-tests changes to a WPT GitHub repository? [Y/n]: y I coundn't proceed the process smoothly(failed with some error in fact), so I wonder I don't have to care about that wpt export failure message.
I've read https://lists.webkit.org/pipermail/webkit-dev/2017-December/029837.html, and succeeded to update wpt test case with PR form using it. So current state of this patch is that if this patch get r+, the included wpt test can be merged too into wpt github repo.
Comment on attachment 371036 [details] Complete the WPT test's export process Clearing flags on attachment: 371036 Committed r245966: <https://trac.webkit.org/changeset/245966>
All reviewed patches have been landed. Closing bug.