Opened 11 years ago
Last modified 9 years ago
#11111 confirmed Bug
[Safari] Range.enlarge test failing in Safari 7.0 and Chrome 37 — at Version 4
| Reported by: | Olek Nowodziński | Owned by: | |
|---|---|---|---|
| Priority: | Normal | Milestone: | |
| Component: | General | Version: | 4.0 |
| Keywords: | Cc: |
Description (last modified by )
Test /dt/core/dom/range/enlarge.html#test_enlarge_element12 fails in latest Safari 7.0 that comes with OSX 10.9. It's also reproducible in Chrome 37 (see #12221).
The root of the problem is the following condition: https://github.com/ckeditor/ckeditor-dev/blob/major/core/dom/range.js#L1039
In this case sibling is empty <b></b> (<p>Test <i><b></b>[Enlarge]</i></p>). Unlike previous versions, Safari 7 considers sibling.$.offsetWidth to be 1 at that point, while in other browsers it is 0. So the conclusion is that empty elements may have some width, which seems to be a regression in browser's engine.
That particular test fails since CKeditor 4.0 in Safari 7 which makes it clear that we're facing Safari regression.
Related bug in Chrome: #12221.
Change History (5)
Changed 11 years ago by
| Attachment: | emptyOffsetWidth.html added |
|---|
Sample to reproduce the issue without editor
comment:2 Changed 11 years ago by
I pushed a commit to major (7d732b3) that marks that particular test as a regression in Safari.
comment:4 Changed 10 years ago by
| Description: | modified (diff) |
|---|---|
| Status: | new → confirmed |
| Summary: | [Safari] Range.enlarge test failing in Safari 7.0 → [Safari] Range.enlarge test failing in Safari 7.0 and Chrome 37 |
The same regression made up to Chrome - #12221.
I submitted a bug to Webkit: https://bugs.webkit.org/show_bug.cgi?id=124037.
Also reported for Chrome 37: https://code.google.com/p/chromium/issues/detail?id=395136