Created attachment 92766 [details] Test case that fails. See the example. The number of nested span elements is proportional to the amount of incorrect space required on the RHS. Simply resize the window and you will see that the span elements are wrapped with plenty of space to spare. This appears to be incorrect behavior, works fine in FF, but also broken in Chrome (uses WebKit too right?). Thanks Samuel
Created attachment 92767 [details] A screen shot comparing Firefox to Safari to show the incorrect behavior.
Created attachment 92769 [details] Compare display: inline and display: inline-block I have been doing some more testing, and I found that if I add "display: inline-block;" to the top span, it shows how one might expect the wrapping to look. Unfortunately, inline-block isn't suitable because it doesn't break into separate pieces like a normal inline span.
Created attachment 96040 [details] Patch.
This isn't a regression from earlier padding changes for bidi, right?
(In reply to comment #4) > This isn't a regression from earlier padding changes for bidi, right? No. I have a build of QtWebKit 2.1 from almost a year ago, and I can reproduce the error there too.
Comment on attachment 96040 [details] Patch. Looks OK to me. I would like mitz or leviw to comment before you land though.
(In reply to comment #6) > (From update of attachment 96040 [details]) > Looks OK to me. I would like mitz or leviw to comment before you land though. LGTM! Thanks for the heads up :)
(In reply to comment #7) > (In reply to comment #6) > > (From update of attachment 96040 [details] [details]) > > Looks OK to me. I would like mitz or leviw to comment before you land though. > > LGTM! Thanks for the heads up :) Thank you both :)
Comment on attachment 96040 [details] Patch. Clearing flags on attachment: 96040 Committed r88175: <http://trac.webkit.org/changeset/88175>
All reviewed patches have been landed. Closing bug.