We need to change the new text field so the outline is exactly like the old one.
Created attachment 7445 [details] patch with detailed change log
Created attachment 7455 [details] patch with detailed change log -- also deals with blue/graphite appearance
Comment on attachment 7455 [details] patch with detailed change log -- also deals with blue/graphite appearance r=me, assuming you talk to Hyatt about the strict mode issue at some point :)
Has this landed? Something recently caused the pixel tests to regress wrt focus rings: http://build.webkit.org/results/post-commit-pixel-powerpc-mac-os-x/128/
(In reply to comment #4) > Has this landed? No, but I'm landing it right now, along with new pixel test results. > Something recently caused the pixel tests to regress wrt focus rings: > http://build.webkit.org/results/post-commit-pixel-powerpc-mac-os-x/128/ That was the WebKitSystemInterface change, which was a precursor to this one. I knew it would change the appearance of focus rings; sorry, I forgot to mention that to you.
outline-offset is an actual CSS property. By changing it to accept negative values, we now violate the spec. We need to log this as an issue in our bug database so that we can lobby to get the spec changed. If not, we will have to change outline-offset to -khtml-outline-offset before shipping.
In Radar as <rdar://problem/4497131>
Actually nvm. It looks like the restriction does not exist any more in the current spec.
*** This bug has been marked as a duplicate of 8093 ***