Test case: data:text/html,<!doctype html> <span style="border-width: initial"></span> <script> var span = document.querySelector("span"); span.style.borderWidth = ""; span.style.removeProperty("border-width"); span.style.removeProperty("borderWidth"); alert(span.style.cssText); </script> In Chrome 15 dev, this alerts "border-width: initial; ". Opera 11.50 does something similar. Firefox 8.0a2 alerts the empty string, as expected. A similar test in IE9 (it doesn't like the data URL, so I used Live DOM Viewer) also alerts the empty string. Chrome's behavior is very unexpected. Chrome behaves the same for border-color. In fact, if you do data:text/html,<!doctype html> <span style="border: 0"></span> <script> var span = document.querySelector("span"); span.style.border = ""; alert(span.style.cssText); </script> it alerts "border-style: initial; border-color: initial; ", which suggests these two properties have some peculiar bug.
Created attachment 117789 [details] Proposed patch
Comment on attachment 117789 [details] Proposed patch Attachment 117789 [details] did not pass chromium-ews (chromium-xvfb): Output: http://queues.webkit.org/results/10701556 New failing tests: fast/dom/HTMLMeterElement/meter-element-markup.html fast/dom/HTMLProgressElement/progress-element-markup.html
Created attachment 117800 [details] Patch
Comment on attachment 117800 [details] Patch Clearing flags on attachment: 117800 Committed r101970: <http://trac.webkit.org/changeset/101970>
All reviewed patches have been landed. Closing bug.
The test-cases from comment #0 still fail in Google Chrome 17.0.963.26 (Official Build 116225) dev. about:version reports WebKit 535.11 (@103967), and 103967 is greater than 101970. Do the test-cases in comment #0 actually work correctly in WebKit trunk?
(In reply to comment #6) > The test-cases from comment #0 still fail in Google Chrome 17.0.963.26 (Official Build 116225) dev. about:version reports WebKit 535.11 (@103967), and 103967 is greater than 101970. Do the test-cases in comment #0 actually work correctly in WebKit trunk? I'm not familiar enough with the Chrome release process to know what's up with that. FWIW it works in WebKit trunk with Safari, and with the latest Chrome Canary build. :)
So, marking RESOLVED again.