As discussed in http://codereview.chromium.org/4678006/, Chromium has a plain to reuse Firefox's messages. Firefox has some type-specific valueMissing messages.
Created attachment 74208 [details] Patch
Comment on attachment 74208 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=74208&action=review > WebCore/html/ValidityState.cpp:46 > - if (valueMissing()) > - return validationMessageValueMissingText(); > if (patternMismatch()) > return validationMessagePatternMismatchText(); Do we need to check valueMissing before pattermMismatch?
Comment on attachment 74208 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=74208&action=review >> WebCore/html/ValidityState.cpp:46 >> return validationMessagePatternMismatchText(); > > Do we need to check valueMissing before pattermMismatch? Oh, good point. If a page has <input required pattern="[0-9]{4}"> and the field is empty, we should show valueMissing message, not patternMismatch message.
Created attachment 74302 [details] Patch 2
Comment on attachment 74302 [details] Patch 2 Clearing flags on attachment: 74302 Committed r72363: <http://trac.webkit.org/changeset/72363>
All reviewed patches have been landed. Closing bug.