15710 2012-01-25 18:12:15 +0000 Serialization of inline style should be specified for function lists (e.g., transform) 2013-08-23 19:05:14 +0000 1 1 1 Unclassified CSS CSSOM unspecified All All NEW P2 normal --- 1 ayg zcorpan ayg cmarrin dino eoconnor glenn smfr public-css-bugzilla oldest_to_newest 63107 0 ayg 2012-01-25 18:12:15 +0000 data:text/html,<!DOCTYPE html> <script> document.head.style.transform = 'RotatE( 45dEg )scalex(2)'; document.documentElement.textContent = document.head.style.transform </script> This outputs "rotate(45deg) scaleX(2)" in IE9, Chrome 17 dev, and Opera Next 12.00 alpha. In Firefox 12.0a1, it outputs "Rotate(45deg) scalex(2)", not normalizing the capitalization of the transform functions. It should be specified that Firefox's behavior is incorrect. 63114 1 ayg 2012-01-25 19:23:57 +0000 More thorough test-case: data:text/html,<!DOCTYPE html> <script> document.head.style.transform = 'RotatE( 3rAd )scalex(2) translateX(10%)'; document.documentElement.textContent = document.head.style.transform </script> IE9, Chrome 17 dev: rotate(3rad) scaleX(2) translateX(10%) Firefox 12.0a1: RotatE(3rad) scalex(2) translateX(10%) Opera Next 12.00 alpha: rotate(171.89deg) scaleX(2) translateX(10%) I suggest the IE/Chrome behavior as standard. Units should be preserved, unlike Opera; and case should be normalized, unlike Gecko. Gecko bug: https://bugzilla.mozilla.org/show_bug.cgi?id=721136 64040 2 ayg 2012-02-13 17:31:29 +0000 This should probably be in CSSOM, actually, not the transforms spec. 64612 3 dschulze 2012-02-26 15:31:05 +0000 (In reply to comment #2) > This should probably be in CSSOM, actually, not the transforms spec. Can we move it to product CSSOM? 64830 4 ayg 2012-03-01 17:06:52 +0000 Now that CSSOM is being actively edited, I'm happy to move it there: http://dvcs.w3.org/hg/cssom/raw-file/tip/Overview.html/#serializing-css-values That should say how to serialize a function. The serialization should be something like: first the function's name, in whatever case is used in the specification (usually but not always lowercase); then "("; then the arguments, serialized according to their type, separated by ", "; then ")". Glenn, does that sound good to you? 64836 5 glenn 2012-03-01 17:30:37 +0000 (In reply to comment #4) > Now that CSSOM is being actively edited, I'm happy to move it there: > > http://dvcs.w3.org/hg/cssom/raw-file/tip/Overview.html/#serializing-css-values > > That should say how to serialize a function. The serialization should be > something like: first the function's name, in whatever case is used in the > specification (usually but not always lowercase); then "("; then the arguments, > serialized according to their type, separated by ", "; then ")". > > Glenn, does that sound good to you? sounds good; have this on my edit list already! 92520 6 zcorpan 2013-08-23 19:05:14 +0000 Currently CSSOM switches based on component value type, so the following (at least) are missing <transform-list> <transform-function> <translation-value> (I'm not sure we can use more abstract types like css-syntax's component value, since we need special rules for e.g. <color>.)