I agree it's mainly for mobile. >see if you can put in a flag so folks can get back to the old behavior:
I propose the new "device-diagonal" selector, which doesn't change with the orientation. So the test could be: - if device-width or device-height is used, update styles on orientation change - if device-diagonal is used, don't update styles on orientation - if no device-xxx is used, don't update styles on orientation change. WDYT? Maurice -----Message d'origine----- De : Alex Harui [mailto:aha...@adobe.com] Envoyé : mercredi 19 mars 2014 15:40 À : dev@flex.apache.org Objet : Re: Advice needed on FLEX-33994 IIRC, pixel density is not reported accurately on the desktop, so computing inches and cms will have a certain amount of inaccuracy. That said, see if you can put in a flag so folks can get back to the old behavior if they find this one too slow, and go for it. Large desktop UI's with tons of widgets could be slow, but the main point is for mobile where there are probably fewer widgets at a time. -Alex On 3/19/14 7:32 AM, "Maurice Amsellem" <maurice.amsel...@systar.com> wrote: >Hi, > >I am preparing for https://issues.apache.org/jira/browse/FLEX-33994 ( >media size CSS ) and have a few interrogations: > >1) Specs: this is how I intend to implement the new media selectors: > >@media selectors: >device-width, device-height > >Supports the following units: >(no unit), px, pt, in, cm, dp >- no unit / px => physical pixels >- pt => pixels-equivalent at 72 ppi >- cm & in => self explanatory >- dp => pixels-equivalent at 160 ppi (bucket ppi) > >If I understand correctly the explanations from Justin, "dp" uses the >bucket DPIs, not the real ppis. >So I will use the real device ppi to compute pt, inch and cm pixel >equivalent and bucket-ppi to compute the pixel equivalent for "dp" unit. > >2) Supporting resize / orientation change: >Ideally, when resizing the SWF (on desktop) or changing the device >orientation / going full screen on mobile, the media queries should be >re-evaluated, like in HTML, with potentially different results. > >I am not an expert of the Style Management in Flex, but I am worried >that this would be very slow , especially on mobile, and may even break >the application. > >Since the initial need was to be able to distinguish mobile / tablet / >screen / tv, and there are other means of handling orientation change >in the UI , I thought it could be acceptable to compute width / height >independently of the actual device orientation (eg. always in portrait >mode). > >Another possibility , which is not W3C, is to provide new >"device-diagonal" measure, which is a common measurement for display >sizes Eg. >@media (max-device-diagonal: 5in) { > /* phone css*/ >} >@media (min-device-diagonal: 5in) and (max-device-diagonal: 11in) { > /* tablet css*/ >} >@media (min-device-diagonal: 11in) and (max-device-diagonal: 24in) { > /* screen*/ >} >@media (min-device-diagonal: 24in) { > /* TV*/ >} >What do you think? > >Maurice > > > > >
