We will need to check with the original committer where the code came from. I could not find anything similar so rather than not be "supported" I switched locally to a spring based solution. It also is far superior code than was previously supplied.
The tablet renders currently as a mobile, which we can change if needed to render normal. Cheers Greg On 3 September 2014 14:41, Glen Mazza <glen.ma...@gmail.com> wrote: > I'm not comfortable with this change at the present, I think it is too > soon for us to move to three device support (now including tablets) and not > a good allocation of resources, at a time that multiple device checking is > nicely going out the window due to responsive themes and usage of media > queries. What we presently have, i.e., check for "Mobile" in the UA > string, then check a device listing, and then fallback to standard theme if > mobile unavailable will work for the vast majority of blogs today. And > such simplicity saves us time, allowing us to add more important features > that grab more bloggers than we'd lose by not separately supporting > tablets. Three-device support is going to require code changes throughout > the system to support, it's not just bringing in these few classes. > > I was hoping we could just update our list of devices we presently have > and just go with that--update one file alone. (Where did that original > source come from?) There are many sources for this information, even > JQuery will probably work because it's MIT-licensed. Let's consider > whether we need three-device support later, once we get user demand for it > (and your solution looks fine for it), but I'd rather we not be maintaining > something that our present user base isn't asking for. > > Glen > > On 09/03/2014 03:14 AM, Greg Huber wrote: > >> Checking the spring-mobile license it uses >> http://www.apache.org/licenses/LICENSE-2.0. >> >> So it looks ok to use, I will add a version which uses >> DeviceResolverRequestFilter and LiteDeviceResolver to determine the >> browser >> type (also its easily maintained by spring! ;) ) and which we can easily >> switch to. >> >> I have added the code Committed revision 1622172. If it is OK I will >> update roller accordingly. >> >> Cheers Greg >> >> >> >> On 2 September 2014 10:57, Glen Mazza<glen.ma...@gmail.com> wrote: >> >> No, we still support multiple renditions (i.e., basic-mobile) to be >>> defined if that's what the blogger wants, for single-rendition the >>> blogger >>> can use either a responsive theme or even a non-responsive one (my >>> smartphone just shrinks the image if it's non responsive, I can enlarge >>> it >>> and view chunks of the blog page.) >>> >>> The older code, if there was just the standard rendition defined, would >>> make a copy of it and make the copy the mobile rendition, requiring the >>> theme user to have to maintain two sets of templates even if they were >>> desired to stay identical (e.g., a responsive theme). When I took that >>> out >>> -- no copies unless two renditions are defined in the theme.xml -- I >>> apparently didn't get the code right for the standard theme to be the >>> default one. I'll get it fixed. >>> >>> As for the "browser user agent", I'm not sure if that "deviceType" >>> parameter is something that a Roller page creates once in a browser or >>> something all browsers supply regardless of the website that they are on, >>> Googling isn't bringing up much on that parameter so I'm assuming the >>> former. I'm pretty much new to this particular topic. >>> >>> Glen >>> >>> On 09/02/2014 02:45 AM, Greg Huber wrote: >>> >>> If there is no "mobile" on the theme.xml for the theme it used to show >>>> the >>>> default, so maybe something has changed. >>>> >>>> The browser user agent is used to determine if its a mobile device. >>>> What >>>> I >>>> do is to use the jquery mobile logic i.e. LiteDeviceResolver, I can >>>> update >>>> roller but am not sure on the licensing etc on copying jquery code. As >>>> you >>>> mentioned previously the preferred method now would be to use a >>>> responsive >>>> design, rather than a separate theme, so this is kind of parked? >>>> >>>> Cheers Greg >>>> >>>> >>>> On 2 September 2014 01:49, Glen Mazza<glen.ma...@gmail.com> wrote: >>>> >>>> Hi Team, I noticed today with Roller 5.1 the blogs are not rendering >>>> on >>>> >>>>> smartphones (at least mine, I have a Windows 8 smartphone that uses IE >>>>> as >>>>> its browser) except for the combo basic-mobile theme, the only one that >>>>> provides explicit "mobile" rendition types. For the others, Roller >>>>> just >>>>> returns a blank screen or a 404 or similar error page. To test, for my >>>>> website I created 5 empty blogs, one for each theme we offer: >>>>> >>>>> https://web-gmazza.rhcloud.com/frontpage/ >>>>> https://web-gmazza.rhcloud.com/gaurav/ >>>>> https://web-gmazza.rhcloud.com/testdual/ (basic-mobile). >>>>> https://web-gmazza.rhcloud.com/frontpage/ >>>>> https://web-gmazza.rhcloud.com/fauxcoly/ >>>>> >>>>> What I would like to have Roller do -- and I had incorrectly assumed >>>>> was >>>>> already being done -- was for Roller to fall back to the "standard" >>>>> rendition type when the "mobile" rendition was not available, correct >>>>> anyway if you're using a responsive theme. Searching through the code I >>>>> think the only change I need to do is in class RollerVelocity[1], for >>>>> those >>>>> getTemplate() methods that take a deviceType parameter, to attempt to >>>>> get >>>>> the standard rendition type as a fallback if the mobile deviceType was >>>>> requested and is not available. I'll test it. Until a Roller 5.1.1 is >>>>> out, users should be able to duplicate renditions in their theme.xml, >>>>> defining the standard one as also the mobile one. >>>>> >>>>> Couple of other concerns, in our MobileDeviceRepository class, our >>>>> device >>>>> listing[2] used as a backup to determine if mobile is necessary may be >>>>> out-of-date, I think I can Google something more recent. Also, just to >>>>> confirm, line #88 of that same file, checks the user agent "deviceType" >>>>> parameter for "standard" or "mobile" to determine the type, but that >>>>> parameter is not normally sent by a browser, correct? >>>>> >>>>> Regards, >>>>> Glen >>>>> >>>>> [1]http://svn.apache.org/viewvc/roller/trunk/app/src/main/ >>>>> java/org/apache/roller/weblogger/ui/rendering/ >>>>> velocity/RollerVelocity.java?revision=1583506&view=markup#l96 >>>>> [2]http://svn.apache.org/viewvc/roller/trunk/app/src/main/ >>>>> java/org/apache/roller/weblogger/ui/rendering/mobile/ >>>>> MobileDeviceRepository.java?revision=1611764&view=markup#l34 >>>>> >>>>> >>>>> >>>>> >