There's no absolute requirement for submitting a patch of any sort, but the more testing you've done, the safer we'll feel about applying it. And, a unit test would be extremely wonderful - very few people (committers or not) add unit tests for their work, so you'd be at the top of the game. :)
-- Adam On 5/12/07, Safurudin Mahic <[EMAIL PROTECTED]> wrote:
The cache issue I last wrote about had nothing to do with Trinidad, so nevermind it. As I wrote last time, only minor changes was necesarry to get PPR to work with Opera, both on Desktop and on the N800. I could provide a patch via JIRA, but I haven't done any formal tests with it other than testing it with different versions of the browser. I'm also not sure about the guidelines of providing patches - is unit testing a necesarry part of the patch? --Safi Safurudin Mahic skrev: > Your reply encouraged me to try a hack solution. > I've managed to successfully test the PPR panelPageRoot and poll components > with following browsers: > > Opera 8.54 Windows Desktop > userAgent string "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; nb) > Opera 8.54 " > > Opera 9.20, Windows Desktop > userAgent string "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; nb) > Opera 9.20" > > Opera 8.5 Nokia N800, Linux > userAgent String "Mozilla/4.0 (compatible; MSIE 6.0; X11; Linux armv6l; > U) Opera 8.5 [no_NO] Maemo browser 0.7.11 RX-34_2007SE_3.2007.10-7" > > Basically the reason for Opera is not rendered as PPR capable or with > partially broken PPR capabilities > is that the browser up to version 9.0 always spoofed itself as MSIE. > > What I did, was to rearrange some of the if-elses in Core.js and > AgentFactoryImpl.java. and instead > of Opera beeing detected falsly as MSIE because of its spoof, I've > hardcoded it to be detected as Gecko > by the JavaScript. > > I'm not sure if this breaks any compability whatsoever, but it works for > PPR at least, which was my goal. > > I have however another issue now, at least on the N800 device, I'm not > sure how to exactly test this > on the desktop: > > Every time a PPR is issued, it seems as the device downloads the script > file as well, which is rather unnecessary - could this > have anything to do with the browser not respecting header > repsonse/expiry date of the script file? > > -- Safi > > Adam Winer skrev: > >> The detection and support isn't there - we'd love a patch from >> someone with this device who could implement and test this >> functionality. >> >> -- Adam >> >> >> On 5/3/07, Safurudin Mahic <[EMAIL PROTECTED]> wrote: >> >>> Hi, >>> >>> I am using trunk builds of Trinidad, it seems as the framework isn't >>> capable of detecting the Opera browser (mobile edition, 8.5) on the >>> Nokia N800 device as capable of doing PPR, page refreshs that were >>> supposed to be PPR are actually full page refreshs. >>> >>> The framework is able to detect a regular desktop Opera 8.5 browser >>> as PPR-capable. >>> >>> I believe it should be able to detect the Opera mobile as >>> PPR-capable also, because both the mobile and the desktop browsers >>> support the XmlHttpRequest JS object model, and I've found the N800 >>> device >>> to be working on several AJAX-enabled sites. >>> >>> >>> Hope someone could provide an answer, or even better, a possible >>> workaround for correct detection. >>> >>> >>> Best regards, >>> Safurudin Mahic >>> >>> >>> > >