On 06/02/12 15:58, lkcl luke wrote: > On Mon, Feb 6, 2012 at 3:45 PM, Phil Charlesworth > <[email protected]> wrote: > >> On 06/02/12 13:24, lkcl luke wrote: >> >>> On Mon, Feb 6, 2012 at 12:39 PM, Phil Charlesworth >>> <[email protected]> wrote: >>> >>> >>>> On 05/02/12 16:12, lkcl luke wrote: >>>> >>>> >>>>> thank you to rasiel for this one. >>>>> >>>>> http://code.google.com/p/pyjamas/issues/detail?id=661 >>>>> >>>>> ugh! new platform needed, at some point (IE8/9). this will do for now. >>>>> >>>>> could someone verify that this is also borked for pyjd (MSHTML) engine >>>>> when IE9 is installed? >>>>> >>>>> >>>> I'm puzzled about this because I've never seen this bug either in pyjs >>>> or pyjd on IE9. I've just created a test app for pyjd with a >>>> three-button radio group and it ran without any errors. >>>> >>>> >>> ok, then that's good enough. whew. >>> >>> hmm... it could be related to quirks / strict mode. or, maybe the >>> code in IE does something weird, does some sort of translation and >>> HTML5 compliance that they simply cannot put into the MSHTML engine >>> itself due to having to ensure backwards-compatibility. >>> >>> there are some really odd things that have to be done in the webkit >>> javascript bindings - stuff that simply doesn't exist in the >>> underlying c++ HTML5 engine to which those JS bindings are added. >>> automatic integer-to-string and vice-versa conversion, stripping off >>> of "px" on certain parameters etc. whereas the underlying >>> corresponding function that's called in c++ only takes e.g. an int. >>> >>> so... yeah, it may sound odd but it kinda makes sense. >>> >>> thanks for checking phil. >>> >>> l. >>> >> There is more. I have just removed createInputRadio() from DOM.ie6.py >> and built a pyjs version of my RadioTest application and it still >> doesn't produce any errors in any of the possible combinations of >> Browser Mode and Standards Mode in IE9. Ah, I've just noticed he used >> -O in compiling! So I've recompiled with that option. Still no errors! >> >> I guess the override is not doing any harm but it's not obvious that >> it's doing anything really different for IE9. >> > bizarre. > > hmmm... when you say browser mode and standards mode, is that by > putting a DTD XHTML jobbie at the top of the html loader file? Sorry, I should have called it Document Mode not Standards Mode. If you select Developer Tools from the menu after clicking the gear-wheel, you get some buttons at the bottom of the window, two of which are Browser Mode and Document Mode, so you can select anything from IE7, IE8, IE9 or even IE9 Compatibility View by the Browser Mode button and Quirks Mode, IE7 Standards, IE8 Standards and IE9 Standards with the Document Mode button. As you say, you can play tunes with DOCTYPEs and things like <meta http-equiv="X-UA-Compatible" content="IE=5" /> as well or instead. My test application just had <html>, so it went into quirks mode by default but the buttons must override whatever would otherwise be used, Otherwise they would be pointless. P.
