On Jan 13, 3:23 pm, Diego Perini <diego.per...@gmail.com> wrote: > Kangax, > I could extend a bit the testing to ensure type is "unknown" but I > believe it is unfeasible since at some point that property will be > "number" in IE.
Perhaps. Although, IE8 still returns "unknown" for `typeof` of certain objects and it seems unlikely that they change internal representation anytime soon. > > That property can be tested for "unknown" only in rare and specific > occasions, in that case it should be done before anything else, after > DOM has finished loading that property will be a number, in case of > static pages, in case of dynamic pages that will depend on the headers > sent by the server...(no-cache, pragma et all). > > This is the reason I kept testing for just !== "undefined". > > If we are presuming that somebody with some bad code can fold our > checks, yes that will always be true, both if we use sniffing, > inference or feature testing. Really, people could also overwrite the > createElement, or prototype the HTMLIFrameElement to open a window > instead but we should draw a line about what is necessary and > consistent. True. We *do* need to draw a line at some point, but wouldn't you want to use a test that is tolerant to false positives produced by 3rd party scripts? For example, no 3rd party script can spoof `typeof` result to return "unknown" instead of, say, "object". Javascript/DOM simply do not allow such level of control. The same is true for conditional compilation syntax extension - i.e. /*...@cc_on */ - which every version of IE supports till this day and is very unlikely to drop. Using such tests still doesn't "save" you from changes in later versions of browsers; If IE starts being more standard-compliant, the test will return false positives. It nevertheless prevents you from 3rd party collisions (and so gives you one less potential annoyance to worry about). > > Why would somebody use the "fileSize" property if the 3/4 of the > browsers have this property natively....and why would they want to add You'll be surprised ; ) > that to FF / Opera / Webkit only and not to Internet Explorer, I > believe we are crossing the border of usefulness here... I don't mean to stress this subject any longer. I simply remember how assumptions like this lead to problems in the future (when someone implements something that you would never expect - e.g. `document.fileSize` - for whatever reasons the author has). [...] -- kangax --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "jQuery Development" group. To post to this group, send email to jquery-dev@googlegroups.com To unsubscribe from this group, send email to jquery-dev+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/jquery-dev?hl=en -~----------~----~----~----~------~----~------~--~---
