I believe this code has just been migrated from an earlier hack for Safari. We used to just clone the event object only for Safari but now we clone it for all browsers. The best thing that can be done to help us make sure we don't browser sniff unless we absolutely have to is to create a new ticket with a test case and even better a patch.
Thanks :) -- Brandon Aaron On 8/30/07, Michael Geary <[EMAIL PROTECTED]> wrote: > > > > > In this case you are fixing a bug that happens to appear in one > > > browser, but why limit the fix to only that browser? What > > > if another browser based on Safari comes out but has a > > > different user agent string and isn't recognized by jQuery? > > > Wouldn't you still want the bug to be fixed? What harm > > > would it cause to leave out the browser check? > > > Why fix something that isn't broken? You can play the > > hypothetical both ways without satisfaction. What if another > > browser comes out that breaks when this fix is applied? It's > > a pragmatic solution to a problem that exists today. > > In general, I agree with you. But look at the specific code again: > > // check if target is a textnode (safari) > if (jQuery.browser.safari && event.target.nodeType == 3) > event.target = originalEvent.target.parentNode; > > The purpose of this code is to detect when event.target is a text node and > substitute the parent node instead. > > The only reason to make this specific to Safari would be if we *want* to > allow event.target to be a text node in other browsers. But we don't want > event.target to be a text node, in any browser. We'd always want to get > the > parent node in such a situation, even if it happened in some unknown > future > browser. > > -Mike > >
