Archive for the 'cross browser' Category

Facebook login broken in Safari 4 beta?

Saturday, May 30th, 2009

Here’s a fun, major example of a cross-browser bug in the wild. It seems that, for certain users of the Safari 4 beta (latest update after the Mac OSX release), logging into Facebook is totally hosed. If I try to log in, Safari gives me a cursor with a lovely blue ball. It doesn’t even spin like the beach ball of death.

Running the Safari Developer tools error console yields the following errors on Facebook’s end:

Unsafe JavaScript attempt to access frame with URL http://www.facebook.com/home.php? from frame with URL
http://static.ak.facebook.com/common/redirectiframe.html. Domains, protocols and ports must match.

Unsafe JavaScript attempt to access frame with URL http://www.facebook.com/home.php? from frame with URL
http://0.channel14.facebook.com:80/[snip...]. Domains, protocols and ports must match.

Basically, it looks like Facebook’s trying to send a request during the login process from a subdomain of their site. Safari’s flagging it as breaking the AJAX “same origin” URL policy and preventing the request from completing. By acting correctly - Facebook’s requests are indeed breaking the same origin policy - Safari’s also blocking me (and a whole bunch of other people) from being able to login. (Now, I could be wrong here with the exact cause of the behavior, but the bottom line is something on Facebook is still totally breaking login in Safari 4.)

Is the new Safari being too strict? Why can I login in Firefox and Google Chrome? Is this an odd use case in a browser with a small market share (I’m guessing less than 4% of Facebook’s base) that wouldn’t be caught during testing as it’s not part of Facebook’s test platform, or is it a legitimate cross-browser bug that will be solved?

Clearing my cache (holding SHIFT and pressing the refresh button in the address bar or holding CONTROL + F5 / CONTROL + FN + F5 on a macbook) seemed to do the trick: the requests still showed up as failing, but I was able to log in. But, I’m still interested to see if this winds up being a fairly high-profile example of cross-browser development’s pitfalls on a really big stage.

Very, very odd.

IE and Disappearing Google Maps API Routes

Thursday, January 15th, 2009

I ran into this one earlier today. Essentially, you use the Google Maps API’s GDirections object to load some directions into your map. Occasionally, without warning or any pattern I can see, the little blue-ish route polyline will begin disappearing in IE 6 and IE 7. There seems to be no rhyme or reason to it, although I suspect that it has to do with DOM modifications after the map redraws.

Speculation aside, here’s a fix that I dug out of the Google Maps API Group. Make sure your document is using a valid DOCTYPE (duh) and then add the following XML namespacing to your HTML tag:


<html xmlns=”http://www.w3.org/1999/xhtml” xmlns:v=”urn:schemas-
microsoft-com:vml”>
</html>

I didn’t have access to that part of the document as I’m working in a CMS, but found that this JavaScript version worked:

var el = document.getElementsByTagName('head')[0];
el.setAttribute('xmlns','http://www.w3.org/1999/xhtml');
el.setAttribute('xmlns:v', 'urn:schemas-microsoft-com:vml');

That should fix the problem. However, it may pop-up again if you clear the map results and any overlays you have, and plot points on that same map again. Really odd. Does Google clobber that namespace at some point, after centering or drawing a map? I haven’t been able to test to find out. My only thought would be to set those attributes after every redraw or map load, but that could be a bit of overkill.

Still, if you can promise only a single load of a set of directions, the above should save you. Here’s the Google Groups thread for posterity.

IEBlog : Microsoft CSS Vendor Extensions

Tuesday, September 9th, 2008

As you may know, all browsers have a set of CSS features that are either considered a vendor extension (e.g. -ms-interpolation-mode), are partial implementations of properties that are fully defined in the CSS specifications, or are implementation of properties that exist in the CSS specifications, but aren’t completely defined. According to the CSS 2.1 Specification, any of the properties that fall under the categories listed previously must have a vendor specific prefix, such as ‘-ms-’ for Microsoft, ‘-moz-’ for Mozilla, ‘-o-’ for Opera, and so on.

IEBlog : Microsoft CSS Vendor Extensions

Nice to see IE jump on the vendor extension prefix bandwagon. Although it does break backwards compatibility to an extent (annoying), it will allow browsers to play better together.

Also, nice to see that they’re doing the “implement draft CSS specs” as prefixed, too.

Tags: ,

Weeeeeird IE6 overflow bug

Friday, July 18th, 2008

Just ran into a situation with a strange CSS overflow bug in IE6. The basic set up is a li containing a bunch of divs, one of which has a 0px height and overflow set to hidden (I’m going to be expanding it later in an accordion).

In this case, IE6 ignores the overflow setting on the div, and defaults to some kind of overflow:visible situation.

I messed around with a bunch of style attributes before finally trying to set overflow:hidden on the parent li. This worked like a charm.

So, if you’re ever in a situation where overflow isn’t being honored correctly, try setting it on a parent element.

Tags: , ,

Google, Yahoo spiders can now crawl through Flash sites

Tuesday, July 1st, 2008

As anyone who has had the pleasure of doing web design and development through marketing agencies knows, Flash tends to be wildly popular among clients and wildly unpopular among, well, pretty much everyone else. Part of the reason for this is because Flash is so inherently un-Googleable; anything that goes into a Flash-only website is basically invisible to search engines and therefore, the world. That will no longer be the case, however, as Adobe announced today that it has teamed up with Google and Yahoo to make Flash files indexable by search engines.

Google, Yahoo spiders can now crawl through Flash sites

Score one for Flash. Will this cause Flash adoption to ramp up on major websites, or will it still lag behind HTML’s ease-of-development?

As we all know, one of the biggest arguments for Flash is that it looks the same and performs (relatively) the same in all browsers. “Searchability” was a huge minus; you’d almost inevitably need to create another HTML-based site to stay in the search engines. With that solved, Adobe takes another step towards becoming a really ubiquitous platform, although I still think they need to open source their development tools to really, really take over. (Will all this destroy HTML? Nope. HTML still has Flash beat in the learning curve barrier to entry; you still need to be a specialist to build Flash apps, but a generalist can at least dabble in HTML and build a web application.)

Here’s my big question of the day: if Google spiders deep within your Flash movie and finds some sort of content, how does it provide a link to it? Is deeplinking to content going to be a problem, especially when it may discover content that’s not really part of any state you’ve really considered? If there’s a page 5 levels deep in my HTML site, Google can find the page and a URL for it; that is not necessarily true of the same type of content located 5 “views” off the stage in a Flash piece. Looks like we’ll be embedding deep linking information for various program states into more and more of our Flash apps…

Tags: , ,

Quick verdict on IE tester

Thursday, May 29th, 2008

Following up my last post on IE tester, I’ve been using it most of the day in my ordinary development activities, and I’ve drawn the following quick conclusions: this app is excellent for quick, visual QA of an application, but, until toolbars are supported, won’t quite cut it for development.

There are still times when you need the MS IE Developer Toolbar to muck with the DOM or modify styles, and that’s not supported here. If toolbar support makes it in, this could be a really great product, especially if modifying the contents of one window was reflected in all the others, kind of like cascading changes.

Still, this is an early-stage piece of software that already should cut down QA cycles a bit.

My DebugBar | IETester / HomePage

Thursday, May 29th, 2008

IETester is a free WebBrowser that allows you to have the rendering and javascript engines of IE8 beta 1, IE7 IE 6 and IE5.5 on Vista and XP, as well as the installed IE in the same process.

My DebugBar | IETester / HomePage

Side-by-side IE6, 7, and 8 is fantastic. Looks like there’s a few minor bugs, but I’ll try it out, if only to get rid of the 7 versions of IE I have installed here. :-)

Tags: , ,

IE6 in SP3 …

Friday, May 16th, 2008

IE6 in SP3 has the same version of jscript (@_jscript_version) number as IE7. (5.7 in this case). If you’re using jscript to target conditional compilation, heads up.

Tags: , ,

pi-js - Google Code

Tuesday, April 8th, 2008

pi.debugger is a cross browser web development tool allows you to examine web page’s structural and debug with console.

pi-js - Google Code

A formidable entry into the debug toolbars. I played with it for a few minutes and it doesn’t appear to have "live edit" of the DOM or CSS like Firebug (and to a lesser extent IE Developer Toolbar), but it does have a really, really nice console.

Recommendation: keep Firebug and add this to IE and Safari; now you have a console in all your browsers. Great app.

Tags: ,

Get Your Copy of IE8

Thursday, March 6th, 2008

Download the Internet Explorer 8 beta and check out all the stuff I blogged about yesterday. (Via the mighty Slashdot.)

Tags: ,

I'm Reading…
Search This Site
You are currently browsing the archives for the cross browser category.

AddThis Feed Button

Need great hosting?

Categories