W3C home > Mailing lists > Public > public-html@w3.org > October 2014

Re: Document.createElement: null or HTML namespace?

From: Boris Zbarsky <bzbarsky@mit.edu>
Date: Wed, 08 Oct 2014 08:16:31 -0700
Message-ID: <5435554F.7020106@mit.edu>
To: Robin Berjon <robin@w3.org>, HTML WG <public-html@w3.org>
On 10/8/14, 8:10 AM, Boris Zbarsky wrote:
> Probably at least 3-4 years ago when we last checked.

Ah, to be precise as of 2011-09-06 there was a library in use on the web 
(e.g. on virginamerica.com, which has since done a site redesign, so we 
could _hope_ it's not using it anymore) with the following behavior:

1)  Sniff the UA string for "applewebkit" and if present use sync XHR.
2)  Otherwise create a document (via either createDocument or MSCOM
     stuff) and use document.load().

See the discussion in https://www.w3.org/Bugs/Public/show_bug.cgi?id=14037

So at least as of then, to be web-compatible you had to either have a UA 
string claiming to be webkit or implement load() in one of its incarnations.

We could try adding telemetry in Gecko to see whether we're still 
hitting pages that use this, I guess...

Received on Wednesday, 8 October 2014 15:17:01 UTC

This archive was generated by hypermail 2.4.0 : Saturday, 9 October 2021 18:46:10 UTC