The WHATWG Blog

Please leave your sense of logic at the door, thanks!

WHATWG Weekly: HTML canvas version 5 has arrived

The StringEncoding proposal is getting closer to consensus. It now consists of a TextEncoder and a TextDecoder object that can be used both for streaming and non-streaming use cases. This is the WHATWG Weekly.

Some bad news for a change. It may turn out that the web platform will only work on little-endian devices, as the vast majority of devices is little-endian today (computers, phones, …) and code written using ArrayBuffer today assumes little-endian. Boris Zbarsky gives a rundown of options for browsers on big-endian devices. Kenneth Russell thinks the situation can still be saved by universal deployment of DataView and sufficient developer advocacy.

Over the past couple of weeks the canvas element 2D API has gotten some major new features. Ian Hickson wrote a lengthy email detailing the canvas v5 API additions. Path primitives, dashed lines, ellipses, SVG path syntax, text along a path, hit testing, more text metrics, transforming patterns, and a bunch more.

Posted in Weekly Review | Comments Off on WHATWG Weekly: HTML canvas version 5 has arrived

WHATWG Weekly: Path objects for canvas and creating paths through SVG syntax

Jonas Sicking proposed an API for decoding ArrayBuffer objects as strings, and encoding strings as ArrayBuffer objects. The thread also touched on a proposal mentioned here earlier, StringEncoding. This is the mid-March WHATWG Weekly.

Revision 7023 added the Path object to HTML for use with the canvas element, and the next revision made it possible to actually use it:

var path = new Path()
path.rect(1,1,10,10)
context.stroke(path)

A new method addPathData() (introduced in revision 7026) can be used to construct canvas paths using SVG path data. Revision 7025 meanwhile added ellipse support to canvas.

Tune in next week for more additions to canvas.

Posted in Weekly Review | 2 Comments »

WHATWG Weekly: http+aes URL scheme, control Referer, …

Apple's Safari team provided feedback to the Web Notifications Working Group. That group, incidentally, is looking for an active editor to address that and other feedback. Opera Mobile shipped with WebGL support. This is March's first WHATWG Weekly.

Simon Pieters overhauled much of HTML5 differences from HTML4 and the document now provides information on added/changed APIs, differences between HTML and W3C HTML5, content model changes, and more.

Ian Hickson introduced a new URL scheme named http+aes (and also https+aes) in revision 7012 that allows for hosting private data on content distribution networks. Revision 7009 by the way, added the necessary hooks for the DOM mutation observers feature to HTML.

A new "referrer" metadata name for the meta element has been proposed on the WHATWG Wiki. This allows for controlling the Referer header on outgoing links.

Posted in Weekly Review | Comments Off on WHATWG Weekly: http+aes URL scheme, control Referer, …

WHATWG Weekly: New canvas API goodies

A draft for the SPDY protocol has been submitted, the W3C HTML WG mailing list goes crazy over media DRM. This is the WHATWG Weekly.

In response to feedback Adam Barth changed the getRandomValues() method to return the array the method modifies. The method is part of the window.crypto proposal.

Ian Hickson has been busy updating the Canvas Wiki page with proposals for dashed lines, ellipsis, hit regions, using SVG path syntax for paths, and path primitives. Updates to HTML itself seem imminent.

Posted in Weekly Review | Comments Off on WHATWG Weekly: New canvas API goodies

WHATWG Weekly: Unicode for the platform?

In less than a year we reached another arbitrary milestone. HTML is another thousand revisions further, now over 7000 (not quite 9000). This is the WHATWG Weekly.

Over on public-script-coord@w3.org, the mailing list used by TC39 (responsible for JavaScript) and the WebApps WG to coordinate development of JavaScript, IDL, and APIs, Brendan Eich launched a mega thread on full Unicode for ES6. The entire platform is currently build around 16-bit code units, which are not quite sufficient to encompass all code points. Some code points therefore require two code units, but string manipulation, length information, etc. is all in code units, making it difficult to deal with code points that require two (in practice nobody seems to bother much). The idea is to introduce some kind of switch which when used would let you deal with code points exclusively, rather than code units.

HTML did not change much last week as its editor was playing in the snow. The DOM meanwhile now has mutation observers defined, the replacement for mutation events. Adam Klein did all the heavy lifting and yours truly cleaned it up a bit. An introduction to DOM events has been added as well.

Posted in Weekly Review | Comments Off on WHATWG Weekly: Unicode for the platform?