923913 - Make Element.innerHTML [Pure]?

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Olli Pettay [:smaug][[e-mail]]]

bz had the idea to make innerHTML (getter) [Pure].
Trying to figure out whether that is actually possible.

Comment 1

[Olli Pettay [:smaug][[e-mail]]]

Attachment: WIP

https://tbpl.mozilla.org/?tree=Try&rev=658096800cd8

Comment 2

[Olli Pettay [:smaug][[e-mail]]]

Attachment: test

On my machine time goes from ~235ms to ~8ms

Comment 3

[Olli Pettay [:smaug][[e-mail]]]

hmm, perhaps .innerHTML may throw but .outerHTML doesn't.
Will test and figure out why we have the code paths we have.

Comment 4

[Olli Pettay [:smaug][[e-mail]]]

Or vice versa, outerHTML may, but innerHTML may not... Anyhow, will try to investigate tomorrow.

Comment 5

[Olli Pettay [:smaug][[e-mail]]]

So we do serialize for example comment which has -- in the data.

Comment 6

[Olli Pettay [:smaug][[e-mail]]]

And so does blink.

Comment 7

[Olli Pettay [:smaug][[e-mail]]]

I haven't managed to make any browser throw from innerHTML getter
http://mozilla.pettay.fi/moztests/innerhtml-01.xhtml (taken from https://github.com/whatwg/domparsing/blob/master/tests/innerhtml-01.xhtml)
and http://mozilla.pettay.fi/moztests/xhtml_innerHTML_outerHTML.xhtml

Comment 8

[Olli Pettay [:smaug][[e-mail]]]

Attachment: patch

https://tbpl.mozilla.org/?tree=Try&rev=9eb5ac196105

Comment 9

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 814445 [details] [diff] [review]
patch

Make GetMarkup return void, and r=me

Comment 10

[Olli Pettay [:smaug][[e-mail]]]

Attachment: v2

Comment 11

[Olli Pettay [:smaug][[e-mail]]]

https://tbpl.mozilla.org/?tree=Try&rev=dfdca5ae83fe

Comment 12

[Olli Pettay [:smaug][[e-mail]]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/772be3f89d8e

Comment 13

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/772be3f89d8e

Comment 14

[Jean-Yves Perrier [:teoli]]

In order to know if we need to update documentation or not, I have one question:
Does this change the behavior users see? Or is it just adding semantics acknowledging that innerHTML will never throw?

Comment 15

[Olli Pettay [:smaug][[e-mail]]]

This should not change the semantics. (But the spec may change to say empty string is returned in
case of XML document and non-serializable DOM.)