Add datetime attribute alternative following W3C recommendations

[RoiEXLab]

Picky, I know.
In your README.md you mentioned the possible use of a data-timeago attribute instead of the "W3C-illegal" datetime attribute (see W3C Recommendation) which we couldn't actually use before this PR.
This PR adds the possibility to use the data-timeago attribute for the automatic rendering while maintaining backwards compatibility.

[hustcc]

sorry for confused, that is a typo of readme. Maybe you can pr for this.

after 2.x.x, data-timeago was not supported, but 1.x.x supports.

[RoiEXLab]

This is a PR^^

[RoiEXLab]

A small question btw.: Is there any difference when using document.getElemensByClassName('yourclass') instead of document.querySelectorAll('.yourclass') ?

[likerRr]

@RoiEXLab http://stackoverflow.com/questions/14377590/queryselector-and-queryselectorall-vs-getelementsbyclassname-and-getelementbyid

[RoiEXLab]

@RoiEXLab http://stackoverflow.com/questions/14377590/queryselector-and-queryselectorall-vs-getelementsbyclassname-and-getelementbyid

I'll treat that as a "there is a difference, but if it works it doesn't matter"

Thanks

[hustcc]

I means in v1.x.x, timeago.js supportes data-timeago attribute. But after v2.x.x, data-timeago attribute do not be supported.

The data-timeago mentioned inreadme.md is a typo. I will fixed soon.

---

I think there are no differences between querySelectorAll against getElementsByClassName when using timeago.js.

[RoiEXLab]

@hustcc I figured. But I think having a data-timeago attribute as alternative is a good idea.
The changes in this PR cause timeago.js to use the data-timeago attribute if it's available, and fallback to the current behaviour using the datetime attribute if it's not available.

[likerRr]

@hustcc I agree with @RoiEXLab . By accepting the PR timeago.js will become more compatible with the W3C.

Moreover i think we should leave only behavior provided by this PR, without falling back to current custom attribute datetime.

[hustcc]

ok~ compatible with jquery maybe will be perfact.

if(node.dataset.timeago) return node.dataset.timeago; // js
if(node.data) return node.data['timeago']; //jq
if (node.getAttribute) return node.getAttribute(ATTR_DATETIME); // js
if(node.attr) return node.attr(ATTR_DATETIME); //jq

[likerRr]

@hustcc is it a joke? 😄 We should provide compatibility with standard, not with some library until it's a plugin for the library

[hustcc]

not jquery plugin, not jquery plugin, not jquery plugin. 😄😄😄

before the PR, we use it below:

<time datetime='2016-11-09 12:12:12'></time>

timeago.render(document.getElementsByTag('time'));
// or jq selector
timeago.render($(tag));

After be compatibility with standard by data-*, then we can use it as below:

<time data-timeago='2016-11-09 12:12:12'></time>

We can not select the render javascript DOM below:

timeago.render(document.getElementsByTag('time'));

Can not be used with jquery selector as the README mentioned.

I think compatibility with W3C standard is OK, but we should keep the consistent usage.

Pool english maybe can not express what I mean -_-!!

[likerRr]

@hustcc thx, I understand what you mean, but let me correct you a little.

After PR you will still be able to query elements by data-* attribute in plain javascript:

document.querySelectorAll('[data-timeago]')

and in jquery:

$('[data-timeago]')

[likerRr]

But.. hm.. honestly I don't see the difference between:

<time datetime='2016-11-09 12:12:12'></time>
// and
<time data-timeago='2016-11-09 12:12:12'></time>

They both are queried the same way:

document.getElementsByTag('time')

[RoiEXLab]

if(node.dataset.timeago) return node.dataset.timeago; // js
if(node.data) return node.data['timeago']; //jq

The problem I see with this snippet, is that if jQuery is supported and the data-timeago attribute is not set the method will return an empty string without checking if the datetime attribute is set.

Nonetheless I'd like to point out that @hutscc as owner of this repo has write access to my branch - allowing her to edit my PR without me

[likerRr]

@hustcc @RoiEXLab nice to see it merged, good job

[RoiEXLab]

@likerRr @hustcc thanks!

[hustcc]

😄

[RoiEXLab]

@hustcc I noticed, the dist folder isn't updated yet. I'd appreciate if you could rerun minifiers etc.

[hustcc]

done, include ro locale and this pr.