You can either use a <div> or <span> and just be careful not to style it, or set a class on all the elements you want and use document.getElementsByClassName() or document.querySelectorAll().

We are using DIVs now, one problem with this method is the content is arbitrary, sometimes containing only block elements, sometimes only inline, sometimes both. So this complicates things in that the CMS needs to decide on what sort of wrapper to use, block or inline. Also the group or wrapper elements can be accessed by designers through CSS which is not desirable, we want them to style the content based on the semantics of the content, not the CMS.

Adding classes and using document.getElementsByClassName() or similar works but is not without issues, it adds to the file size when pages are complex with multiple widgets being used, and makes replacing the nodes with the updated content a little messier, although it's probably the way we'll go as these issues are easier to deal with than the issues mentioned above.

Many web-sites are now data driven, and constructed by ever more sophisticated CMS, it'd be nice to help CMS developers create clean standards based semantic markup, so I would've thought that it's worth considering.



Thanks for your time anyway.

Regards

Andreas Baitis

We can't have an element that's in the DOM but has no style.  That would require significant changes to all sorts of things, which just aren't worth it.

Would I be crazy to suggest that <span> just be allowed to contain block elements?  It has the correct visual behavior, right?  Why is it invalid as a block wrapper to start with?

(In reply to comment #3)
> Would I be crazy to suggest that <span> just be allowed to contain block
> elements?  It has the correct visual behavior, right?  Why is it invalid as a
> block wrapper to start with?

Its default display value interacts badly with block descendants, for one.

How so?

Inlines that contain blocks are split around the blocks into multiple inline boxes.

And this differs how from the desired behavior here?  In what case would $foo and <span>$foo</span> render differently, assuming that $foo is a snippet of HTML markup in the body with all tags properly closed?  (You do have to make sure to properly close tags, because </span> doesn't auto-close <p> and such.)

If you had any styles defined for "span", like borders, you'd get some unexpected rendering.

The use case according to comment 0 is that there are no styles defined.

(In reply to comment #9)
> The use case according to comment 0 is that there are no styles defined.

The use case is that sometimes designers will style the SPAN or DIV because they can, hence the request for the GROUP tag which can not be styled, it's sole purpose would be to group it's children.

That is not going to happen.  All elements are styleable, no way are we going to change that without *extremely* good reason.

(In reply to comment #11)
> That is not going to happen.  All elements are styleable, no way are we going
> to change that without *extremely* good reason.

OK thanks for everyones time anyway, I had to ask of course :)

If you would like to withdraw the request, please mark the bug RESOLVED INVALID.  That way the editor won't have to spend his time reading it, and can read other people's bugs instead.

(In reply to comment #13)
> If you would like to withdraw the request, please mark the bug RESOLVED
> INVALID.  That way the editor won't have to spend his time reading it, and can
> read other people's bugs instead.

mass-moved component to LC1