Fixes #2576 by saving all attributes on the img tag.

[TimDix]

No description provided.

[joostrijneveld]

Note that, for an empty attrString, this results in two consecutive spaces. This is why the Travis build is failing.

[TimDix]

Thanks, tweaking now.

[TimDix]

Alright, this latest 'fail' I think is the result of a poorly written test. The test asserts that

array('Test<img src="http://www.dummyco.com/path/to/server/index.php/download_file/view_inline/1">', 'Test<concrete-picture fID="1" alt="" style="" />'),

This is failing because the new code doesn't arbitrarily add empty attributes. I'd propose changing to to:

array('Test<img src="http://www.dummyco.com/path/to/server/index.php/download_file/view_inline/1">', 'Test<concrete-picture fID="1" />'),
array('Test<img src="http://www.dummyco.com/path/to/server/index.php/download_file/view_inline/1" alt="Woohoo" style="display: block" />', 'Test<concrete-picture fID="1" alt="Woohoo" style="display: block" />'),

[TimDix]

Alternatively, if there's a good reason why the empty attributes should be there, I can force those to always output.

[MrKarlDilkington]

👍

I tested the code and it applies the attributes to the img tag on save as expected.

Regarding any confusion relating to the placement of attributes.

http://www.w3.org/html/wg/drafts/html/master/semantics.html#the-picture-element

"While all of them contain source elements, the sourceelement's src attribute has no meaning when the element is nested within a picture element, and the resource selection algorithm is different. As well, thepicture element itself does not display anything; it merely provides a context for its contained img element that enables it to choose from multiple URLs"

[aembler]

No, there's no need to output the empty attributes. We probably just need to update the test so that it passes and then we'll get the pull request applied.

[MrKarlDilkington]

@aembler

Will this fix be part of the next release?

[aembler]

Thanks!