Falsely detecting empty spacer columns

[stephtr]

Empty columns can be used for styling, like <div class="col-xs-1 col"></div> with col defining a border or background-color. In this case W009's are detected even though the "empty" column can't be replaced.

[cvrebert]

Seems like you should typically be able to accomplish the same thing without spacer columns by moving some styles from the column to the row. What's your real-world example where spacer columns seem necessary?

Also, W009 is a Warning rather than an Error precisely because it may have false-positives.

[stephtr]

For example http://codepen.io/stephtr/full/grOmQe/
A real-world example would be if you include colored blocks into something like http://itsashapechristmas.co.uk/.

I reported it mainly because the twbs-lmvtfy bot treats these warnings as errors, but this is intentionally (cvrebert/lmvtfy#91).

[cvrebert]

@stephtr The links in your comment are broken. Markdown links can be written as either just the raw URL, or as [link text](http://your.url.goes/here/)

[cvrebert]

X-Ref: twbs/bootstrap#11130

[stephtr]

I corrected the links (http://codepen.io/stephtr/full/grOmQe/ and http://itsashapechristmas.co.uk/ ).

[cvrebert]

http://itsashapechristmas.co.uk/

I'm not getting any W009 errors on that page:

[stephtr]

That website does not even use Bootstrap (it was intended to show where empty columns could be useful). I meant if someone would build a website like that with Bootstrap and insert colored spaces then there would be a W009.

My suggestion would be to check not only if there are col-..-.. classes mentioned in the html code and if the element has no content but also if there are no custom classes (or classes other than col-..-..) mentioned.
For example <div class="col-xs-6 col-md-1"></div> would be a W009 rather sure, but something like <div class="col-xs-1 red-square"></div> is propably a false-positive.

[cvrebert]

probably a false-positive.

Eh, it's an imperfect heuristic either way. What if we just tweaked the warning message text to clarify that it's uncertain?
As for the bot, I reopened cvrebert/lmvtfy#91

[stephtr]

Another option would be to ignore that border case and leave the warning message as it is (there probably are much more true-positives than false-positives).