.table-responsive regression

[alecpl]

After #23665 I see one problem. A table with table table-responsive class does not fit the screen width on a big screen. Previously the display:block was added only on small widths, now it applies always.

For your consideration if it's worth to change or it's as expected, but note it's a regression.
I'll be fine with any decision.

[pat270]

I think this bug appeared when .table-responsive was moved to the table element. The quickest fix is to go back to the way Bootstrap 3 did it and wrap the table in <div class="table-responsive"></div>.

[andresgalante]

Hey @alecpl Thanks a lot for reporting this.

Responsive tables are extremely hard to solve on a component level since there is no universal solution for them. It all depends on the data the user has to display. I like to call them the "holy grail" of components 😄. Having said that I don't see any usecase that can benefit from the docs introduced on #23665.

To me it generates unclear documentation and comments. Duplicates width: 100% that's already defined on.table and uses max-width media queries which aren't ideal on a mobile first framework.

While reviewing this issue I also noticed 2 things: Our docs have a warning about overflow-y for .table-responsive that doesn't exist on the code. And we are setting width: 100%; max-width: 100%; on .table which can probably be just min-width: 100%.

@mdo @XhmikosR What do you guys think? If there is a really good reason to keep the table-reponsive-* classes I am more than happy to clean it up and write better docs. Otherwise I'd just remove them and review tables docs and scss.

[XhmikosR]

@andresgalante: I'd say this is purely @mdo's choice. If it were up to me I'd keep .table-responsive and just fix the docs.

[mdo]

Yup, let's any docs inaccuracies or shortcomings. We can also call out max-width vs min-width nearly everywhere else.

[mdo]

PR at #24438 improves docs for this by including demos and better explaining how they work.