Tested the issue on Ubuntu 14.04, Windows 7, Mac 10.11.5 using 52.0.2743.75 as per above steps.Observed the same behavior in M-51(51.0.2704.106), M-52(52.0.2743.75), firefox.

Please find attached screenshot and update if anything missed here in triaging the issue.

liamenoc@Could you please provide actual and expected behavior screencast for further triaging the issue.

Deleted: 628927.png 995 KB

	628927.png 995 KB View Download

ssamanoori@ The issue described is below the big picture of the man. The two small pictures are aligned to be in columns, but since v. 52, they align on top of each other.

Here is a test: http://codepen.io/anon/pen/QEQpGy

I also attached two screenshots. One is Chromium 51, other 52.

Deleted: 52.png 230 KB

	52.png 230 KB View Download

Deleted: 51.png 225 KB

	51.png 225 KB View Download

More about columns rendering indifferently.

Using -webkit-column-break-inside: avoid behaves inconsistent when the number of elements inside the container varies.

See the differences between 51 and 52: http://codepen.io/jochemnabuurs/pen/ugDJB

https://css-tricks.com/almanac/properties/b/break-inside/#comment-1585497

v. 52 produces the same behaviour as Firefox bug: https://css-tricks.com/almanac/properties/b/break-inside/#li-comment-1602153

Meaning: The columns are buggy/inconsistent in chromium, but beta version, 52, is different from the past.

Thank you for providing more feedback. Adding requester "ssamanoori@chromium.org" for another review and adding "Needs-Review" label for tracking.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

We've recently made a number of changes to better support multicol and match the spec so I suspect this is Working as intended but I'll leave it open for Morten to confirm.

It is possible the new implementation follows the spec better, hence the "bugs", but bugs with break-inside: avoid when the number of elements inside the container varies remain - http://codepen.io/jochemnabuurs/pen/ugDJB

Response to comment #3:

This changed because we now implement the orphans and widows properties as mandated by the spec, i.e. setting their initial value to 2, instead of non-standard "auto", which was almost the same as setting them to 1. [1]

The test in comment #3 displays the multicol FIGURE children as inline-block, which means that the FIGURE elements end up on lines. And since there are only two lines, we shouldn't allow column breaks between them, because that would violate the orphans/widows requirements.

[1] See  bug 473509 

Response to comment #5:

Bottom margins end up in the next column in some cases. That's  bug 440362 . Additionally, we seem to have a problem with the column balancer here. If we insert the "right" number of LI elements, we over-stretch the multicol container (i.e. making it taller than necessary) (the wording in the test is ".example gets more bottom padding"). I've now reported bug 636055 separately for that problem.

I'm now closing this bug, since it was about too many different problems. If I didn't deal with everything properly, please file separate bugs - one bug per problem.