Confirmed, issue is not visible on Chrome 51.0.2662.0 but present on the current stable 52.0.2743.116. 

Able to reproduce the issue on windows 7, Linux Ubuntu 14.04 and Mac 10.11.6 using chrome version 52.0.2743.116 and 54.0.2838.0.
This is regression issue broken in M52.Please find the bisect information as below

Narrow Bisect::
Good :: 52.0.2715.0   ---  (build revision 388964)
Bad:: 52.0.2716.0   ---  (build revision 389538)

CHANGELOG URL:  https://chromium.googlesource.com/chromium/src/+log/723a4b536fa9a2f0071c510d14c36918cfee83a9..a74b8ccb63e527837449638b13e2c1534c40bddf

Unable to find the suspect from the above CL.Could any one from dev team please look into this issue.

Thanks,

Oh, I really like the sample app. It's pretty. Though it is a bit big, so there are lots of moving parts and it is not obvious why it's failing. I'll bisect a bit more.

Bisected down to https://chromium.googlesource.com/chromium/src/+log/dd5aacea530b6a32cb3fdf358093a091d600d08e..f0e5c5078762140f3f423681ec5e70babef4b24f

Probably r389061?

Yep, adding the following piece of CSS to the test makes the problem go away:

* { orphans:1; widows:1; }

The test case is too complicated for me to comment on any further (but I suspect it uses inline-blocks for things that SHOULD be blocks, so that you get tall lines, making it hard to satisfy the default orphans and widows requirements), but without a further reduction, I'll just mark it as a dup.

Thanks for your prompt responses everyone, I'll update our css to include that fix. You're right, we're probably heavily abusing CSS to make this work, and simplifying it down was difficult from our end too. It's trying to calculate the page layout to do a print preview in browser before the user gets to the _actual_ print preview screen.

Will pass onto our designers that the prettiness was appreciated, that'll make their day!

Thanks again,
Mathew