Manual regression range:
Good Build: 50.0.2656.0 
Bad Build:  50.0.2658.0 

Narrow bisect:
https://chromium.googlesource.com/chromium/src/+log/45605a80083f18230d5359890f6b8735310b25af..e2376628c5fe171c364499a8f39c5d0f73e7fcfe?pretty=fuller&n=10

Suspecting: 376732 or 376737 ?
Kindly help to re-assign, if your changes are not cause for this issue.

Note: Will update Linux info soon

Deleted: Actual_Reload.mov 2.4 MB

Actual_Reload.mov 2.4 MB Download

Deleted: Expected_Reload.mov 3.9 MB

Expected_Reload.mov 3.9 MB Download

Adding release block label, please undo if not the case. Also seeing this on Ubuntu OS

Can reproduce. See Issue 500657 for a similar issue on Android.

So what I can observe right now:
- on Linux M49 reloading causes the New Tab page to appear briefly before showing the error/warning page.
- on Linux build from trunk the New Tab page appears briefly before showing the error/warning page.

Since the time during which the new tab page is shown is very short, it is possible not to actually see it fully when doing the experiment manually.

This is because the interstitial page code has the old interstitial be hidden when navigating to a new interstitial, and until the new interstitial has been committed we are showing what was there before (ie the New Tab page). The code doing that has been in place since 2010, so this is not a regression.

Considering that interstitials are due for a refactoring (see issue 392354), reassigning to the owner.

felt@: I was able to reproduce an actual "navigation" to the new tab page yesterday.

That is, the same as "Actual_Reload.mov". Is that an expected regression?

lgarron, what did you have to do to repro? a few refreshes?