1. sounds like something that should be fixed
2. I can look into this but as pointed out, this seems likely to be impossible to fix perfectly. The attached screenshot doesn't look horrific to me.

last oddity is  bug 608976 

Yeah, we may not be able to do better with the spacing, but I'm hoping either there's always a pixel of padding somewhere we can count on having, or some other thing that's fixable :/

I can't repro the green looking glass, but I did notice that we were using two different glasses for search and keyword search, which should probably be the same now that there's no visible chip.

Make sure you're getting the "remote" NTP?  I'm betting this has to do with how, under the hood, the NTP is HTTPS, and we're just fetching the color for "HTTPS" without special-casing the color in whatever the same special cases are where we set the icon to the glass and the omnibox to empty.

You're right that we should combine those other glasses into one.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/7d14a9887968d5e4ff4710339b4848c39802ee6a

commit 7d14a9887968d5e4ff4710339b4848c39802ee6a
Author: estade <estade@chromium.org>
Date: Wed Jun 01 03:06:28 2016

Change selected keyword search icon to the same icon used for default
omnibox search.

BUG= 616191 
TEST=looking glass should be the same size when tab to search is activated
as when the omnibox is being used for a default search engine search

Review-Url: https://codereview.chromium.org/2023013003
Cr-Commit-Position: refs/heads/master@{#397036}

[modify] https://crrev.com/7d14a9887968d5e4ff4710339b4848c39802ee6a/chrome/browser/ui/views/location_bar/selected_keyword_view.cc
[modify] https://crrev.com/7d14a9887968d5e4ff4710339b4848c39802ee6a/ui/gfx/BUILD.gn
[delete] https://crrev.com/b26313ef9d24c94f72c87f450e54999404e55e51/ui/gfx/vector_icons/keyword_search.1x.icon
[delete] https://crrev.com/b26313ef9d24c94f72c87f450e54999404e55e51/ui/gfx/vector_icons/keyword_search.icon

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/9aba6c9ce7bf118e44acd4158b5597cf9f91fd24

commit 9aba6c9ce7bf118e44acd4158b5597cf9f91fd24
Author: estade <estade@chromium.org>
Date: Fri Jun 03 21:22:35 2016

Fix color of looking glass when searching on remote NTP.

BUG= 616191 

Review-Url: https://codereview.chromium.org/2028933002
Cr-Commit-Position: refs/heads/master@{#397805}

[modify] https://crrev.com/9aba6c9ce7bf118e44acd4158b5597cf9f91fd24/chrome/browser/ui/views/location_bar/location_bar_view.cc
[modify] https://crrev.com/9aba6c9ce7bf118e44acd4158b5597cf9f91fd24/components/toolbar/toolbar_model_impl.cc

Can r397805 be merged into M52? The color of the looking glass is incorrect there

Actually sorry, forget my comment on #8, I can no longer reproduce it on M52

I have a fix for the layout issues (comment 0 item 2).

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/f08f7b065f97fcfd408743f2158ab08af49679ef

commit f08f7b065f97fcfd408743f2158ab08af49679ef
Author: pkasting <pkasting@chromium.org>
Date: Wed Jun 15 23:57:43 2016

Adjust omnibox and dropdown text position to be correct.

There were a variety of issues here.  It appears the textfield does not actually
put 1 px before the edit text (verified using wide letters like 'w' and
scrolling the field until they were clipped on the right).  The RTL-only pre-MD
code (which I wrote long ago) was incorrect too, AFAICT; it's just misleading
because the edit reserves space for the cursor on the right side.  (Maybe the
textfield has changed since I wrote that, or it dated from the CRichEditCtrl
days.  I don't remember.)

The trailing padding incorrectly subtracted the edge thickness.  This wasn't
previously visible on MD (see next paragraph), but on non-MD it brought the end
of the text closer to the visible edge.  But if we wanted that effect, we should
have made the trailing padding e.g. "edge thickness + 1" instead of
"item padding - edge thickness".  With the larger item padding in MD this was
that much weirder.  Changing this left the same padding on the right side of the
omnibox as everywhere else in the location bar.

The trailing padding on MD was set to 0 for no obvious reason (this dates back
about a year).  This led to clipping the right edge on higher scale factors.
Using the normal calculation fixed this.

The dropdown icon width was calculated using the pre-MD raster icons even for
MD.  This was fixed to use the same size the location bar uses by means of a new
oracle function on that class.  This moved the dropdown text closer to the icon.

BUG= 616191 
TEST=Type in the omnibox; the text in the dropdown should line up horizontally with the text in the omnibox.

Review-Url: https://codereview.chromium.org/2070133004
Cr-Commit-Position: refs/heads/master@{#400049}

[modify] https://crrev.com/f08f7b065f97fcfd408743f2158ab08af49679ef/chrome/browser/ui/views/location_bar/location_bar_view.cc
[modify] https://crrev.com/f08f7b065f97fcfd408743f2158ab08af49679ef/chrome/browser/ui/views/location_bar/location_bar_view.h
[modify] https://crrev.com/f08f7b065f97fcfd408743f2158ab08af49679ef/chrome/browser/ui/views/omnibox/omnibox_result_view.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/a4123685765478a14349e959be9ca83545b4c06c

commit a4123685765478a14349e959be9ca83545b4c06c
Author: pkasting <pkasting@chromium.org>
Date: Thu Jun 16 05:52:26 2016

Remove IconLabelView layout constants.

With the new IconLabelView design, we just want to use
LOCATION_BAR_HORIZONTAL_PADDING everywhere, except for the trailing padding in
non-MD, which I'm making a class-scoped constant for now (we can delete it
entirely when non-MD goes away).

This increases the internal spacing in MD from 5 to 6, which is correct for the
new style.

BUG= 616191 
TEST=none

Review-Url: https://codereview.chromium.org/2075443002
Cr-Commit-Position: refs/heads/master@{#400092}

[modify] https://crrev.com/a4123685765478a14349e959be9ca83545b4c06c/chrome/browser/ui/layout_constants.cc
[modify] https://crrev.com/a4123685765478a14349e959be9ca83545b4c06c/chrome/browser/ui/layout_constants.h
[modify] https://crrev.com/a4123685765478a14349e959be9ca83545b4c06c/chrome/browser/ui/views/location_bar/icon_label_bubble_view.cc
[modify] https://crrev.com/a4123685765478a14349e959be9ca83545b4c06c/chrome/browser/ui/views/location_bar/icon_label_bubble_view.h
[modify] https://crrev.com/a4123685765478a14349e959be9ca83545b4c06c/chrome/browser/ui/views/location_bar/keyword_hint_view.cc
[modify] https://crrev.com/a4123685765478a14349e959be9ca83545b4c06c/chrome/browser/ui/views/omnibox/omnibox_result_view.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/21db7774b3bae9eec33c7e645672e9aa73d13c4d

commit 21db7774b3bae9eec33c7e645672e9aa73d13c4d
Author: pkasting <pkasting@chromium.org>
Date: Thu Jun 16 16:16:40 2016

Change IconLabelBubbleView layout to specify padding beside the separator.

Specifying the total extra padding as before was confusing, because the value
inherently depended on the external spacing after the view.  Going this route
means that no matter how the spacing after the view changes, we won't need to
change the padding constant in sync.

This changes the net padding on either side of the separator from 9 DIP to 8
DIP, which matches the mocks.  (Note that until the omnibox text position is
tweaked, in another CL to come, it may not appear to be 8 DIP away.)

BUG= 616191 
TEST=none

Review-Url: https://codereview.chromium.org/2064323005
Cr-Commit-Position: refs/heads/master@{#400165}

[modify] https://crrev.com/21db7774b3bae9eec33c7e645672e9aa73d13c4d/chrome/browser/ui/views/location_bar/icon_label_bubble_view.cc
[modify] https://crrev.com/21db7774b3bae9eec33c7e645672e9aa73d13c4d/chrome/browser/ui/views/location_bar/icon_label_bubble_view.h