Tweaks for promo messages on NTP

Tweaks and fixes for NTP extension promo.

Don't show promo for extensions on Mac, as they're not available yet. Decrement promo counter even when NTP is not home page, and do not remove puzzle piece when promo line is closed. When message bar is closed, make it zip off bottom of screen instead of just disappearing.

BUG=27814, 25258, 27815
TEST= see various bugs.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=32176

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=32413

[Miranda Callahan, 2009-11-16 16:16:25]

On 2009/11/16 16:13:45, Miranda Callahan wrote:
>

[Miranda Callahan, 2009-11-16 16:59:39]

Aaron, I added the other two NTP promo bug fixes to this patch, as they're all
little tweaks.

[Aaron Boodman, 2009-11-16 17:41:11]

LGTM

You should get a confirmation on these changes from a UI lead though before
submitting.

[Miranda Callahan, 2009-11-16 22:13:32]

On 2009/11/16 17:41:11, Aaron Boodman wrote:
> LGTM
> 
> You should get a confirmation on these changes from a UI lead though before
> submitting.

Got LGTM from Glen on this change per email; bringing Erik in because of an
added webkit-transition also requested by Glen, and because some tweaks in this
code address changes Erik had suggested from a patch that was committed
yesterday.

[arv (Not doing code reviews), 2009-11-16 22:20:11]

LGTM with one nit

http://codereview.chromium.org/385135/diff/21/28
File chrome/browser/resources/new_new_tab.css (right):

http://codereview.chromium.org/385135/diff/21/28#newcode710
Line 710: margin-bottom: -40px;
why not bottom: 40px? It seems clearer to animate the bottom property than the
margin.

[Miranda Callahan, 2009-11-17 00:56:22]

Thanks for the help!

http://codereview.chromium.org/385135/diff/21/28
File chrome/browser/resources/new_new_tab.css (right):

http://codereview.chromium.org/385135/diff/21/28#newcode710
Line 710: margin-bottom: -40px;
On 2009/11/16 22:20:11, arv wrote:
> why not bottom: 40px? It seems clearer to animate the bottom property than the
> margin.

Done.

[jam, 2009-11-17 21:04:41]

On 2009/11/17 00:56:22, Miranda Callahan wrote:
> Thanks for the help!
> 
> http://codereview.chromium.org/385135/diff/21/28
> File chrome/browser/resources/new_new_tab.css (right):
> 
> http://codereview.chromium.org/385135/diff/21/28#newcode710
> Line 710: margin-bottom: -40px;
> On 2009/11/16 22:20:11, arv wrote:
> > why not bottom: 40px? It seems clearer to animate the bottom property than
the
> > margin.
> 
> Done.

Looks like this change had a pretty noticeable negative impact on the new tab
warm tests.  Is this known/accepted?  see

http://build.chromium.org/buildbot/perf/xp-release-dual-core/new-tab-ui-warm/...
http://build.chromium.org/buildbot/perf/xp-release-single-core/new-tab-ui-war...
http://build.chromium.org/buildbot/perf/vista-release-dual-core/new-tab-ui-wa...
http://build.chromium.org/buildbot/perf/vista-release-single-core/new-tab-ui-...

[cmp_google, 2009-11-17 21:11:11]

It would be great to run a test on the reference build for new tab ui warm,
too.  I filed http://crbug.com/28028 to track setting this up.

On Tue, Nov 17, 2009 at 1:04 PM, <jam@chromium.org> wrote:

> On 2009/11/17 00:56:22, Miranda Callahan wrote:
>
>> Thanks for the help!
>>
>
>  http://codereview.chromium.org/385135/diff/21/28
>> File chrome/browser/resources/new_new_tab.css (right):
>>
>
>  http://codereview.chromium.org/385135/diff/21/28#newcode710
>> Line 710: margin-bottom: -40px;
>> On 2009/11/16 22:20:11, arv wrote:
>> > why not bottom: 40px? It seems clearer to animate the bottom property
>> than
>>
> the
>
>> > margin.
>>
>
>  Done.
>>
>
> Looks like this change had a pretty noticeable negative impact on the new
> tab
> warm tests.  Is this known/accepted?  see
>
>
>
http://build.chromium.org/buildbot/perf/xp-release-dual-core/new-tab-ui-warm/...
>
>
http://build.chromium.org/buildbot/perf/xp-release-single-core/new-tab-ui-war...
>
>
http://build.chromium.org/buildbot/perf/vista-release-dual-core/new-tab-ui-wa...
>
>
http://build.chromium.org/buildbot/perf/vista-release-single-core/new-tab-ui-...
>
>
> http://codereview.chromium.org/385135
>

[Miranda Callahan, 2009-11-18 03:18:45]

On 2009/11/17 21:11:11, cmp_google.com wrote:
> It would be great to run a test on the reference build for new tab ui warm,
> too.  I filed http://crbug.com/28028 to track setting this up.
> 
> On Tue, Nov 17, 2009 at 1:04 PM, <mailto:jam@chromium.org> wrote:
> 
> > On 2009/11/17 00:56:22, Miranda Callahan wrote:
> >
> >> Thanks for the help!
> >>
> >
> >  http://codereview.chromium.org/385135/diff/21/28
> >> File chrome/browser/resources/new_new_tab.css (right):
> >>
> >
> >  http://codereview.chromium.org/385135/diff/21/28#newcode710
> >> Line 710: margin-bottom: -40px;
> >> On 2009/11/16 22:20:11, arv wrote:
> >> > why not bottom: 40px? It seems clearer to animate the bottom property
> >> than
> >>
> > the
> >
> >> > margin.
> >>
> >
> >  Done.
> >>
> >
> > Looks like this change had a pretty noticeable negative impact on the new
> > tab
> > warm tests.  Is this known/accepted?  see
> >
> >
> >
>
http://build.chromium.org/buildbot/perf/xp-release-dual-core/new-tab-ui-warm/...
> >
> >
>
http://build.chromium.org/buildbot/perf/xp-release-single-core/new-tab-ui-war...
> >
> >
>
http://build.chromium.org/buildbot/perf/vista-release-dual-core/new-tab-ui-wa...
> >
> >
>
http://build.chromium.org/buildbot/perf/vista-release-single-core/new-tab-ui-...
> >
> >
> > http://codereview.chromium.org/385135
> >
> 

Looking into this more deeply -- but does anyone have an idea what could have
caused the regression at r32170?  There are no commits listed, yet it spiked
>10ms; after my reversion, it's still back at that plateau, ~12ms above the
baseline...

[Miranda Callahan, 2009-11-18 03:21:36]

On 2009/11/18 03:18:45, Miranda Callahan wrote:
> On 2009/11/17 21:11:11, http://cmp_google.com wrote:
> > It would be great to run a test on the reference build for new tab ui warm,
> > too.  I filed http://crbug.com/28028 to track setting this up.
> > 
> > On Tue, Nov 17, 2009 at 1:04 PM, <mailto:jam@chromium.org> wrote:
> > 
> > > On 2009/11/17 00:56:22, Miranda Callahan wrote:
> > >
> > >> Thanks for the help!
> > >>
> > >
> > >  http://codereview.chromium.org/385135/diff/21/28
> > >> File chrome/browser/resources/new_new_tab.css (right):
> > >>
> > >
> > >  http://codereview.chromium.org/385135/diff/21/28#newcode710
> > >> Line 710: margin-bottom: -40px;
> > >> On 2009/11/16 22:20:11, arv wrote:
> > >> > why not bottom: 40px? It seems clearer to animate the bottom property
> > >> than
> > >>
> > > the
> > >
> > >> > margin.
> > >>
> > >
> > >  Done.
> > >>
> > >
> > > Looks like this change had a pretty noticeable negative impact on the new
> > > tab
> > > warm tests.  Is this known/accepted?  see
> > >
> > >
> > >
> >
>
http://build.chromium.org/buildbot/perf/xp-release-dual-core/new-tab-ui-warm/...
> > >
> > >
> >
>
http://build.chromium.org/buildbot/perf/xp-release-single-core/new-tab-ui-war...
> > >
> > >
> >
>
http://build.chromium.org/buildbot/perf/vista-release-dual-core/new-tab-ui-wa...
> > >
> > >
> >
>
http://build.chromium.org/buildbot/perf/vista-release-single-core/new-tab-ui-...
> > >
> > >
> > > http://codereview.chromium.org/385135
> > >
> > 
> 
> Looking into this more deeply -- but does anyone have an idea what could have
> caused the regression at r32170?  There are no commits listed, yet it spiked
> >10ms; after my reversion, it's still back at that plateau, ~12ms above the
> baseline...

(seems to only be happening on
http://build.chromium.org/buildbot/perf/xp-release-dual-core/new-tab-ui-warm/...)

[Miranda Callahan, 2009-11-18 18:52:14]

Eradicated the source of ntp test slowdown: it was the localized strings being
set for "firstview" (which is never used) in new_tab_ui.cc.  The removal of that
single line is the only difference between this patch set and the previous one,
and on my machine eliminates the 20% startup regression.

[Miranda Callahan, 2009-11-18 18:53:03]

+tony

On 2009/11/18 18:52:14, Miranda Callahan wrote:
> Eradicated the source of ntp test slowdown: it was the localized strings being
> set for "firstview" (which is never used) in new_tab_ui.cc.  The removal of
that
> single line is the only difference between this patch set and the previous
one,
> and on my machine eliminates the 20% startup regression.