Fix DCHECK.

Fix DCHECK.

SuggestExactInput() was being called with the wrong value for |trim_http|.  We
want to trim HTTP only when the user hasn't explicitly typed it, but instead we
were trying to trim it only _if_ the user explicitly typed it.

This has been wrong since this code was introduced in
https://codereview.chromium.org/58553003/ .

BUG=603839
TEST=In a debug build, use the omnibox to navigate to "http://ab cd" (no quotes).  This should not DCHECK.

Committed: https://crrev.com/25a43970447cb8760dde96b9ec6875371dbd07f8
Cr-Commit-Position: refs/heads/master@{#399014}

[Peter Kasting, 2016-06-08 01:21:33]

pkasting@chromium.org changed reviewers:
+ mpearson@chromium.org

[Peter Kasting, 2016-06-08 01:21:34]

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
File components/omnibox/browser/omnibox_controller.cc (right):

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
components/omnibox/browser/omnibox_controller.cc:43: const bool was_open =
popup_ && popup_->IsOpen();
These changes avoid a segfault when reaching this from test code.

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
File components/omnibox/browser/omnibox_edit_model.cc (right):

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
components/omnibox/browser/omnibox_edit_model.cc:762: TemplateURLService*
template_url_service = client_->GetTemplateURLService();
Another change to avoid segfaulting in test code.

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
File components/omnibox/browser/omnibox_edit_unittest.cc (right):

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
components/omnibox/browser/omnibox_edit_unittest.cc:213: 
Will remove this blank line before landing

[Mark P, 2016-06-08 20:35:57]

generally lgtm, with minor comments below

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
File components/omnibox/browser/omnibox_edit_unittest.cc (right):

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
components/omnibox/browser/omnibox_edit_unittest.cc:213: 
On 2016/06/08 01:21:34, Peter Kasting wrote:
> Will remove this blank line before landing

Remember me.

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
components/omnibox/browser/omnibox_edit_unittest.cc:382: TEST_F(OmniboxEditTest,
HasHTTP) {
nit: AlternateNavHasHttp?
Otherwise this test, with lack of comments, and lack of the context I know from
reading the changelist description and the bug, is pretty obscure.

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
components/omnibox/browser/omnibox_edit_unittest.cc:387:
AutocompleteMatchType::URL_WHAT_YOU_TYPED);
For realism, can you please make this a search provider match for
query-what-you-typed?  After all, that's what the typical case is / the case
mentioned in the bug.

[Peter Kasting, 2016-06-08 20:52:24]

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
File components/omnibox/browser/omnibox_edit_unittest.cc (right):

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
components/omnibox/browser/omnibox_edit_unittest.cc:382: TEST_F(OmniboxEditTest,
HasHTTP) {
On 2016/06/08 20:35:57, Mark P wrote:
> nit: AlternateNavHasHttp?
> Otherwise this test, with lack of comments, and lack of the context I know
from
> reading the changelist description and the bug, is pretty obscure.

Honestly, I need to figure out better whether the param I'm changing can have
any visible effect; maybe I should just always set it to false.

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
components/omnibox/browser/omnibox_edit_unittest.cc:387:
AutocompleteMatchType::URL_WHAT_YOU_TYPED);
On 2016/06/08 20:35:57, Mark P wrote:
> For realism, can you please make this a search provider match for
> query-what-you-typed?  After all, that's what the typical case is / the case
> mentioned in the bug.

Good idea.

[Peter Kasting, 2016-06-09 19:26:56]

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
File components/omnibox/browser/omnibox_edit_unittest.cc (right):

https://codereview.chromium.org/2046263002/diff/1/components/omnibox/browser/...
components/omnibox/browser/omnibox_edit_unittest.cc:382: TEST_F(OmniboxEditTest,
HasHTTP) {
On 2016/06/08 20:52:24, Peter Kasting wrote:
> On 2016/06/08 20:35:57, Mark P wrote:
> > nit: AlternateNavHasHttp?
> > Otherwise this test, with lack of comments, and lack of the context I know
> from
> > reading the changelist description and the bug, is pretty obscure.
> 
> Honestly, I need to figure out better whether the param I'm changing can have
> any visible effect; maybe I should just always set it to false.

Turns out I can indeed just set it to false.  Did that, updated this test, and
added more comments to make the purpose clear.

[Mark P, 2016-06-09 20:04:26]

still lgtm

[Peter Kasting, 2016-06-09 20:42:57]

The CQ bit was checked by pkasting@chromium.org

[commit-bot: I haz the power, 2016-06-09 20:43:35]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2046263002/20001

[commit-bot: I haz the power, 2016-06-09 21:11:37]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-06-09 21:12:01]

CQ bit was unchecked

[commit-bot: I haz the power, 2016-06-09 21:13:14]

Description was changed from

==========
Fix DCHECK.

SuggestExactInput() was being called with the wrong value for |trim_http|.  We
want to trim HTTP only when the user hasn't explicitly typed it, but instead we
were trying to trim it only _if_ the user explicitly typed it.

This has been wrong since this code was introduced in
https://codereview.chromium.org/58553003/ .

BUG=603839
TEST=In a debug build, use the omnibox to navigate to "http://ab cd" (no
quotes).  This should not DCHECK.
==========

to

==========
Fix DCHECK.

SuggestExactInput() was being called with the wrong value for |trim_http|.  We
want to trim HTTP only when the user hasn't explicitly typed it, but instead we
were trying to trim it only _if_ the user explicitly typed it.

This has been wrong since this code was introduced in
https://codereview.chromium.org/58553003/ .

BUG=603839
TEST=In a debug build, use the omnibox to navigate to "http://ab cd" (no
quotes).  This should not DCHECK.

Committed: https://crrev.com/25a43970447cb8760dde96b9ec6875371dbd07f8
Cr-Commit-Position: refs/heads/master@{#399014}
==========

[commit-bot: I haz the power, 2016-06-09 21:13:15]

Patchset 2 (id:??) landed as
https://crrev.com/25a43970447cb8760dde96b9ec6875371dbd07f8
Cr-Commit-Position: refs/heads/master@{#399014}