Initial switch to Chromium-style CHECK_* and DCHECK_* macros.

Initial switch to Chromium-style CHECK_* and DCHECK_* macros.

R=svenpanne@chromium.org

Committed: https://crrev.com/0deaa4b629faeae6832046c294dc7c034a0a6653
Cr-Commit-Position: refs/heads/master@{#26340}

[Benedikt Meurer, 2015-01-29 16:22:45]

Sven: PTAL
Jochen, Dan: FYI

[Benedikt Meurer, 2015-01-29 16:35:59]

bmeurer@chromium.org changed reviewers:
+ jarin@chromium.org

[Jarin, 2015-01-29 16:59:27]

It is almost there. Two things I do not like:

- please, do not use pointers as bools. For DCHECK(x), x should be bool.

- the checks for equality pf doubles are messy. Could we make it uniform?
(Ideally, CHECK_EQ(d1, d2) should work.)

Also, change of this size deserves an email to the team; I am sure it will
require some rebasing work from everyone.

[Jarin, 2015-01-29 16:59:30]

It is almost there. Two things I do not like:

- please, do not use pointers as bools. For DCHECK(x), x should be bool.

- the checks for equality pf doubles are messy. Could we make it uniform?
(Ideally, CHECK_EQ(d1, d2) should work.)

Also, change of this size deserves an email to the team; I am sure it will
require some rebasing work from everyone.

[Sven Panne, 2015-01-29 17:07:04]

LGTM when the CHECK vs. DCHECK issue is addressed and ASAN is happy, too. ;-)

https://codereview.chromium.org/888613002/diff/20001/src/checks.cc
File src/checks.cc (right):

https://codereview.chromium.org/888613002/diff/20001/src/checks.cc#newcode8
src/checks.cc:8: namespace internal {}  // namespace internal
As discussed offline, this file should die in a separate CL.

https://codereview.chromium.org/888613002/diff/20001/src/jsregexp.cc
File src/jsregexp.cc (right):

https://codereview.chromium.org/888613002/diff/20001/src/jsregexp.cc#newcode5601
src/jsregexp.cc:5601: CHECK(tree()->Insert(current.from(), &loc));
Here and at other places: Shouldn't this be a DCHECK instead of a CHECK? Or do
we intentionally want different behavior here?

[Sven Panne, 2015-01-29 17:11:02]

On 2015/01/29 16:59:27, jarin wrote:
> It is almost there. Two things I do not like:
> 
> - please, do not use pointers as bools. For DCHECK(x), x should be bool.
> [...]

This is Chrome-style, and unless there are strong reasons ("I don't like that"
doesn't count as strong ;-), we should follow that. I think the reasoning behind
this is: Adding useless "!= NULL" just adds noise, and allowing this makes
things like

   if (Foo* f = GetMyCoolFoo()) {
      ... <use f here>
   }

possible, which is A Good Thing (tm), because it keeps scopes small.

[Benedikt Meurer, 2015-01-30 05:29:32]

I'll put back the DCHECK_NOT_NULL

https://codereview.chromium.org/888613002/diff/20001/src/jsregexp.cc
File src/jsregexp.cc (right):

https://codereview.chromium.org/888613002/diff/20001/src/jsregexp.cc#newcode5601
src/jsregexp.cc:5601: CHECK(tree()->Insert(current.from(), &loc));
DCHECK_RESULT has slightly different semantics. Restored that beast for now.

[Benedikt Meurer, 2015-01-30 06:09:13]

> - the checks for equality pf doubles are messy. Could we make it uniform?
(Ideally, CHECK_EQ(d1, d2) should work.)

As per offline discussion: CheckDoubleEq is a temporary hack that will be
removed once these cctests are converted to GTest/GMock.

[Benedikt Meurer, 2015-01-30 06:09:27]

The CQ bit was checked by bmeurer@chromium.org

[commit-bot: I haz the power, 2015-01-30 06:09:42]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/888613002/60001

[Benedikt Meurer, 2015-01-30 06:20:35]

The CQ bit was unchecked by bmeurer@chromium.org

[Benedikt Meurer, 2015-01-30 06:21:05]

The CQ bit was checked by bmeurer@chromium.org

[commit-bot: I haz the power, 2015-01-30 06:21:53]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/888613002/60001

[commit-bot: I haz the power, 2015-01-30 06:25:22]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2015-01-30 06:25:48]

Patchset 4 (id:??) landed as
https://crrev.com/0deaa4b629faeae6832046c294dc7c034a0a6653
Cr-Commit-Position: refs/heads/master@{#26340}