Fix clang compiler warning for lambda variable catpure

Fix clang compiler warning for lambda variable catpure

This is broken with a recent clang compiler.

This warning got introduced by clang (5.0.0) with this commit:
  https://reviews.llvm.org/D28467

Other recent similar fixes in chrome:
  https://codereview.chromium.org/2646553002/diff/20001/test/unittests/heap/slot-set-unittest.cc

BUG=

[etienneb, 2017-02-22 19:07:19]

Description was changed from

==========
Fix clang compiler warning for lambda variable catpure

This is broken with a recent clang compiler.

This warning got introduced by clang (5.0.0) with this commit:
  https://reviews.llvm.org/D28467

Other recent similar fixes in chrome:
 
https://codereview.chromium.org/2646553002/diff/20001/test/unittests/heap/slo...

BUG=
==========

to

==========
Fix clang compiler warning for lambda variable catpure

This is broken with a recent clang compiler.

This warning got introduced by clang (5.0.0) with this commit:
  https://reviews.llvm.org/D28467

Other recent similar fixes in chrome:
 
https://codereview.chromium.org/2646553002/diff/20001/test/unittests/heap/slo...

BUG=
==========

[etienneb, 2017-02-22 19:07:19]

etienneb@chromium.org changed reviewers:
+ primiano@chromium.org

[etienneb, 2017-02-22 19:08:22]

PTAL & CQ

[Primiano Tucci (use gerrit), 2017-02-23 14:51:10]

lgtm

[Primiano Tucci (use gerrit), 2017-02-23 14:51:14]

The CQ bit was checked by primiano@chromium.org

[commit-bot: I haz the power, 2017-02-23 14:51:35]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-02-23 15:04:14]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-02-23 15:04:15]

Try jobs failed on following builders:
  win_chromium_compile_dbg_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_comp...)
  win_chromium_rel_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)

[etienneb, 2017-02-23 15:35:46]

This is fun, we need to choose between which error we want!
MSVC and Clang won't output the same error.

e:\b\c\b\win\src\base\trace_event\memory_dump_manager_unittest.cc(783): error
C3493: 'kPollsToQuit' cannot be implicitly captured because no default capture
mode has been specified

[Primiano Tucci (use gerrit), 2017-02-23 15:49:49]

On 2017/02/23 15:35:46, etienneb wrote:
> This is fun, we need to choose between which error we want!
> MSVC and Clang won't output the same error.
> 
> e:\b\c\b\win\src\base\trace_event\memory_dump_manager_unittest.cc(783): error
> C3493: 'kPollsToQuit' cannot be implicitly captured because no default capture
> mode has been specified
The fact that they cannot agree on how to pass a constant to a lambda is
extremely amusing. :) 
Good luck

[etienneb, 2017-02-23 15:54:03]

etienneb@chromium.org changed reviewers:
+ chrisha@chromium.org, rnk@chromium.org, thakis@chromium.org

[etienneb, 2017-02-23 15:55:19]

On 2017/02/23 15:49:49, Primiano Tucci wrote:
> On 2017/02/23 15:35:46, etienneb wrote:
> > This is fun, we need to choose between which error we want!
> > MSVC and Clang won't output the same error.
> > 
> > e:\b\c\b\win\src\base\trace_event\memory_dump_manager_unittest.cc(783):
error
> > C3493: 'kPollsToQuit' cannot be implicitly captured because no default
capture
> > mode has been specified
> The fact that they cannot agree on how to pass a constant to a lambda is
> extremely amusing. :) 
> Good luck

The example on the original thread is showing this case:
  https://reviews.llvm.org/D28467

"""
This change makes Clang hardly incompatible with MSVC++. Consider the following
program:

#include <stdio.h>

int main(void) {
  const int kDelta = 10000001;
  auto g = [kDelta](int i)
           {
             printf("%d\n", i % kDelta);
           };
  g(2);
}
Clang will warn about the unused lambda capture:

$ clang++ lala.cc -o lala -std=c++14 -Wall -Werror && ./lala
lala.cc:5:13: error: lambda capture 'kDelta' is not required to be captured for
use in an unevaluated context [-Werror,-Wunused-lambda-capture]
  auto g = [kDelta](int i)
            ^
1 error generated.
"""

[Nico, 2017-02-23 15:58:25]

On 2017/02/23 15:55:19, etienneb wrote:
> On 2017/02/23 15:49:49, Primiano Tucci wrote:
> > On 2017/02/23 15:35:46, etienneb wrote:
> > > This is fun, we need to choose between which error we want!
> > > MSVC and Clang won't output the same error.
> > > 
> > > e:\b\c\b\win\src\base\trace_event\memory_dump_manager_unittest.cc(783):
> error
> > > C3493: 'kPollsToQuit' cannot be implicitly captured because no default
> capture
> > > mode has been specified
> > The fact that they cannot agree on how to pass a constant to a lambda is
> > extremely amusing. :) 
> > Good luck
> 
> The example on the original thread is showing this case:
>   https://reviews.llvm.org/D28467
> 
> """
> This change makes Clang hardly incompatible with MSVC++. Consider the
following
> program:
> 
> #include <stdio.h>
> 
> int main(void) {
>   const int kDelta = 10000001;
>   auto g = [kDelta](int i)
>            {
>              printf("%d\n", i % kDelta);
>            };
>   g(2);
> }
> Clang will warn about the unused lambda capture:
> 
> $ clang++ lala.cc -o lala -std=c++14 -Wall -Werror && ./lala
> lala.cc:5:13: error: lambda capture 'kDelta' is not required to be captured
for
> use in an unevaluated context [-Werror,-Wunused-lambda-capture]
>   auto g = [kDelta](int i)
>             ^
> 1 error generated.
> """

Why is this needed? We disable this warning for this reason. If you want to
build chrome with trunk clang you need to set use_llvm_head (or similar, don't
remember the spelling) to get warning flags that work with trunk.

[etienneb, 2017-02-23 16:04:56]

On 2017/02/23 15:58:25, Nico wrote:
> On 2017/02/23 15:55:19, etienneb wrote:
> > On 2017/02/23 15:49:49, Primiano Tucci wrote:
> > > On 2017/02/23 15:35:46, etienneb wrote:
> > > > This is fun, we need to choose between which error we want!
> > > > MSVC and Clang won't output the same error.
> > > > 
> > > > e:\b\c\b\win\src\base\trace_event\memory_dump_manager_unittest.cc(783):
> > error
> > > > C3493: 'kPollsToQuit' cannot be implicitly captured because no default
> > capture
> > > > mode has been specified
> > > The fact that they cannot agree on how to pass a constant to a lambda is
> > > extremely amusing. :) 
> > > Good luck
> > 
> > The example on the original thread is showing this case:
> >   https://reviews.llvm.org/D28467
> > 
> > """
> > This change makes Clang hardly incompatible with MSVC++. Consider the
> following
> > program:
> > 
> > #include <stdio.h>
> > 
> > int main(void) {
> >   const int kDelta = 10000001;
> >   auto g = [kDelta](int i)
> >            {
> >              printf("%d\n", i % kDelta);
> >            };
> >   g(2);
> > }
> > Clang will warn about the unused lambda capture:
> > 
> > $ clang++ lala.cc -o lala -std=c++14 -Wall -Werror && ./lala
> > lala.cc:5:13: error: lambda capture 'kDelta' is not required to be captured
> for
> > use in an unevaluated context [-Werror,-Wunused-lambda-capture]
> >   auto g = [kDelta](int i)
> >             ^
> > 1 error generated.
> > """
> 
> Why is this needed? We disable this warning for this reason. If you want to
> build chrome with trunk clang you need to set use_llvm_head (or similar, don't
> remember the spelling) to get warning flags that work with trunk.

FLAG: llvm_force_head_revision

[etienneb, 2017-02-23 16:06:13]

On 2017/02/23 16:04:56, etienneb wrote:
> On 2017/02/23 15:58:25, Nico wrote:
> > On 2017/02/23 15:55:19, etienneb wrote:
> > > On 2017/02/23 15:49:49, Primiano Tucci wrote:
> > > > On 2017/02/23 15:35:46, etienneb wrote:
> > > > > This is fun, we need to choose between which error we want!
> > > > > MSVC and Clang won't output the same error.
> > > > > 
> > > > >
e:\b\c\b\win\src\base\trace_event\memory_dump_manager_unittest.cc(783):
> > > error
> > > > > C3493: 'kPollsToQuit' cannot be implicitly captured because no default
> > > capture
> > > > > mode has been specified
> > > > The fact that they cannot agree on how to pass a constant to a lambda is
> > > > extremely amusing. :) 
> > > > Good luck
> > > 
> > > The example on the original thread is showing this case:
> > >   https://reviews.llvm.org/D28467
> > > 
> > > """
> > > This change makes Clang hardly incompatible with MSVC++. Consider the
> > following
> > > program:
> > > 
> > > #include <stdio.h>
> > > 
> > > int main(void) {
> > >   const int kDelta = 10000001;
> > >   auto g = [kDelta](int i)
> > >            {
> > >              printf("%d\n", i % kDelta);
> > >            };
> > >   g(2);
> > > }
> > > Clang will warn about the unused lambda capture:
> > > 
> > > $ clang++ lala.cc -o lala -std=c++14 -Wall -Werror && ./lala
> > > lala.cc:5:13: error: lambda capture 'kDelta' is not required to be
captured
> > for
> > > use in an unevaluated context [-Werror,-Wunused-lambda-capture]
> > >   auto g = [kDelta](int i)
> > >             ^
> > > 1 error generated.
> > > """
> > 
> > Why is this needed? We disable this warning for this reason. If you want to
> > build chrome with trunk clang you need to set use_llvm_head (or similar,
don't
> > remember the spelling) to get warning flags that work with trunk.
> 
> FLAG: llvm_force_head_revision

Ok, it seems to work. I'm closing this CL.

"""
      if (llvm_force_head_revision) {
        cflags += [
          # TODO(hans): https://crbug.com/681136
          "-Wno-unused-lambda-capture",

          # TODO(thakis ): https://crbug.com/683349
          "-Wno-user-defined-warnings",
        ]
      }
"""


Thanks nico