Add a new GYP_DEFINE to make _HAS_ITERATOR_DEBUGGING disablable on Windows/De...

Add a new GYP_DEFINE to make _HAS_ITERATOR_DEBUGGING disablable on Windows/Debug
Shouldn't affect anyone unless I enable this flag on the TSan bot.
TEST=Tested with a local waterfall

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=65377

[Timur Iskhodzhanov, 2010-11-08 13:57:57]

Hi gentlemen,

Can you please review this CL?
The next step would be setting win_debug_disable_iterator_debugging=1 for the
TSan/Win bot. That may speed up things there, at least CookieMonster tests.

Thank you,
Timur Iskhodzhanov

[Alexander Potapenko, 2010-11-08 14:09:43]

rubber stamp LGTM

On Mon, Nov 8, 2010 at 4:57 PM,  <timurrrr@chromium.org> wrote:
> Reviewers: Alexander Potapenko, rdsmith, Marc-Antoine Ruel,
>
> Message:
> Hi gentlemen,
>
> Can you please review this CL?
> The next step would be setting win_debug_disable_iterator_debugging=1 for
> the
> TSan/Win bot. That may speed up things there, at least CookieMonster tests.
>
> Thank you,
> Timur Iskhodzhanov
>
> Description:
> Add a new GYP_DEFINE to make _HAS_ITERATOR_DEBUGGING disablable on
> Windows/Debug
> Shouldn't affect anyone unless I enable this flag on the TSan bot.
> TEST=Tested with a local waterfall
>
> Please review this at http://codereview.chromium.org/4688001/show
>
> SVN Base: svn://svn.chromium.org/chrome/trunk/src/
>
> Affected files:
>  M     build/common.gypi
>
>
> Index: build/common.gypi
> ===================================================================
> --- build/common.gypi   (revision 65371)
> +++ build/common.gypi   (working copy)
> @@ -434,6 +434,12 @@
>       # See http://msdn.microsoft.com/en-us/library/47238hez(VS.71).aspx
>       'win_debug_InlineFunctionExpansion%': '',    # empty = default, 0 =
> off,
>       'win_release_InlineFunctionExpansion%': '2', # 1 = only __inline, 2 =
> max
> +      # VS inserts quite a lot of extra checks to algorithms like
> +      # std::partial_sort in Debug build which make them O(N^2)
> +      # instead of O(N*logN). This is particularly slow under memory
> +      # tools like ThreadSanitizer so we want it to be disablable.
> +      # See http://msdn.microsoft.com/en-us/library/aa985982(v=VS.80).aspx
> +      'win_debug_disable_iterator_debugging%': '', # non-empty = disable
>
>       'release_extra_cflags%': '',
>       'debug_extra_cflags%': '',
> @@ -722,6 +728,9 @@
>                 'InlineFunctionExpansion':
>                   '<(win_debug_InlineFunctionExpansion)',
>               }],
> +              ['win_debug_disable_iterator_debugging!=""', {
> +                'PreprocessorDefinitions': ['_HAS_ITERATOR_DEBUGGING=0'],
> +              }],
>             ],
>           },
>           'VCLinkerTool': {
>
>
>

[M-A Ruel, 2010-11-08 14:11:03]

http://codereview.chromium.org/4688001/diff/2001/3001
File build/common.gypi (right):

http://codereview.chromium.org/4688001/diff/2001/3001#newcode442
build/common.gypi:442: 'win_debug_disable_iterator_debugging%': '', # non-empty
= disable
Why not reuse the convention 0/1 here?

[Randy Smith (Not in Mondays), 2010-11-08 14:44:33]

LGTM.

On 2010/11/08 13:57:57, Timur Iskhodzhanov wrote:
> Hi gentlemen,
> 
> Can you please review this CL?
> The next step would be setting win_debug_disable_iterator_debugging=1 for the
> TSan/Win bot. That may speed up things there, at least CookieMonster tests.
> 
> Thank you,
> Timur Iskhodzhanov

[Timur Iskhodzhanov, 2010-11-08 14:45:47]

http://codereview.chromium.org/4688001/diff/2001/3001
File build/common.gypi (right):

http://codereview.chromium.org/4688001/diff/2001/3001#newcode442
build/common.gypi:442: 'win_debug_disable_iterator_debugging%': '', # non-empty
= disable
Done.
Please take another look.
On 2010/11/08 14:11:03, Marc-Antoine Ruel wrote:
> Why not reuse the convention 0/1 here?

[M-A Ruel, 2010-11-08 15:44:09]

lgtm