Turn on UseLibraryDependencyInputs for browser_tests

Turn on UseLibraryDependencyInputs for browser_tests

Incremental linking (UseLibraryDependencyInputs) was never enabled for
browser_tests. I seem to remember the toolchain falling over when I
tried to enable it before, but I just tried it now, and it seems to work
locally (and link quickly). So, I guess we can try it and see if
anything breaks.

R=tommycli@chromium.org
BUG=276953

[tommycli, 2013-09-30 23:47:34]

On 2013/09/30 22:47:55, scottmg wrote:

This actually does not work for me. I get messages like:

[16339/16339] LINK browser_tests.exe
FAILED: cmd /c
D:\cygwin\home\tommycli\chromium\depot_tools\python_bin\python.exe gyp-win-tool
link-wrapper environment.x86 link.exe /nologo /OUT:browser_tests.exe
/PDB:browser_tests.exe.pdb @browser_tests.exe.rsp &&
D:\cygwin\home\tommycli\chromium\depot_tools\python_bin\python.exe gyp-win-tool
manifest-wrapper environment.x86 cmd /c if exist browser_tests.exe.manifest del
browser_tests.exe.manifest &&
D:\cygwin\home\tommycli\chromium\depot_tools\python_bin\python.exe gyp-win-tool
manifest-wrapper environment.x86 mt.exe -nologo -manifest
..\..\build\win\compatibility.manifest
obj\chrome\browser_tests.browser_tests.exe.intermediate.manifest
-out:browser_tests.exe.manifest
test_support_common.fake_drive_service.obj : error LNK2005: "bool __cdecl
drive::`anonymous namespace'::EntryMatchWithQuery(class
google_apis::ResourceEntry const &,class std::basic_string<char,struct
std::char_traits<char>,class std::allocator<char> > const &)"
(?EntryMatchWithQuery@?A0xbb869cf8@drive@@YA_NABVResourceEntry@google_apis@@ABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@Z)
already defined in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "class GURL __cdecl
drive::`anonymous namespace'::RemoveQueryParameter(class GURL const &)"
(?RemoveQueryParameter@?A0xbb869cf8@drive@@YA?AVGURL@@ABV3@@Z) already defined
in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "void __cdecl
drive::`anonymous namespace'::ScheduleUploadRangeCallback(class
base::Callback<void __cdecl(struct google_apis::UploadRangeResponse const
&,class scoped_ptr<class google_apis::ResourceEntry,struct
base::DefaultDeleter<class google_apis::ResourceEntry> >)> const
&,__int64,__int64,enum google_apis::GDataErrorCode,class scoped_ptr<class
google_apis::ResourceEntry,struct base::DefaultDeleter<class
google_apis::ResourceEntry> >)"
(?ScheduleUploadRangeCallback@?A0xbb869cf8@drive@@YAXABV?$Callback@$$A6AXABUUploadRangeResponse@google_apis@@V?$scoped_ptr@VResourceEntry@google_apis@@U?$DefaultDeleter@VResourceEntry@google_apis@@@base@@@@@Z@base@@_J1W4GDataErrorCode@google_apis@@V?$scoped_ptr@VResourceEntry@google_apis@@U?$DefaultDeleter@VResourceEntry@google_apis@@@base@@@@@Z)
already defined in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "void __cdecl
drive::`anonymous namespace'::EntryActionCallbackAdapter(class
base::Callback<void __cdecl(enum google_apis::GDataErrorCode)> const &,enum
google_apis::GDataErrorCode,class scoped_ptr<class
google_apis::ResourceEntry,struct base::DefaultDeleter<class
google_apis::ResourceEntry> >)"
(?EntryActionCallbackAdapter@?A0xbb869cf8@drive@@YAXABV?$Callback@$$A6AXW4GDataErrorCode@google_apis@@@Z@base@@W4GDataErrorCode@google_apis@@V?$scoped_ptr@VResourceEntry@google_apis@@U?$DefaultDeleter@VResourceEntry@google_apis@@@base@@@@@Z)
already defined in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "class
std::basic_string<char,struct std::char_traits<char>,class std::allocator<char>
> __cdecl drive::`anonymous namespace'::Identity(class
std::basic_string<char,struct std::char_traits<char>,class std::allocator<char>
> const &)"
(?Identity@?A0xbb869cf8@drive@@YA?AV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@ABV34@@Z)
already defined in
browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: __thiscall
drive::FakeDriveService::FakeDriveService(void)"
(??0FakeDriveService@drive@@QAE@XZ) already defined in
browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: virtual
__thiscall drive::FakeDriveService::~FakeDriveService(void)"
(??1FakeDriveService@drive@@UAE@XZ) already defined in
browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: bool
__thiscall drive::FakeDriveService::LoadResourceListForWapi(class
std::basic_string<char,struct std::char_traits<char>,class std::allocator<char>
> const &)"
(?LoadResourceListForWapi@FakeDriveService@drive@@QAE_NABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@Z)
already defined in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: bool
__thiscall drive::FakeDriveService::LoadAccountMetadataForWapi(class
std::basic_string<char,struct std::char_traits<char>,class std::allocator<char>
> const &)"
(?LoadAccountMetadataForWapi@FakeDriveService@drive@@QAE_NABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@Z)
already defined in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: bool
__thiscall drive::FakeDriveService::LoadAppListForDriveApi(class
std::basic_string<char,struct std::char_traits<char>,class std::allocator<char>
> const &)"
(?LoadAppListForDriveApi@FakeDriveService@drive@@QAE_NABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@Z)
already defined in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: void
__thiscall drive::FakeDriveService::SetQuotaValue(__int64,__int64)"
(?SetQuotaValue@FakeDriveService@drive@@QAEX_J0@Z) already defined in
browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: static
class GURL __cdecl drive::FakeDriveService::GetFakeLinkUrl(class
std::basic_string<char,struct std::char_traits<char>,class std::allocator<char>
> const &)"
(?GetFakeLinkUrl@FakeDriveService@drive@@SA?AVGURL@@ABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@Z)
already defined in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: virtual
void __thiscall drive::FakeDriveService::Initialize(class
std::basic_string<char,struct std::char_traits<char>,class std::allocator<char>
> const &)"
(?Initialize@FakeDriveService@drive@@UAEXABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@Z)
already defined in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: virtual
void __thiscall drive::FakeDriveService::AddObserver(class
drive::DriveServiceObserver *)"
(?AddObserver@FakeDriveService@drive@@UAEXPAVDriveServiceObserver@2@@Z) already
defined in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: virtual
void __thiscall drive::FakeDriveService::RemoveObserver(class
drive::DriveServiceObserver *)"
(?RemoveObserver@FakeDriveService@drive@@UAEXPAVDriveServiceObserver@2@@Z)
already defined in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: virtual
bool __thiscall drive::FakeDriveService::CanSendRequest(void)const "
(?CanSendRequest@FakeDriveService@drive@@UBE_NXZ) already defined in
browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: virtual
class base::Callback<class std::basic_string<char,struct
std::char_traits<char>,class std::allocator<char> > __cdecl(class
std::basic_string<char,struct std::char_traits<char>,class std::allocator<char>
> const &)> __thiscall
drive::FakeDriveService::GetResourceIdCanonicalizer(void)const "
(?GetResourceIdCanonicalizer@FakeDriveService@drive@@UBE?AV?$Callback@$$A6A?AV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@ABV12@@Z@base@@XZ)
already defined in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: virtual
bool __thiscall drive::FakeDriveService::HasAccessToken(void)const "
(?HasAccessToken@FakeDriveService@drive@@UBE_NXZ) already defined in
browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: virtual
void __thiscall drive::FakeDriveService::RequestAccessToken(class
base::Callback<void __cdecl(enum google_apis::GDataErrorCode,class
std::basic_string<char,struct std::char_traits<char>,class std::allocator<char>
> const &)> const &)"
(?RequestAccessToken@FakeDriveService@drive@@UAEXABV?$Callback@$$A6AXW4GDataErrorCode@google_apis@@ABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@Z@base@@@Z)
already defined in browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: virtual
bool __thiscall drive::FakeDriveService::HasRefreshToken(void)const "
(?HasRefreshToken@FakeDriveService@drive@@UBE_NXZ) already defined in
browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: virtual
void __thiscall drive::FakeDriveService::ClearAccessToken(void)"
(?ClearAccessToken@FakeDriveService@drive@@UAEXXZ) already defined in
browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: virtual
void __thiscall drive::FakeDriveService::ClearRefreshToken(void)"
(?ClearRefreshToken@FakeDriveService@drive@@UAEXXZ) already defined in
browser_tests.fake_drive_service.obj
test_support_common.fake_drive_service.obj : error LNK2005: "public: virtual
class std::basic_string<char,struct std::char_traits<char>,class
std::allocator<char> > __thiscall
drive::FakeDriveService::GetRootResourceId(void)const "
(?GetRootResourceId@FakeDriveService@drive@@UBE?AV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@XZ)
already defined in browser_tests.fake_drive_service.obj
...

[scottmg, 2013-10-01 00:00:28]

OK. Those are probably fixable, but yeah, there was a lot of similar errors in
getting the other targets to turn on ULDI (not sure why I didn't get those when
I just tried locally).

I won't have a chance to work through them soon, but I'm happy to review if you
want to start from these gyp changes as a base.

[tommycli, 2013-10-01 00:01:57]

On 2013/10/01 00:00:28, scottmg wrote:
> OK. Those are probably fixable, but yeah, there was a lot of similar errors in
> getting the other targets to turn on ULDI (not sure why I didn't get those
when
> I just tried locally).
> 
> I won't have a chance to work through them soon, but I'm happy to review if
you
> want to start from these gyp changes as a base.

Sounds good. I probably won't get a chance soon either, but we'll fix it at some
point.

[jam, 2013-11-18 06:08:32]

I just tried with a clean build, and this worked.

How can I reproduce the errors? It's very painful to work on browser_tests
without this, so I'd like to help turn this on.

[scottmg, 2013-11-18 15:51:56]

On 2013/11/18 06:08:32, jam wrote:
> I just tried with a clean build, and this worked.
> 
> How can I reproduce the errors? It's very painful to work on browser_tests
> without this, so I'd like to help turn this on.

Yeah, browser_tests really sucks.

I'll try and build locally today and see if anything complains. The first is
non-incremental, so touch a bunch of random files and build again and see if it
doesn't work. If we don't have any problems then might as well land it and see
if anything has trouble I think.

Generally, it's that an obj is getting included two ways, so it should be
reproducible failures.

[jam, 2013-11-18 16:43:41]

On 2013/11/18 15:51:56, scottmg wrote:
> On 2013/11/18 06:08:32, jam wrote:
> > I just tried with a clean build, and this worked.
> > 
> > How can I reproduce the errors? It's very painful to work on browser_tests
> > without this, so I'd like to help turn this on.
> 
> Yeah, browser_tests really sucks.
> 
> I'll try and build locally today and see if anything complains. The first is
> non-incremental, so touch a bunch of random files and build again and see if
it
> doesn't work.

Just tried this a bunch of times and I didn't get any errors. Linking was super
fast instead of a minute!

> If we don't have any problems then might as well land it and see
> if anything has trouble I think.

sgtm, do you want to sync this cl and commit it?

lgtm

> 
> Generally, it's that an obj is getting included two ways, so it should be
> reproducible failures.

[scottmg, 2013-11-18 16:57:36]

On 2013/11/18 16:43:41, jam wrote:
> On 2013/11/18 15:51:56, scottmg wrote:
> > On 2013/11/18 06:08:32, jam wrote:
> > > I just tried with a clean build, and this worked.
> > > 
> > > How can I reproduce the errors? It's very painful to work on browser_tests
> > > without this, so I'd like to help turn this on.
> > 
> > Yeah, browser_tests really sucks.
> > 
> > I'll try and build locally today and see if anything complains. The first is
> > non-incremental, so touch a bunch of random files and build again and see if
> it
> > doesn't work.
> 
> Just tried this a bunch of times and I didn't get any errors. Linking was
super
> fast instead of a minute!
> 
> > If we don't have any problems then might as well land it and see
> > if anything has trouble I think.
> 
> sgtm, do you want to sync this cl and commit it?

Hm, apparently the bot can repro those failures.
http://build.chromium.org/p/tryserver.chromium/builders/win/builds/126801/ste...

It looks like fake_drive_service.obj gets linked into both browser_tests and
test_support_common, and similar for the others lower down. If you want to chase
those up, feel free otherwise I'll try to get it sorted in the next day or two.

I'm not sure why it doesn't happen locally, the GYP_DEFINES and what it's
building seem the same as what I would be building, but there must be something
that's slightly different
http://build.chromium.org/p/tryserver.chromium/builders/win/builds/126801/ste...
.

[jam, 2013-11-22 17:12:47]

On 2013/11/18 16:57:36, scottmg wrote:
> On 2013/11/18 16:43:41, jam wrote:
> > On 2013/11/18 15:51:56, scottmg wrote:
> > > On 2013/11/18 06:08:32, jam wrote:
> > > > I just tried with a clean build, and this worked.
> > > > 
> > > > How can I reproduce the errors? It's very painful to work on
browser_tests
> > > > without this, so I'd like to help turn this on.
> > > 
> > > Yeah, browser_tests really sucks.
> > > 
> > > I'll try and build locally today and see if anything complains. The first
is
> > > non-incremental, so touch a bunch of random files and build again and see
if
> > it
> > > doesn't work.
> > 
> > Just tried this a bunch of times and I didn't get any errors. Linking was
> super
> > fast instead of a minute!
> > 
> > > If we don't have any problems then might as well land it and see
> > > if anything has trouble I think.
> > 
> > sgtm, do you want to sync this cl and commit it?
> 
> Hm, apparently the bot can repro those failures.
>
http://build.chromium.org/p/tryserver.chromium/builders/win/builds/126801/ste...
> 
> It looks like fake_drive_service.obj gets linked into both browser_tests and
> test_support_common, and similar for the others lower down. If you want to
chase
> those up, feel free otherwise I'll try to get it sorted in the next day or
two.
> 
> I'm not sure why it doesn't happen locally, the GYP_DEFINES and what it's
> building seem the same as what I would be building, but there must be
something
> that's slightly different
>
http://build.chromium.org/p/tryserver.chromium/builders/win/builds/126801/ste...
> .

sure I can take a look

[tommycli, 2014-04-21 19:07:58]

Closing, as a version of this patch was committed successfully months ago. See:
https://codereview.chromium.org/82713004