Reland of https://codereview.chromium.org/697223003.
Some fixes to get this subdirectory to compile for MacViews.
An aura::Window dependency is moved from BrowserView to
BrowserCommandHandlerX11.
Hide Ash dependency from ImmersiveModeController.
Add Mac implementation of NativeBrowserFrame.
Now that we include both BrowserWindow::CreateBrowserWindow definitions
in the build, add a new build flag "mac_views_browser" to select between
Cocoa and Views browser.
BUG=425229
Review URL: https://codereview.chromium.org/713043004
Cr-Original-Commit-Position: refs/heads/master@{#303958}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: f15013419679b0e5860d58464f64f7e65d5c9f1e
Code review where I missed this:
codereview.chromium.org/461633002
This only breaks downstream builds that are still tied to CLD1, such as Chrome
for Android. We shouldn't do the changeover to CLD2 without announcing first to
chromium-dev.
notry=true because this simply returns Android to CLD1, as it has always been.,
BUG=
NOTRY=true
TBR=toyoshim@chromium.org
Review URL: https://codereview.chromium.org/698483003
Cr-Original-Commit-Position: refs/heads/master@{#303710}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: fe3b61660588dce1f1cee9ec2bb6c0b2e09f2303
This change further refines the approach taken in setting the implementation
used for accessing compact language detector at runtime. It is based upon the
pattern used in content_client.h, though adapted for this specific use case.
Fundamentally, there is a switch from having exactly one linked-in class define
a global method to instead having the high-level targets (like the content
shell, chrome shell, and so on) be responsible for setting the implementation
that is desired. This eliminates a lot of the ugly GYP/GN logic, but requires
work for each high-level target.
BUG=367239
Review URL: https://codereview.chromium.org/461633002
Cr-Original-Commit-Position: refs/heads/master@{#303669}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 14d91355848a733b18396d74c97d6b945260b899
All other platforms use other -O2 or -Os. Since Linux even uses the same
compiler as mac now, there isn't a great reason for Mac to use a different
optimization level.
In a static release build, this reduces binary size from 175.1MB to 174MB, so
it's probalby not a huge change anyways. If it is, the perf bots will tell us.
BUG=none
Review URL: https://codereview.chromium.org/714433002
Cr-Original-Commit-Position: refs/heads/master@{#303476}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 9ce1b7510648285b80ff75213384bf4186b8c9dd
Some fixes to get this subdirectory to compile for MacViews.
An aura::Window dependency is moved from BrowserView to
BrowserCommandHandlerX11.
Hide Ash dependency from ImmersiveModeController.
Add Mac implementation of NativeBrowserFrame.
Now that we include both BrowserWindow::CreateBrowserWindow definitions
in the build, add a new build flag "mac_views_browser" to select between
Cocoa and Views browser.
BUG=425229
Review URL: https://codereview.chromium.org/697223003
Cr-Original-Commit-Position: refs/heads/master@{#303299}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: f6158b090dc352608141ba8a9d8f09bbdbf4af88
Passing this flag to grit will prevent it from including all
resource IDs in header files and resource maps. The default
behavior is to output all resource defines, even for whitelisted
elements, so this will force compile errors if iOS depends on
resource IDs that are not whitelisted.
BUG=428947, 425998
Review URL: https://codereview.chromium.org/692233002
Cr-Original-Commit-Position: refs/heads/master@{#303242}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 59d7c372f9091d7b6b6c62789148121301336d6c
This flag causes GRIT to include the same resources in generated .h
files as are included in the PAK files. This seems like an obvious thing
to do, right? Well, before this CL, *every* resource from the GRD file
was listed in the .h file, even resources inside of <if> statements that
evaluated to false. As a result, it was possible to reference a resource
(e.g. IDS_FOO) in a .cc file, even if IDS_FOO was excluded from the PAK
file. What happened at runtime? Crash!
This change brings compile-time safety to resource loading: the crash
mentioned above would now be a compile failure since IDS_FOO is no
longer in the generated .h file.
BUG=428947
Review URL: https://codereview.chromium.org/692733003
Cr-Original-Commit-Position: refs/heads/master@{#303180}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 0548dfae799a5423abd007f2f9d8ea0a1d5d571b
Previously, we were using the non-thumb versions of these libraries.
This should save around 60 kiB in binary size.
This also includes some minor style fix in build/common.gypi
Review URL: https://codereview.chromium.org/695333003
Cr-Original-Commit-Position: refs/heads/master@{#303015}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 3572927ce7c9ce6d528820f9575a93985f9309bf
The way we build host binaries in the android gyp backend currently
doesn't work if we attempt to link against libc++. To avoid problems for
AOSP work to change the default C++ library, explicitly specify that we
want to use libstdc++ (the current default).
BUG=
NOTRY=true
Review URL: https://codereview.chromium.org/705753002
Cr-Original-Commit-Position: refs/heads/master@{#302784}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 8beb670cff7f4c09fb6f48b122efea1dff3e8803
Add a new gyp variable: use_libpci (default value is 1).
Add a new define when use_libpci==1: USE_LIBPCI.
Don't call specific PCI code in gpu_info_collector_linux.cc if USE_LIBPCI isn't defined.
BUG=427401
R=piman@chromium.org
Review URL: https://codereview.chromium.org/692893004
Cr-Original-Commit-Position: refs/heads/master@{#302613}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 65ce5accb99908e292b9766c3da6cf54d6cc337f
We can't fix all instances of this warning at the moment (because of gmock
and because of blink apis, see bug). We've fixed many, but our bots still
show almost 20k warnings due to this warning, most of them in v8. That's already
fixed upstream, but v8 rolls don't happen often. Turn the warning off for now.
BUG=428099
NOTRY=true
Review URL: https://codereview.chromium.org/687343003
Cr-Original-Commit-Position: refs/heads/master@{#302356}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 42c03079517c6b0ea14b0427ccd276157b1bef45
FieldTrialTest.DisableProbability was using Time::NowFromSystemTime - 1 year to generate a date before the build time. Instead use GetBuildTime - 1 year.
TransportSecurityState::IsBuildTimely is using GetBuildTime to determine if the build is older than 10 weeks. Disable this and return a default value if DONT_EMBED_BUILD_METADATA is defined (i.e. The build time is invalid) and we're not doing an official build.
CL based on https://codereview.chromium.org/685123005/ with some additional fixes for the unittests.
BUG=314403
Review URL: https://codereview.chromium.org/695523002
Cr-Original-Commit-Position: refs/heads/master@{#302316}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 153e1b83e8af5e6f9dbd3b79e5e0643ea8bcf72c
This increases the size of ChromeShell.apk by 16.8 KiB in
debug builds and 18.7 KiB in release builds.
BUG=90795, 115320
Review URL: https://codereview.chromium.org/695853002
Cr-Original-Commit-Position: refs/heads/master@{#302291}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 459eb777b9914f375ae246196efbb07ec31ba930
When win_analyze=1 is specified then /analyze is added to the command
line to enable static code analysis with VC++.
WarnAsErrors is set to false to allow reporting of all errors.
Additional, WX- is added to disable WarnAsErrors more robustly, necessary because some projects override WarnAsErrors.
Additionally, many noisy but low-value warnings are suppressed.
Finally, _USING_V110_SDK71_ is un-set when win_analyze is true because
this is incompatible with /analyze.
The change to callback_internal.h suppresses an extremely noisy warning about a potentially incorrect use of sizeof():
base\callback_internal.h(80) : warning C6334: sizeof operator applied to an expression with an operator might yield unexpected results: Parentheses can be used to disambiguate certain usages.
With this change almost all of Chrome can be built with high but useful levels of warnings.
bug=427616
Review URL: https://codereview.chromium.org/676743003
Cr-Original-Commit-Position: refs/heads/master@{#301723}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 4b6961df21fb1475efd3f6713ef5c34b1763d748
Compared to the previous CL, this picks up in breakpad:
Fix breakpad on mips and x86_64 for the NDK r10c update.
Original issue's description:
> Revert of Roll android_tools and breakpad to the latest. (patchset #5 id:80001 of https://codereview.chromium.org/671633003/)
>
> Reason for revert:
> Downstream x64 and mips bots didn't like it.
>
> Original issue's description:
> > Roll android_tools and breakpad to the latest.
> >
> > This picks up in android_tools:
> > Updated SDK build tools to 21.0.1 and removed old versions.
> > Upgrade NDK to r10c.
> >
> > And in breakpad:
> > Update breakpad to support Android NDK r10c
> > Fixing the Mac Inspector build by adding the launch_report dependency to the breakpadUtilities dylib project after it got pulled out from Inspector.mm.
> > Refactor .so name detection logic in minidump/linux_dumper.
> > The process uptime computation when the Windows out-of-process Crash Generation Server is used is done by subtracting the crashing process creation time from the current time when the crash is generated.
> > Linux multiprocess crash reporter: use pipe instead of socketpair in the client.
> > Remove workaround for Linux credential-passing bug.
> >
> > TBR=brettw, danakj
> > NOTRY=true
> > BUG=424680, 425077, 358831
> >
> > Committed: https://crrev.com/79bd5f12620f81dd241984d925b87a888e1dc37b
> > Cr-Commit-Position: refs/heads/master@{#300671}
>
> TBR=brettw@chromium.org,primiano@chromium.org,rmcilroy@chromium.org,aurimas@chromium.org,danakj@chromium.org,fdegans@chromium.org
> NOTREECHECKS=true
> NOTRY=true
> BUG=424680, 425077, 358831
>
> Committed: https://crrev.com/e6319d7f7d5c2fc96c9cb1f5a81ae16bde19f068
> Cr-Commit-Position: refs/heads/master@{#300688}
TBR=brettw@chromium.org,primiano@chromium.org,rmcilroy@chromium.org,aurimas@chromium.org,danakj@chromium.org,ppi@chromium.org
BUG=424680, 425077, 358831
Review URL: https://codereview.chromium.org/676483002
Cr-Original-Commit-Position: refs/heads/master@{#301087}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 27660a3c6e308e9293745697433747e0a4fafae0
install-debian.wheezy.sysroot.py is now use for ARM
also, removing duplication. sysroot location is also
now consistent.
Review URL: https://codereview.chromium.org/664383004
Cr-Original-Commit-Position: refs/heads/master@{#300746}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 41d0e08a38f6cc0cb32a2596953a7262c4a9ab02
Reason for revert:
Downstream x64 and mips bots didn't like it.
Original issue's description:
> Roll android_tools and breakpad to the latest.
>
> This picks up in android_tools:
> Updated SDK build tools to 21.0.1 and removed old versions.
> Upgrade NDK to r10c.
>
> And in breakpad:
> Update breakpad to support Android NDK r10c
> Fixing the Mac Inspector build by adding the launch_report dependency to the breakpadUtilities dylib project after it got pulled out from Inspector.mm.
> Refactor .so name detection logic in minidump/linux_dumper.
> The process uptime computation when the Windows out-of-process Crash Generation Server is used is done by subtracting the crashing process creation time from the current time when the crash is generated.
> Linux multiprocess crash reporter: use pipe instead of socketpair in the client.
> Remove workaround for Linux credential-passing bug.
>
> TBR=brettw, danakj
> NOTRY=true
> BUG=424680, 425077, 358831
>
> Committed: https://crrev.com/79bd5f12620f81dd241984d925b87a888e1dc37b
> Cr-Commit-Position: refs/heads/master@{#300671}
TBR=brettw@chromium.org,primiano@chromium.org,rmcilroy@chromium.org,aurimas@chromium.org,danakj@chromium.org,fdegans@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=424680, 425077, 358831
Review URL: https://codereview.chromium.org/676453002
Cr-Original-Commit-Position: refs/heads/master@{#300688}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: e6319d7f7d5c2fc96c9cb1f5a81ae16bde19f068
This picks up in android_tools:
Updated SDK build tools to 21.0.1 and removed old versions.
Upgrade NDK to r10c.
And in breakpad:
Update breakpad to support Android NDK r10c
Fixing the Mac Inspector build by adding the launch_report dependency to the breakpadUtilities dylib project after it got pulled out from Inspector.mm.
Refactor .so name detection logic in minidump/linux_dumper.
The process uptime computation when the Windows out-of-process Crash Generation Server is used is done by subtracting the crashing process creation time from the current time when the crash is generated.
Linux multiprocess crash reporter: use pipe instead of socketpair in the client.
Remove workaround for Linux credential-passing bug.
TBR=brettw, danakj
NOTRY=true
BUG=424680, 425077, 358831
Review URL: https://codereview.chromium.org/671633003
Cr-Original-Commit-Position: refs/heads/master@{#300671}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 79bd5f12620f81dd241984d925b87a888e1dc37b
This turns Chrome Shell's AndroidManifest.xml into a template which gets
processed at build time to produce the final AndroidManifest.xml that's
used to build the APK.
BUG=163751
NOTRY=true
Review URL: https://codereview.chromium.org/657443002
Cr-Original-Commit-Position: refs/heads/master@{#300409}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: d5366cb1ad27e6bf1ae659f374a9b9dd3186761d
To allow the chromium WebView to build unbundled, we need to stop
linking to system libraries and instead link to the prebuilt libraries
in the NDK.
This change rolls gyp to include the required makefile generation
changes, and adjusts the chromium side to match:
1) Stop setting use_system_stlport and instead use the NDK stlport by
specifying LOCAL_NDK_STL_VARIANT := stlport_static
2) Remove support for the use_system_stlport flag entirely as WebView
was the only user.
3) Don't try to link standard libraries like -lc -lm as the NDK does
this automatically.
4) Don't include the OpenAL headers from frameworks/wilhelm; these are
already part of the NDK with no special include path needed.
BUG=409851
Review URL: https://codereview.chromium.org/571953002
Cr-Original-Commit-Position: refs/heads/master@{#300212}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 3bc122f4da5cecfc6ffac44bce3fddf432afa174
This reverts commit 96256bfea5c496aa70b1913555b1e1c7f166c426, returning
the webview build to using Chrome's copy of ICU.
BUG=409851
NOTRY=true
Review URL: https://codereview.chromium.org/661983002
Cr-Original-Commit-Position: refs/heads/master@{#300092}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 2f017213ed940754da2bbc24fadd5386fe82e3fc
This CL is to build base/ for nacl_helper_nonsfi.
The library is similar to base_nacl, but slightly different:
- For rand_util, rand_util_posix should be used, instead of
rand_util_nacl, because nacl_helper_nonsfi will be running under
Linux directly.
- MessageLoopForIO should be based on MessagePumpLibevent rather than
MessagePumpDefault, to support IPC.
- GetKnownDeadTerminationStatus, GetTerminationStatus,
UnixDomainSocket::SendMsg and RecvMsg are included, as these are used
to implement nacl_helper_nonsfi binary.
- GLIB is not supported. It is unnecessary for nacl_helper_nonsfi.
Note that this library is not used yet from any binary, because this CL
is just a preparation, but the library is built actually.
BUG=358465
TEST=Ran trybot. Implement nacl_helper_nonsfi on top of this CL, and made sure
it is working.
Review URL: https://codereview.chromium.org/659513004
Cr-Original-Commit-Position: refs/heads/master@{#300075}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 2b720d21260715614abea6f241874cc779f01332
iOS does not have this requirement, but we've been installing the 10.6
SDK on the buildbot slaves to make the 'find_sdk.py --verify' call
happy. However, iossim no longer builds with 10.6 so it's time to update
the mac_sdk_min for iOS to 10.8 and remove the false requirement for
official builds.
BUG=418602
Review URL: https://codereview.chromium.org/660783004
Cr-Original-Commit-Position: refs/heads/master@{#299904}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 3a721cdeaa08279e20010c88e282b4a12bcc3ee5
Update the plugin cmake files to explicitly depend on Clang. When we moved
to the approach of building tools and plugins inside the LLVM build in
https://codereview.chromium.org/615083006/, that dependency was lost. We should
find a better way to do this eventually.
Also move the code to clear the CHROME_TOOLS_SHIM_DIR earlier in the script
to avoid having it affect the bootstrap build.
BUG=420674
NOTRY=true
Review URL: https://codereview.chromium.org/630163005
Cr-Original-Commit-Position: refs/heads/master@{#299766}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 51709a2ce53da80b7814b9244fb874e185e50566
To avoid causing problems for people working on using clang by default
in AOSP, explicitly specify that the android webview build doesn't want
to use clang for target binaries. The current default is false, so this
doesn't change behaviour in the normal case.
BUG=
NOTRY=true
Review URL: https://codereview.chromium.org/661433003
Cr-Original-Commit-Position: refs/heads/master@{#299691}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 6c1c44664f6b3cf7782142135be6206aeb356217
This isn't in general recommended by Android,
and it's causing problems with some services that
need to establish white/black-listing.
BUG=422491
Review URL: https://codereview.chromium.org/647483004
Cr-Original-Commit-Position: refs/heads/master@{#299334}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: ad608226402b04d953f5c7b13966e6a453e20a9e
Current AOSP versions use clang to build host code by default and
https://codereview.chromium.org/516723004 configured common.gypi to
assume this, but this breaks building against older releases where clang
is not the default. Force use of clang by emitting "LOCAL_CLANG := true"
into generated host makefiles so that it will work no matter what
release we are using.
BUG=421873
Review URL: https://codereview.chromium.org/644683004
Cr-Original-Commit-Position: refs/heads/master@{#298942}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 981f132f9f354a9c3cd1bb45bb1f95f8d006f5fa
andrewhayden
andresantoso
derat
Sadrul Habib Chowdhury
thakis
glider
raymes
oshima
lliabraa
newt
fdegans
vitalybuka
torne
bedupont
sebmarchand
peter
earthdok
brucedawson
scroggo
baixo
thestig
sbc
ppi
aurimas
hidehiko
hans
aruslan
maniscalco