This gets the //chrome subtree down to a fairly small number of build
errors on Android. Most of the changes are making features not enabled
on android like extensions and plugins properly guarded. This also adds
a few jni targets and fixes up some bitrotted bits.
R=brettw@chromium.org, cjhopman@chromium.org
Review URL: https://codereview.chromium.org/571863002
Cr-Original-Commit-Position: refs/heads/master@{#295231}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 2ed4d695d6cf2917fc0e52f767d629fbefa7a1f9
- Use bundled binutils by adding to cflags -B
- Enables dwarf-fission based on bundling (maybe correct?)
- Makes differ normalize -B paths so they can be compared between gyp/gn.
R=brettw@chromium.org
BUG=335824
Review URL: https://codereview.chromium.org/568613002
Cr-Original-Commit-Position: refs/heads/master@{#295185}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 416440522fa2e5034c76a2adbe069bc7c6faaf55
This updates Chromium's clang version to r216630 with
r216684 cherry-picked to fix an ASan issue.
It also changes the build script for Clang to use CMake
instead of Autoconf. The ASan team say this configuration
is better tested, and it also makes us consistent with
the Windows Clang build which already uses CMake.
BUG=400849
R=thakis@chromium.org
TBR=brettw
Review URL: https://codereview.chromium.org/578623002
Cr-Original-Commit-Position: refs/heads/master@{#295099}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 3853ffa861dc43537f23c21a3fc4104c6f08c3a2
This is a reland of dd7edfa67 which was reverted in 80a9a88f4c for build
failures. It also includes a BoringSSL roll and build changes to avoid adding a
static initializer.
This is a much much larger change than the diff suggests. If it breaks
something, please revert first and ask questions later.
BUG=338885
TBR=brettw
Review URL: https://codereview.chromium.org/568643003
Cr-Original-Commit-Position: refs/heads/master@{#294589}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 71c13e6ac01d5cdc49ca40e8e9301d5142d85018
- Don't remove -fno-ident from gyp build (was added about 3y ago, seemingly unnecessary for current clang)
- Add -fno-slp-vectorize to GN to make builds match (both to be removed after next clang roll)
- Make usage of -Wno-reserved-user-defined-literal match (required for dbus pre-trusty)
- Add -march=x86-64 to gn cflags
- Default symbol_level based on is_debug, and turn it off by default in Release (same as gyp)
Also, fiddle with output of gyp_flag_compare.py.
R=brettw@chromium.org
BUG=335824
Review URL: https://codereview.chromium.org/562383002
Cr-Original-Commit-Position: refs/heads/master@{#294489}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 29682b848742042388ebd29c9df775f074c1414c
OS_CHROMEOS is defined in //build/config/linux:sdk per brettw
recommendation.
Typo fixed in ui/ozone/BUILD.gn
Some source-list fixes in content/browser/BUILD.gn that take both
use_x11 and use_ozone into account.
BUG=None
TEST=None
Review URL: https://codereview.chromium.org/558343002
Cr-Original-Commit-Position: refs/heads/master@{#294426}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: db5203ee72bc04974fde7bc990a76e9ca6449f6f
This enables all blink targets and nearly all content targets in the
android build. This produces a content shell APK, but not a fully
operational one yet. This also flips on a bunch of other targets pulled
in by deps.
R=brettw@chromium.org
Review URL: https://codereview.chromium.org/561623002
Cr-Original-Commit-Position: refs/heads/master@{#294247}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 060f2e31fd19dd8d5b029cd281aed05a510c545f
This is a much much larger change than the diff suggests. If it breaks
something, please revert first and ask questions later.
BUG=338885
Review URL: https://codereview.chromium.org/539163002
Cr-Original-Commit-Position: refs/heads/master@{#294233}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: dd7edfa679492a6feb3ef2e1630f05da82a1b31f
This fixes differences in -Wno-unused-function and
-Wno-unneeded-internal-declaration.
It also changes the diff'ing script to accept
-Wendif-labels
-Wextra
-Wsign-compare
when in GN but not gyp, because they're more conservative. These
mostly happen when GN declares a lib chromium_code, but gyp doesn't.
Also, remove some defines from the dont_care, as I'm not sure that
we don't care yet.
And some minor tweaks to diff script output.
R=brettw@chromium.org
BUG=335824
Review URL: https://codereview.chromium.org/555343002
Cr-Original-Commit-Position: refs/heads/master@{#294229}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: e8d3adb185373b15d593855f623fb5e28ef7878c
This changes the "test" template to create a shared_library (instead of
an executable) on Android. After this, we can actually run
base_unittests with
`build/android/test_runner.py gtest -s base_unnittests`
as normal (though may need to CHROMIUM_OUT_DIR and BUILDTYPE as
appropriate).
This requires adding the following targets:
//testing/android:native_test_native_code
//testing/android:native_test_util
//testing/android:native_test_jni_headers
//tools/android/md5sum:md5sum
//tools/android/md5sum:md5sum_bin
//tools/android/md5sum:md5sum_prepare_dist
//tools/android/md5sum:md5sum_copy_host
Also, makes it so that native executables are stripped (just like shared
libraries). Adds a simple create_native_execuatable_dist template that
sets up a dist directory for the executable (see
build/android/gyp/native_app_dependencies.gyp).
BUG=359249
TBR=rlarocque
Review URL: https://codereview.chromium.org/557463002
Cr-Original-Commit-Position: refs/heads/master@{#294032}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: e65ead8fb1a1976682356cf824445476584e70f0
Some packages used in linux builds are not used on CrOS, and
indeed are not even installed in the CrOS build chroot. Exclude
them so that the GN build doesn't try to run pkg-config on
packages that aren't installed.
BUG=388412
TEST=Create a target that builds //mojo/public, build with os==chromeos
R=brettw
Review URL: https://codereview.chromium.org/556733002
Cr-Original-Commit-Position: refs/heads/master@{#293977}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: c397c15b4db8d22528410af612fa2b154bbdb635
build/config/linux/BUILD.gn:
* Usage of generate_library_loader for libbrlapi was incorrect; there
* was nothing provided for "functions". This is now an empty list.
chrome/chrome_common.gypi
chrome/common/BUILD.gn
* Reference to a removed file, chrome_version_info_linux.cc
chromeos/BUILD.gn
components/pairing/BUILD.gn
* Usage of removed proto_{in,out}_dir in proto_library rules.
content/gpu/BUILD.gn
* Typo in reference to third_party/libva:libva_config
ui/views/BUILD.gn
* Attempt to filter source files out of
* gypi_values.views_unittests_sources by subdirectory path. Fortunately
* that variable already doesn't include the files that were supposed
* to be filtered out, so the rule is simply removed.
BUG=388412
TEST=Create a target that builds //mojo/public, build with os==chromeos
R=brettw@chromium.org
Review URL: https://codereview.chromium.org/550423002
Cr-Original-Commit-Position: refs/heads/master@{#293964}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 418e04caaf5701fac99ae766ab60f2e65279fccd
Also remove a few redundant consts that result in warnings like:
../../content/common/cursors/webcursor_aurax11.cc:20:1: error: 'const' type qualifier on return type has no effect [-Werror,-Wignored-qualifiers]
const ui::PlatformCursor WebCursor::GetPlatformCursor() {
^~~~~~
TBR=davemoore@chromium.org,sadrul@chromium.org
R=brettw@chromium.org, thakis@chromium.org,rch@chromium.org
BUG=393046, 335824
Review URL: https://codereview.chromium.org/538333002
Cr-Original-Commit-Position: refs/heads/master@{#293930}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 4c2d33ac8c061cf27a594473d6f9a450ee6c3c89
Tune global Android ARM cflags for performance.
Update cflags in gn to keep up with gyp.
BUG=407545
CQ_TRYBOTS=tryserver.chromium.linux:android_aosp,android_clang_dbg,android_dbg
Review URL: https://codereview.chromium.org/375483003
Cr-Original-Commit-Position: refs/heads/master@{#293913}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 6b767135e631b94cd096c78f6622be28a4af277e
This adds support for android_apk targets without any java included
directly in the target (only included through deps).
This adds the following targets:
//base:base_javatests
//content/public/android:content_javatests
//content/shell/android:content_shell_test_apk
//media/android:media_android_imageformat_list (moved)
//media/android:media_java (moved+fixed)
BUG=359249
Review URL: https://codereview.chromium.org/525533003
Cr-Original-Commit-Position: refs/heads/master@{#293897}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 394dfd1052e53e64083daef7cc66055557256dcb
The content_shell_apk does not have the full native part that it should.
It builds, installs, starts, loads the native library, and then crashes
when it tries to access missing native parts.
This requires fixes to a bunch of little things:
1. Should only extract java files to compile from srcjars.
2. The way we were creating the single combined resources zip was wrong
because we were clobbering xml files (we should only clobber images, xml
files need to be handled more carefully). Instead of trying to properly
handle xml files, the combined zip now just puts all the different
dependencies in different folders (0/, 1/, etc) and then when packaging
we detect that type of zip and pass all the subdirs to aapt.
3. java_cpp_template with multiple sources needs to have a unique
depfile path for each source
4. jni_generator needs to clear the classlist variable in each loop of
foreach (otherwise gn complains about assigning to non-empty list)
5. a couple parts of apk native lib handling was not skipped for an apk
with no native lib
This adds a bunch of targets:
//base:base_java_test_support
//content/public/android:content_java_resources
//content/public/android:content_java
//content/public/android:content_strings_grd
//content/public/android:gesture_event_type_java
//content/public/android:page_transition_types_java
//content/public/android:popup_item_type_java
//content/public/android:result_codes_java
//content/public/android:selection_event_type_java
//content/public/android:speech_recognition_error_java
//content/public/android:top_controls_state_java
//content/public/android:screen_orientation_values_java
//content/public/android:content_gamepad_mapping_java
//content/public/android:content_jni_headers
//content/shell/android:content_shell_jni_headers
//content/shell/android:libcontent_shell_content_view (partial)
//content/shell/android:content_shell_java_resources
//content/shell/android:content_shell_java
//content/shell/android:content_shell_apk_resources
//content/shell/android:content_shell_apk
//net/android:net_java
//net/android:remote_android_keystore_aidl
//net/android:net_java_test_support
//net/android:net_javatests
//net/android:net_errors_java
//net/android:certificate_mime_types_java
//net/android:cert_verify_status_android_java
//net/android:private_key_types_java
//net/android:net_unittests_apk (disabled)
TBR=yfriedman,thestig
BUG=359249
Review URL: https://codereview.chromium.org/507523002
Cr-Original-Commit-Position: refs/heads/master@{#293671}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: cd7694afb928b265e64be5448bb281ad86771f34
Adds the android_aidl template to support Android aidl.
Adds a DEPRECATED_java_in_dir to android_library and android_apk. This
will greatly ease the transition from gyp to gn.
Adds BUILD.gn for third_party/eyesfree (using the two new features
added).
TBR=rsleevi
BUG=359249
Review URL: https://codereview.chromium.org/512923002
Cr-Original-Commit-Position: refs/heads/master@{#293386}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: dad5f427256371787b536ccdcf08d6493bebb379
It also hooks up libpdf which chrome requires at startup.
Adds the enable_hidpi and use_athena build flags.
Adds a locales list.
This makes the chrome target actually runnable from a GN build.
R=jamesr@chromium.org
Review URL: https://codereview.chromium.org/541153002
Cr-Original-Commit-Position: refs/heads/master@{#293360}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 6a90cafc8e2b0f1e1bc6a8f7e69470ac8acfa253
The GLIB_DISABLE_DEPRECATION_WARNINGS define disables the deprecation
warnings which is caused by the usage of the g_settings_list_schemas
method. Moving the define to the build scripts avoids any future build
errors (-Werror) and warning reports.
BUG=391250
TEST=ninja -C out/Release chrome
Review URL: https://codereview.chromium.org/367113004
Cr-Original-Commit-Position: refs/heads/master@{#292608}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 3a31d8e495b1d04a01f9e6c57fac608fee0285e3
Previously "component" was a built-in function that forwarded to source_set, static_library, or shared_library depending on the value of component_mode. And test was a built-in function that forwarded to executable.
These can be expressed as templates instead. This patch removes the built-in versions and implements the templates in the master build config file.
Forwarding all of the variables for these is somewhat tedious, but I still prefer not to add a new concept for this, since it will be rarely used and less clear what is happening.
With the updated BUILDCONFIG.gn file, the build will work with both the old and new GN version (the built-in version of component and test take precedence). When the new GN binary is pushed, we can remove the references to component_mode.
BUG=
R=hclam@chromium.org
Review URL: https://codereview.chromium.org/516703005
Cr-Original-Commit-Position: refs/heads/master@{#292500}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 60a5bf4a8a5d1a2dad8e7a67f81e72756e886266
Reason for revert:
Tentatively reverting because it made the sizes bot grumpy (http://build.chromium.org/p/chromium/buildstatus?builder=Linux&number=52715)
Original issue's description:
> linux: Always use (64bit) gold for linking, even on 32bit.
>
> All 32bit bots now have 64bit kernels and can run 64bit gold. So always use
> that. We already require running 64bit binaries for clang, and the official
> bots already used to do this for gold. This makes linking on 32bit bots faster,
> and simplifies our configurations.
>
> BUG=396333
> TBR=brettw
>
> Committed: c7864673b1TBR=thestig@chromium.org,brettw@chromium.org,thakis@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=396333
Review URL: https://codereview.chromium.org/482173003
Cr-Original-Commit-Position: refs/heads/master@{#291905}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 277720168a1fd1bd3147912fc81c88f6dc0f1c69
All 32bit bots now have 64bit kernels and can run 64bit gold. So always use
that. We already require running 64bit binaries for clang, and the official
bots already used to do this for gold. This makes linking on 32bit bots faster,
and simplifies our configurations.
BUG=396333
TBR=brettw
Review URL: https://codereview.chromium.org/503873004
Cr-Original-Commit-Position: refs/heads/master@{#291904}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 7cd5e7a3a82ebad8bf53e1236229d1094f8dafc6
The previous code removed symbols from clang builds based on what GYP used to
do. But GYP changed this logic when clang was set to the default. GN was then
left with clang on by default, but no symbols.
This adds some additional GYP flag parity work.
R=scottmg@chromium.org
Review URL: https://codereview.chromium.org/501173002
Cr-Original-Commit-Position: refs/heads/master@{#291803}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 7927079955d2767e1fa812fe58432fcb9d6250aa
Since GN runs its scripts in a different current directory than GYP, I had to update the remoting html template script to be able to handle this situation.
Also includes many misc linking fixes:
- Sets enable_nacl to false.
- Add renderer context menu component
- Adds many missing files
R=sergeyu@chromium.org
Review URL: https://codereview.chromium.org/498773005
Cr-Original-Commit-Position: refs/heads/master@{#291731}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 1c693998d75f78e94862c324afba3815356ecaf1
Mostly just a refactoring of build rules to support prebuilt jars.
Adds support for proguard preprocessing of both prebuilt and
non-prebuilt java libraries.
Adds definitions of all (but for android.jar) of the prebuilt jars that
we need from the sdk.
BUG=359249
Review URL: https://codereview.chromium.org/484813002
Cr-Original-Commit-Position: refs/heads/master@{#291464}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 684b22957e09c08e917a63390b04980a67da4fa1
First, change the NativeLibraries template to use preprocessor defines
instead of includes. This allows us to delete all the corresponding
headers.
Then, change the ordered libraries file to be a dict with both the list
of library paths and the java-style list of library names.
Update a bunch of things that used the ordered libraries file to instead
take the list of libraries directly as an argument (as a FileArg in
every case).
Add calculating the list of libraries to write_build_config and add it
to the build config (gn can use this then instead of a separate ordered
libraries file).
Add a simple copy step in gn to set up the native libraries directory
for an apk. Use this directory when creating the apk.
BUG=359249
NOTRY=true
Review URL: https://codereview.chromium.org/456493002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@290963 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This pulls buildtools to get GN 290714 and updates to the new style of toolchain definitions in that revision.
Unfortunately, this new version doesn't support getting the outputs of excutables, which made the android unit test template a bit less automatic. We can consider how to best fix this in the future.
R=jamesr@chromium.org
Review URL: https://codereview.chromium.org/485833003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@290894 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
Reason for revert:
Breaks builds where use_udev==0 since certain targets just assume udev and never check against use_udev.
Udev package should be wrapped in a use_udev confitional altogether and all the locations using Udev should check against use_udev.
Original issue's description:
> Run pkg-config for Udev only if use_udev==1
>
> Some embedded platforms do not have udev support. The change will skip
> running pkg-config if use_udev=0.
>
> BUG=none
>
> Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=288770TBR=brettw@chromium.org
NOTRY=true
Review URL: https://codereview.chromium.org/469763002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@289280 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
When gn's desc command is used with --tree, it prints a dependency tree. But recursively printing the dependency tree of a high-level target in Chrome prints too much output to be useful because it's printing, e.g. base and it's deps over and over (//chrome/browser-s --tree is 3210166 lines).
With this patch, a target will be recursed into the first time it's printed, but subsequent times the children will be elided with "..." This brings down //chrome/browser to 2050 lines. The previous behavior is still available (it could be useful in certain cases) via "--tree --all".
This also updates an optimization comment in the the optimization setup (optimize_max isn't just Windows-only).
R=dalecurtis@chromium.org
Review URL: https://codereview.chromium.org/437033002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@287424 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
On Android, we detect neon support at runtime and switch implementations
based on that. This requires that some parts of skia are compiled with
-mfpu=neon.
To support this, the -mfpu flag is moved out of the 'compiler' config
and into its own config. A target can then remove this config and supply
its own -mfpu flag (without having to duplicate all the other stuff in
the 'compiler' config).
BUG=359249
Review URL: https://codereview.chromium.org/371853007
git-svn-id: http://src.chromium.org/svn/trunk/src/build@286038 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This adds two templates, android_apk and unittest_apk. unittest_apk is
a simple wrapper of android_apk for packaging gtests into apks.
android_apk currently just does the java compilation, processing
resources, and build_config stuff for apks.
BUG=359249
Review URL: https://codereview.chromium.org/379333003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@286031 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
It doesn't run yet. This disables WebRTC since that's not done yet.
Libvpx isn't done either. This patch exports the disable flag for this from media to dependent targets like the GYP build does. It adds one extra check for this flag in Pepper. I assume everybody that compiles without libvpx also disables plugins so this has never come up before.
Disable some warnings in Mesa.
The resulting binary does not yet run.
R=jamesr@chromium.org
Review URL: https://codereview.chromium.org/419913003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@285962 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This fixes the GN build for the ui/ozone component (and tests). The
following build should work going forward:
gn gen out_gn_ozone --args="use_ozone=true"
ninja -C out_gn_ozone -j1000 -k50 ozone_egl_demo ozone_unittests
BUG=396316
TEST=compile & run ozone_egl_demo, ozone_unittests
TBR=darin
Review URL: https://codereview.chromium.org/409163004
git-svn-id: http://src.chromium.org/svn/trunk/src/build@285362 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This adds the missing conditions to the chrome/common GN build file (previously
only the basic files were there). Some refactoring of the GYP file to make this
possible.
Renames the enable_printing flag to printing_mode since this is an integer
tri-state.
Minor improvements to chrome/renderer
R=jamesr@chromium.org
Review URL: https://codereview.chromium.org/400243003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@284508 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This changes the syntax from:
@(filepath:key0:key1)
to:
@FileArg(filepath:key0:key1)
This should make it much easier for someone who encounters this syntax
to figure out what it is (currently a `git grep FileArg` returns only
results about/using this feature).
Also changes it to use a simple regex search instead of adhoc stuff.
BUG=359249
Review URL: https://codereview.chromium.org/377803003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@283484 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This adds support for android resources to gn.
Two new templates are introduced:
java_string_grd: This is like gyp's build/java_strings_grd.gypi. It runs
grit and generates Android strings.xml files and then zips them
together.
android_resources: This is the target for android resources. It is
mostly a wrapper around process_resources.py. This is *not part* of
android_library (like it is in gyp). Making these two things separate is
more like facebook's BUCK and google's internal build rules (and gyp's
java.gypi and java_apk.gypi have gotten way too big and complicated).
Changes to the actual build scripts are very minor except for the added
support for andoid_resources to write_build_config.py. Building
resources requires getting all the transitive resource dependencies, and
so this reuses the simple sorted transitive dependency thing from
write_ordered_libraries.py.
TBR=rch@chromium.org
BUG=359249
Review URL: https://codereview.chromium.org/361633002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@283293 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
The problem was that the toolchain definitions had an is_clang block in them, but this is only evaluated in the context of the default toolchain (so false for and Android build). But then when we were re-evaluating the build config for the host compile, it was forcing clang to true which was causing the clang flags to be set.
This changes the way we default the is_clang value. Forcing it to true prevents it from being overridden manually. In some cases, like mac, this is pointless. In other cases like Linux, we may want to turn it on and off. Changing this definition allows us to toggle it on and off for different toolchains as we desire.
Ideally I think the way of defining toolchains would be different to make this a bit more flexible, I'll think about this. But this will work for now.
R=thakis@chromium.org
Review URL: https://codereview.chromium.org/385823002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@282685 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
New components:
- rappor
- data_reduction_proxy
- captive_portal
- translate
- usb_service
- web_modal
Removes the _IMPLEMENTATION define for web_modal. It is unused (it is a static library, not a component).
Adds third_party/cld and third_party/cld_2 BUILD files.
Separates out the include dir setting of "//" and the root gen directory so. cld_2 needed to not have this due to base file name conflicts.
R=jamesr@chromium.org
Review URL: https://codereview.chromium.org/382663002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@282438 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
The //build directory in GN shouldn't refer to anything outside of itself so that non-Chrome projects can copy it and re-use the build configuration. speechd relies on third_party/speech-dispatcher. This patch just moves the loader to the speech-dispatcher target, which makes more sense anyway.
TBR=grunell@chromium.org
Review URL: https://codereview.chromium.org/377403002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@282138 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
components/autofill
components/metrics
components/policy
components/variations
third_party/libphonenumber
third_party/mt19937ar
ui/gfx/ipc
These are all requried to compile chrome/browser in some form or another.
Fills out all grit_defines (also required for chrome/browser) and adds associated flags for them.
Add the ability for generated_extensions_api to not generate a bundle.
Remove mt19937ar from .gitignore (it's just checked into the tree)
R=ajwong@chromium.org
Review URL: https://codereview.chromium.org/365793002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@281057 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This is the GN equivalent of build/java.gypi. It is a template for
creating an Android library including java code and Android resources.
It currently only compiles java files (including files in srcjars from
srcjar targets like java_cpp_template) and zips them together in a .jar
and creates the corresponding .jar.TOC.
Some of the things still to do: proguard_preprocess, android_lint,
emma coverage, dex, everything resources.
Adds android_java_library rule for base_java, guava, and jsr-305.
This add the --java-srcjars argument to javac.py. This will accept a .zip of
.java files and include those files in the compilation. This approach is
preferred over using the --src-gendirs option.
Many of the parts of building Android stuff (libraries, resources, apks)
require knowledge of the dependents of that thing. Examples: javac classpath,
for resources aapt needs to know about all dependents, dexing for an apk
needs to know about all java code going into that apk. For gyp, this is done
primarily with all_dependent_settings. There is then some of this logic in
two particular steps (dexing and proguard). These steps, when building an
instrumentation apk, need to exclude the things in the tested apk and this is
done by having the tested apk essentially write a file saying what it did in
those steps and the test apk reading that file and excluding stuff.
In GN, all_dependent_settings doesn't really work. This change introduces a new
way of calculating and using this information. Specifically .build_config files
and build_utils.ExpandFileArgs(). The build_config file for a target contains
the information that depends on dependents. The logic in write_build_config and
the logic in the template specification are very much tied together (in some
sense, write_build_config is just the part of the template specification
that can actually inspect the dependency graph). With
build_utils.ExpandFileArgs() all the other build scripts are essentially
unaware of the .build_config files and can just be written in a (mostly)
straightforward way.
A large part of the information calculated by the build_config is finding input
files to later actions. This requires that those later actions writes a depfile
that contains any inputs that are specified by the build_config (in the case of
this change, javac and the classpath files). Since a action's script shouldn't
really know about the build_config file and what information it got
from that, it is safest for the action to write *all* of its inputs into the
depfile (but to be correct it only has to write those that aren't explicitly
specified in the build files).
Depends on: https://codereview.chromium.org/341823003/
BUG=359249
Review URL: https://codereview.chromium.org/269943005
git-svn-id: http://src.chromium.org/svn/trunk/src/build@280995 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
These deps are some of those required to compile chrome/browser (so mostly those containing resources and protobufs that have generated headers). chrome/browser still has some dependencies left.
Add third_party libs libaddressinput and dom_distiller_js.
Add components dom_distiller and policy.
Separate the generate_library_loader into a separate .gni file.
Rename components_strings to just strings.
R=ajwong@chromium.org
Review URL: https://codereview.chromium.org/362583002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@280916 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This adds a lot of global enable flags for various features used by the browser.
It is not turned on because it does not compile yet. It needs some include directories and generated sources from dependencies that aren't done yet.
It is basically a line-by-line translation of the current chrome_browser.gypi. As such, it looks pretty ugly but I think keeping these in sync will be easiest.
I put in the obvious names of the GN deps. There are surely a bunch that exist in the GN build that aren't here yet but I'm going to save this for later passes.
R=thestig@chromium.org
Review URL: https://codereview.chromium.org/356303002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@280693 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This adds the --depfile option to a bunch of android's build scripts
(the ones used by gn). It also adds a simple function in build_utils to
calculate the paths of the imported python files.
Currently, the written depfiles just contain that list of python files
(which is much more accurate than the way this is handled in gyp).
BUG=359249
Review URL: https://codereview.chromium.org/341823003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@279546 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
Strictly speaking this flag isn't here to force a deterministic build, it's here to help us to make sure that no build metadata ends up in the build artifacts to help us guarantee a deterministic build.
(See the discussion in https://codereview.chromium.org/324403006/ for more details)
FTR, Nico said he'll take care of getting rid of __DATE__ and __TIME__ so this
new define will not be necessary for that.
On the other hand we may have issues with absolute path in __FILE__ and whatever
unforeseen C++ issue. This would require #ifdef'ing them out. This is not
something devs would want to do normally so this flag is off by default.
Eventually we'll want to fix the build system to not have to hack __FILE__ out
by ensuring they are *always* relative paths but until then, this variable can
be used to make progress
R=maruel@chromium.org, thakis@chromium.orgTBR=brettw@chromium.org
BUG=314403
Review URL: https://codereview.chromium.org/330813003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@278053 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
There's more optimization done on a PGO build than on a regular one, this results in some "Unreachable code" warnings that we don't see normally.
But probably anything that this would catch that wouldn't be caught in a normal build isn't going to actually be a bug, so the incremental value of C4702 for PGO builds is likely very small and blocks the setup of the PGO bots.
BUG=380175
Review URL: https://codereview.chromium.org/315323002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@276737 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
Change some targets to source sets. These are small ones that are currently depended on by other source sets. This avoids some duplicate symbols.
duplicated.
Fix some conditions around the touch_device in ui/base.
Lots of Skia build fixes. It turns out our skia library does not use the "util" target, and some of the files in "util" reference functions that it seems are never defined in any file. This removes the ones in our list that weren't in skia_library.gypi to make accessibility_unittests link.
Update gesture recogniser file lists.
BUG=377890
TBR=scottmg
Review URL: https://codereview.chromium.org/305993002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@274022 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
They are necessary otherwise grit won't define IDS_WEB_FONT_FAMILY and
IDS_WEB_FONT_SIZE from app_locale_settings.grd
Which is necessary, otherwise we get the following error when trying to
compile web_ui_util.cc:
../../ui/base/webui/web_ui_util.cc:130:28: error: 'IDS_WEB_FONT_FAMILY'
was not declared in this scope
../../ui/base/webui/web_ui_util.cc:131:26: error: 'IDS_WEB_FONT_SIZE'
was not declared in this scope
BUG=367591
TEST=gn gen out/Debug_gn && ninja -C out/Debug_gn ui_base
R=brettw@chromium.org
TBR=ben (for ui/base change)
Review URL: https://codereview.chromium.org/296933017
git-svn-id: http://src.chromium.org/svn/trunk/src/build@274019 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This adds ppapi-related targets except the tests to the GN build.
Did some minor tweaks to the .gypi ppapi lists. Some files were duplicated between the trusted and common lists which causes errors in GN (it doesn't uniquify the lists like GYP). Common is used in every place trusted is, so I just removed the duplicates from the trusted list.
Fixed some warnings in skia and tcmalloc. Added the ability to undefine NOMINMAX on Windows.
Fixed some src/ui Windows compilation issues.
BUG=
R=scottmg@chromium.org
Review URL: https://codereview.chromium.org/305513003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@273016 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This is the GN version of build/java_cpp_template.gypi.
It defines a template that wraps an action_foreach that generates .java
files using the host C preprocessor.
The major difference in the GN version is that it takes all the
generated java files and zips them together in a single .srcjar. When
such a target is included in the srcjar_deps of a java library, the
.java files in the .srcjar will be treated much like files listed in
that libraries java_sources (particularly they will be compiled and
included in the .jar/.dex).
Depends on https://crrev.com/264773014/
BUG=359249
Review URL: https://codereview.chromium.org/264923007
git-svn-id: http://src.chromium.org/svn/trunk/src/build@272775 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
Adds a protobuf target and a template for defining protos. This handles doing host builds of protoc when cross-compiling.
The command buffer sevice wants more optimization on Android, so this patch generalizes the solution. Normally Android, like Windows, optimizes for size. This now adds support on Android for speed optimization just like Windows does. Mac & Linux always optimize for speed.
Removes the declare_args call around the gcc_version. This was parameterized to be different if is_android was set, which isn't valid because build parameters can't vary across toolchains. Since this value is derived from the environment, I just removed the declare_args call.
BUG=
R=piman@chromium.org
Review URL: https://codereview.chromium.org/289953004
git-svn-id: http://src.chromium.org/svn/trunk/src/build@271941 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
Parameterize webrtc enabling.
Add use_seccomp_pbf flag
Add a webkit version target
Work on ui/base GN build, especially on Windows. This target is a mess (in GYP). This patch tries to make it make a bit more sense while fixing the errors I found.
Fix ui/gl target for non-Linux platform. A bunch of X11 stuff was not parameterized.
Add unscaled resources target
BUG=
R=scottmg@chromium.org
Review URL: https://codereview.chromium.org/287343006
git-svn-id: http://src.chromium.org/svn/trunk/src/build@271724 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
Adds a meta-target "//base/allocator" that forwards to tcmallor or nothing depending on the build configuration.
Write tcmalloc target definition.
Write base unit tests prefs, and some support targets from base.gypi.
Fix a bunch of added files in the base target that were exposed by linking the unit tests.
Removes toolkit_uses_gtk which was removed from the GYP build.
BUG=
R=scottmg@chromium.org
Review URL: https://codereview.chromium.org/288203003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@271079 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This adds special handling for variables that begin with underscores, which is inspired by Dart. Such variables are not imported when doing an import, which gives .gni files a way to have intermediate private variables that won't pollute the scopes of the files that include them.
This also applies to the root build config, which can have private values now.
Adds some missing unused variable checks. This was disabled because processing imports would mean all imported variables were unused, and files not using all of them would get unused variable errors. This adds the option to mark merged values on a scope as used, which is used for imported values.
BUG=341738
R=cjhopman@chromium.org
Review URL: https://codereview.chromium.org/287693002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@271078 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
Some minor tweaks in STL usage in GN itself to enable it to compile using the Android version of the STL. Enable this in the build (not so much because we need a GN binary on Android, but for build verification purposes).
Moved the executable_ldconfig config from the linux file to the gcc one since its shared between the Linux and Android builds. Added "-Bdynamic" and "-Wl,-z,nocopyreloc" to this on Android.
Moved some sysroot path components from sysroot.gni to android/config.gni (which sysroot uses) so it can be shared with the toolchain definitions.
Added the android_full_debug build flag. Made the "optimize off" mode of the build match the GYP build's "light optimization" on Android contingent on this flag.
Pulls out the optimize and optimize_max shared flags into one list to avoid duplication.
Adds a bunch of linker optimization flags that should be passed on non-Mac Posix platforms, and turns on dead code stripping for Mac builds.
Adds functionality to the gcc toolchain template to be able to insert strings before and after the libs. Adds a wrapper template for android toolchains that sets these accordingly to get the gross Android crtbegin/end files inserted in the right place on the linker line.
Made the android_ndk_root variable relative to the source root rather than the system root. Uses of this now rebase according to their own needs which makes some of the arguments a lot easier to follow.
Build file updates for base and libevent for Android. Implement ashmem library.
The only change on desktop linux is the addition of -Wl,--fatal-warnings to the linker line.
R=ajwong@chromium.org, cjhopman@chromium.org, scottmg@chromium.org
Review URL: https://codereview.chromium.org/275703003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@270138 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
Previously "gn args" printed build arguments. That has now moved to "gn args <outdir> --list" along with a new --short mode that's a bit more convenient for some uses.
Now "gn args" runs your editor on the build arguments for the given output directory, and re-generates the build given those arguments. This is an easier way to manage changing build arguments for a build (previously you would have to re-type everything or edit build.ninja manually).
"gn gen" now always uses the existing arguments for the given output dir, unless "--args" is manually specified (giving the old behavior of just using those). This also allows a more convenient way for a user to recover from a borked build (sincetimes I ran into a state where something was missing that prevent ninja from even starting enough to rebuild the build).
I removed the "show" and "refresh" ninja phony rules since the new commands cover those cases.
This patch adds some additional tracing to build startup since I noticed it was missing when trying to figure out why the args command was so slow (I fixed the main reason, it was with new code I added).
Added proper escaping for printing string values and unit tests for these.
Two minor build file fixes.
R=scottmg@chromium.org
Review URL: https://codereview.chromium.org/265693003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@268042 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This adds extension detection for lower-case ".s" assembly files (used in NSPR) in addition to the existing handling of uppercase ".S" ones.
Minor fixes for build files for Mac for dependencies and files.
Proper handling of USE_NSS and addition of USE_OPENSSL_CERTS that match the GYP build.
R=rsleevi@chromium.org
Review URL: https://codereview.chromium.org/242173002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@265064 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This adds a template for using the jni_generator. Creates the
base_jni_headers target and updates the base target to depend on it and
actually use the sources that required it.
Adds //third_party/android_tools:cpu_features required by one of the
base files that also required jni.
Disable a bunch of third_party targets for Android builds.
BUG=359249
Review URL: https://codereview.chromium.org/237973002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@264695 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This fixes a lot of minor mistakes (mostly missing/extra files and flags) for the Mac GN build.
I separated out some clang flags into a config for extra clang warnings. Several of the third party libraries needed to remove this.
Removes the use_nss flag and uses !use_openssl. This is a result of discussion with rsleevi.
Removes extra duplicate net build file from secondary tree.
ui/gesture_events seems to be getting compiled in GN with more strict warnings than in GYP. Rather than fix this, I fixed the warning in the gesture recognizer unit test. It was returning a const copy (the const is pointless when you're copying).
This also removes a bunch of old GYP integration stuff that was left in the GN build.
R=scottmg@chromium.org
Review URL: https://codereview.chromium.org/239543013
git-svn-id: http://src.chromium.org/svn/trunk/src/build@264626 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This does the net target. I started doing some more that are commented out at the bottom, and there are a few more after that do do. The deps of these targets started to build up to an unreasonable level and this should be a good checkpoint.
Fix grit rule.
Fix SSL dependent configs
Add TLD cleanup
Add gconf and gio targets for Linux.
Add sources filtering for ChromeOS, .rc, and .mm files. Remove built-in code that removes .rc and .mm files (this wasn't quite complete so causes problems).
BUG=
R=scottmg@chromium.org
Review URL: https://codereview.chromium.org/236713002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@263967 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This fixes enough issues so that we can compile and link some libraries, but
it is not yet a working build.
Work remaining:
- base needs JNI support ('base_jni_headers', etc.)
- we need to figure out what should link and/or be APKs to run the tests
- we need to adjust GN to only build in the host config, not the target config.
R=brettw@chromium.org, cjhopman@chromium.org
TBR=jam
BUG=360936
Review URL: https://codereview.chromium.org/227673008
git-svn-id: http://src.chromium.org/svn/trunk/src/build@263440 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This also removes some patterns from the sources assignment filter, and adds scary comments not to make it bigger. The GYP version of this is out of control so I want to set a clear policy of what is included and not.
I removed X-related stuff from the filter (there are only about 50 files around the tree) and added manual rules for the affected files.
BUG=
R=scottmg@chromium.org
Review URL: https://codereview.chromium.org/235983002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@263394 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This makes the GN build of Skia compile and adds it to the GN build.
Previously, the configuration of the GN Skia tried to match the confusing array
of skia targets. This new version dispenses with that and just adds all files
(except SSE ones) to one target. I'm not even sure if it's necessary to split
out the SSE ones, but it seems nice since that target will get more complicated
when we add Arm and MIPS support.
I audited the defines again and added some to the main build config that had
been added to the GYP build recently.
This improves some operator error messages that I found confusing when I had errors in my file.
BUG=
R=djsollen@google.com, scottmg@chromium.org
Review URL: https://codereview.chromium.org/214883005
git-svn-id: http://src.chromium.org/svn/trunk/src/build@262779 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This adds an error check when invoking templates which caused really confusing messages if the template invocation encountered an error (because we'd continue running).
Hooks up the provider for programatically defined variables in template invocations so those can be used.
Sets the current directory in a template invocation to be that of the invoking file.
No longer define the target-related programatic variables in an import. Using these in an import will give the directory relative to the import, which is probabyl not what you want.
Fix the Windows build by adding a missing library. Add a warning not to add more to the main list (this added .lib is pretty obscure).
BUG=
R=cjhopman@chromium.org
Review URL: https://codereview.chromium.org/226223006
git-svn-id: http://src.chromium.org/svn/trunk/src/build@262524 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This is enough for this to work OK:
[gn-2013]d:\src\cr\src>gn gen out_gn
[gn-2013]d:\src\cr\src>ninja -C out_gn gn
[gn-2013]d:\src\cr\src>out_gn\gn gen out_gn2
[gn-2013]d:\src\cr\src>ninja -C out_gn2 gn
...
[297/297] LINK gn.exe
I'm not sure if we need all the deleted stuff for something
else, but I guess we can revive it when we need it.
(There's some dependency problem with linking atm, as
"ninja -C out_gn gn" always relinks, but I think it's unrelated.)
R=brettw@chromium.org
BUG=297677
Review URL: https://codereview.chromium.org/204663004
git-svn-id: http://src.chromium.org/svn/trunk/src/build@258327 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
A bunch of mac builder bots are still on 10.6 (http://crbug.com/353529)
> roll clang 198389:202554
>
> This change contains several big changes, many of the caused by upstream
> llvm enabling c++11 in their codebase and thus having higher minimum
> requirements for their toolchain and runtime environment:
>
> * clang on OS X now requires libc++, which means it'll only run on OS X 10.7+,
> so change the deployment target from 10.5 to 10.7 (we were planning on
> requiring 10.8 or 10.9, see bug 335325, but since several people are waiting
> on a new clang, let's go to 10.7 only for now and do more as a follow-up)
> * On OS X, the compiler is only usable with libc++ if libc++ is checked out
> and `make install` is run, so do that for the phase 1 compiler.
> (Chrome doesn't use libc++ yet, so we don't do that for the bootstrapped
> compiler.)
> * clang on linux now needs libstdc++ 4.7+ which isn't present on precise.
> So we now include a newer libstdc++ in the compiler bundle (meaning
> this clang should still run on precise).
> * clang is broken for android at r202554. To unblock oilpan, merge three
> later changes into this roll, llvm r202793, r203601, and r203635
>
> Regular, less tricky clang roll stuff:
> * Disable the new warning -Wabsolute-value for now (crbug.com/351479
> for turning it on)
> * Disable -Wno-tautological-pointer-compare for libxml where it now
> fires.
> * Since this bundles a new version of the oilpan plugin, bump its suffix to _3
> * Change plugin code to track upstream api change
> * Remove plugin flag check-url-directory as that's now always on
>
> Due to the c++11 stuff, it's likely that this change will get reverted a few times.
>
> BUG=351479, 287029,341352,333968,335325
> TBR=hans@chromium.org
>
> Review URL: https://codereview.chromium.org/195623002TBR=thakis@chromium.org
Review URL: https://codereview.chromium.org/203313004
git-svn-id: http://src.chromium.org/svn/trunk/src/build@257620 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This change contains several big changes, many of the caused by upstream
llvm enabling c++11 in their codebase and thus having higher minimum
requirements for their toolchain and runtime environment:
* clang on OS X now requires libc++, which means it'll only run on OS X 10.7+,
so change the deployment target from 10.5 to 10.7 (we were planning on
requiring 10.8 or 10.9, see bug 335325, but since several people are waiting
on a new clang, let's go to 10.7 only for now and do more as a follow-up)
* On OS X, the compiler is only usable with libc++ if libc++ is checked out
and `make install` is run, so do that for the phase 1 compiler.
(Chrome doesn't use libc++ yet, so we don't do that for the bootstrapped
compiler.)
* clang on linux now needs libstdc++ 4.7+ which isn't present on precise.
So we now include a newer libstdc++ in the compiler bundle (meaning
this clang should still run on precise).
* clang is broken for android at r202554. To unblock oilpan, merge three
later changes into this roll, llvm r202793, r203601, and r203635
Regular, less tricky clang roll stuff:
* Disable the new warning -Wabsolute-value for now (crbug.com/351479
for turning it on)
* Disable -Wno-tautological-pointer-compare for libxml where it now
fires.
* Since this bundles a new version of the oilpan plugin, bump its suffix to _3
* Change plugin code to track upstream api change
* Remove plugin flag check-url-directory as that's now always on
Due to the c++11 stuff, it's likely that this change will get reverted a few times.
BUG=351479, 287029,341352,333968,335325
TBR=hans@chromium.org
Review URL: https://codereview.chromium.org/195623002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@257618 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This hooks up the rpath for Mac executables. This fixes a bug in iOS GYP generation where the architectures were specified as a string instead of a list.
Also fixes some base files that were out-of-sync from the GYP build.
BUG=345021
TBR=thakis
Review URL: https://codereview.chromium.org/179873008
git-svn-id: http://src.chromium.org/svn/trunk/src/build@253952 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This makes the -D defines for the default desktop Linux build of re2 in GN match that of GYP.
This moves the SSL configuration stuff into the root build directory. It's better to match the GYP build exactly for ease of landing even though this is a stupid dependency.
Fixes the help text that still references the implicit "gen" command (I removed this).
R=scottmg@chromium.org
Review URL: https://codereview.chromium.org/170533003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@252981 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This is for the base shared library target on desktop Linux.
The remaining difference on Linux is that GYP's base build specifies -Wno-write-strings on Linux. This dates back to the first GYP file. I don't know why we would specify this for only base, and it seems not to trigger any warning without this flag, so I didn't add it.
This changes the pkg-config invocation from --libs to --libs-only-L and --libs-only-l. The output seems slightly different (--libs was producing an extra --export-dynamic for gmodule-2.0 that aren't present when invoking the other way, which is what the GYP build does).
R=thakis@chromium.org
Review URL: https://codereview.chromium.org/165873004
git-svn-id: http://src.chromium.org/svn/trunk/src/build@252409 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
Updates all rebase_path calls to use the new parameter ordering.
Use the new getenv function and delete the Python script we had to get the home directory.
Takes advantage of the new ability to have the default value of declare_args blocks be dependent on other values. This simplifies some code.
R=scottmg@chromium.org
Review URL: https://codereview.chromium.org/164773005
git-svn-id: http://src.chromium.org/svn/trunk/src/build@252065 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This stops doing the host build as a completely separate build and
instead identifies the proper host toolchain inside of the "regular"
build that we should use for host targets (if any). This should help
unify GN's concept of host/target builds with GYPs a bit more.
BUG=
Review URL: https://codereview.chromium.org/152683004
git-svn-id: http://src.chromium.org/svn/trunk/src/build@250689 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This properly sets the DEPLOYMENT_TARGET in the generated GYP file (previously this value wa salways hardcoded to empty -- whoops!).
Sets the simulator properly. Previously when not using an explicit SDK the Xcode generator would still end up with a simulator build. This turns the flag into a tri-state so gyp_chromium can set it explicitly when an SDK is specified, but otherwise we can trigger default behavior based on Xcode/Ninja.
TBR=scottmg
Review URL: https://codereview.chromium.org/155983003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@249643 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
In GYP Mac, most targets got a default set of mac libraries by virtue
of depending on base, which due to a GYP bug pushed the libraries to
all targets depending on base, even separate shared libraries.
In GYP Windows the set of default libraries on Windows is global and
this keeps the same list and behavior.
The lists of libraries could use some cleanup, btu this patch just
attempts to keep the status quo in a cleaner way rather than fixing the
unreasonable list of libraries (especially on Windows).
R=scottmg@chromium.org, scottmg
Review URL: https://codereview.chromium.org/131513026
git-svn-id: http://src.chromium.org/svn/trunk/src/build@249600 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This hooks up Clang in the pure GN build (not the hybrid GYP mode).
Clang gave some warnings about compiler flags which made me realize that we were giving linker flags to the compiler in some cases. So I enhanced our pkg-config wrapper to to add libraries to the cflags (missing else), return ldflags separately, and also strip out -pthread which was getting included over and over.
BUG=
Review URL: https://codereview.chromium.org/147923010
git-svn-id: http://src.chromium.org/svn/trunk/src/build@248477 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
Separates out some "SDK" related setup on Mac land Linux like we do on Windows. I'm trying to make the "compiler" BUILD file just compiler warnings and CPU options, and have system library stuff on the platform directories.
This adds the capability for GN to produce GYP files on Mac that vary according
to the GYP generator as well as target-vs-host. I added a bunch of logic to the
GN iOS build to set up stuff accordingly based on my current knowledge of
what's required.
Sadly, this means we now have an 8-way GN build (all combinations of
debug/release, host/target, and xcode/ninja). I did some refactoring of the GYP
code in GN to make this less unreasonable.
I checked that the GYP files look the way I want, but I didn't actually test
the resulting builds yet. There is still likely to be some conditions wrong or
things not being set properly. I'm going to follow up with a second pass based on actual testing.
I believe, however, that with this new GYP generator code in GN, we can express
in the .gn files what we need to do the iOS build.
Review URL: https://codereview.chromium.org/149163005
git-svn-id: http://src.chromium.org/svn/trunk/src/build@248476 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This hooks up detection for the "-arch" flag on Mac to set the GYP "ARCH" xcode variable. GN then removes the -arch argument from the compiler args, since GYP will then re-add it based on the ARCH value. Previously, not doing this resulting in mutliple "-arch" arguments to the compiler since GYP would always insert its own.
Disables some warnings on Windows for the re2 target to match the GYP build. The third warning (4018) that the GYP build sets is disabled globally so there's no need to do it for this target.
Hooks up some iOS SDK stuff.
BUG=336667
TBR=thakis@chromium.org
Review URL: https://codereview.chromium.org/141433015
git-svn-id: http://src.chromium.org/svn/trunk/src/build@247206 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This supports specifying the output_dir generator flag in the gn-generating-gyp
mode. The value of the output_dir flag is mapped to a gyp variable called gyp_output_dir
and a GN argument by the same name. References from gyp to gn-generated gyp files must
use this variable and references from BUILD.gn files must use the gn argument.
BUG=335760
TBR=brettw for build/config/BUILDCONFIG.gn (looked at in person)
Review URL: https://codereview.chromium.org/142223002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@245800 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This emulates optimize=max from GYP. One difference is that this applies for all build types. The GYP one only applied optimize=max for the Official build. I can't think of why this should be the case, and it seems confusing to have such arbitrary official build differences.
R=bbudge@chromium.org
Review URL: https://codereview.chromium.org/138903002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@244839 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This makes make_global_settings closer to the GYP version. Previously it had relied on GYP concatenating two lists when running with goma, but it doesn't actually behave this way, resulting in unexpected output.
TBR=scottmg
Review URL: https://codereview.chromium.org/132833002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@244104 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
Adds arm version and some android configuration build flags.
This adds most of the logic from common.gypi to the Android GN build.
This is currently missing the crtbegin/end stuff and won't actually make real Android builds. The logic in this patch is just the initial conversion that will require testing and several more passes of fixes.
R=torne@chromium.org, torne
Review URL: https://codereview.chromium.org/121173005
git-svn-id: http://src.chromium.org/svn/trunk/src/build@243681 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
Refactors the toolchain definitions by adding an include file to help defining GCC toolchains. This eliminates most of the duplication on Linux and Android. Adds make_global_settings for most Linux and Mac builds that should match the GYP build (this will require some more passes to get every place to be an exact match).
Fixes Android build issues in WebKit. Fixes the definition of is_android.
Renames the Linux toolchains to use the GN name of the CPU architecture so one can refer to them without conditionals. Remove conditionals from BUILDCONFIG.gn
Adds the gyp_header variable to the help (it wasn't getting inserted into the list of available variables).
R=thakis@chromium.org
Review URL: https://codereview.chromium.org/117863003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@242674 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This adds support similar to the GYP scripts for running pkg-config in the presence of sysroots.
It also adds the linker flags by calling the existing horrible sysroot_ld_path.sh shell script that extracts stuff from the sysroot's configuration.
With this change, official 32-bit linux builds work in the checked in 32-bit Debian sysroots.
R=piman@chromium.org
Review URL: https://codereview.chromium.org/116803003
git-svn-id: http://src.chromium.org/svn/trunk/src/build@241397 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
This passes the official build and branding flags to the GN build. Currently, this is used to hook up the proper sysroot for compiling on Linux for official builds. The 32-bit branded build does not compile currently due to pending pkgconfig issues.
Fix a bug where assertions in the build config would not get reported properly. This was because I forgot to actually throw the error at the toplevel.
BUG=
R=piman@chromium.org
Review URL: https://codereview.chromium.org/115643002
git-svn-id: http://src.chromium.org/svn/trunk/src/build@240918 4ff67af0-8c30-449e-8e8b-ad334ec8d88c
James Robinson
scottmg
Dirk Pranke
Hans Wennborg
davidben
cmasone
mkosiba
Brett Wilson
cjhopman
fdegans
thakis
dalecurtis
pgal.u-szeged
vitalybuka
Sadrul Habib Chowdhury
noms
cjhopman@chromium.org
cjhopman@chromium.org
brettw@chromium.org
dnicoara@chromium.org
jamesr@chromium.org
thakis@chromium.org
garykac@chromium.org
scottmg@chromium.org
pkasting@chromium.org
spang@chromium.org
derat@chromium.org
aurimas@chromium.org
sky@chromium.org
kangil.han@samsung.com
anton@chromium.org
Gordana.Cmiljanovic@imgtec.com
sebmarchand@chromium.org
tfarina@chromium.org
ch.dumez@samsung.com
mvanouwerkerk@chromium.org
henrika@google.com
fdegans@chromium.org
hans@chromium.org
ckocagil@chromium.org
dpranke@chromium.org
mithro@mithis.com
jam@chromium.org
newt@chromium.org
petarj@mips.com
bbudge@chromium.org
piman@chromium.org