Commit graph

66 commits

Author SHA1 Message Date
Victor Chang
b54f5aa359 "module_current" and "system_server_current" should contain ART's @SystemApi(MODULE_LIBRARIES)
Before this fix, compiling a java_library against sdk_version:
"module_current" can't use the @SystemApi(MODULE_LIBRARIES) provided
by the ART module because the system module "core-current-stubs-system-modules"
contains only the public APIs.

Use the new system module with module lib APIs.

Bug: 183097033
Test: m droid
Change-Id: I274e2710d1ff34e896aa620bfafb4481180c53b5
2021-06-29 22:05:58 +01:00
Andrei-Valentin Onea
b75031e918 Merge "Replace source based system server stubs" 2021-04-28 15:06:22 +00:00
Paul Duffin
d3c1513df9 Add tempPathForRestat to improve consistency
Previously, there were two approaches used to construct the temporary
path needed by the commitChangeForRestat method and neither was
suitable for general use. One was:
    android.PathForOutput(ctx, outputPath.Rel()+".tmp")
The other was:
    combinedAidl.ReplaceExtension(ctx, "aidl.tmp")

The first approach would not work if the supplied path had been created
by PathForModuleOut() or similar as it would drop the module directory.
That could lead to the same path being used for multiple rules.

The second approach would not work for files with a different
extension.

The tempPathForRestat combines the two approaches so it can be used
generally with any WritablePath.

Bug: 179354495
Test: verified that the ninja rules that used these files were not
      changed by these changes.
Change-Id: I4439dea0a823512c281eeb1366522fb49dceb4e3
2021-04-21 23:26:57 +01:00
Jiyong Park
9231537fe2 SdkSpec is fully using ApiLevel
Previously, SdkSpec was constructed only from the user string. It didn't
make use of the Config struct where information about the latest stable
SDK version, etc. is recorded. As a result, the build system couldn't
check if the sdk version "current" is referring to the in-development
(i.e.  not-yet-frozen) SDK version or the latest stable version.
"current" was always assumed to be in-development (IsPreview() returns
true) even when Platform_sdk_final == true.

As the first step for fixing that, this change requires
android.EarlyModuleContext to be passed when constructing SdkSpec from
the user string.

In the following changes, "current" will be mapped to either
FutureApiLevel (10000) or one of the FinalApiLevels() depending on
whether the platform SDK was finalized or not.

Bug: 175678607
Test: m
Change-Id: Ifea12ebf147ecccf12e7266dd382819806571543
2021-04-08 11:27:24 +09:00
Jiyong Park
54105c48f4 SdkSpec = Scope + ApiLevel
SdkSpec.Version was an int type. Now it becomes ApiLevel type which
is a better abstraction of the version (or api level).

Bug: 1655587
Test: m

Change-Id: I4d67b9b9eae45f653b6af4f5b73da9e091b3dfab
2021-04-05 09:32:06 +09:00
Jiyong Park
f1691d2a2c Move java.sdkSpec to the android package
... in preparation for making the handling of sdk versions consistent
across java and cc modules.

Bug: 175678607
Test: m
Change-Id: I598f0454bce9b7320621022115412fbe97403945
2021-04-03 08:25:12 +09:00
Andrei Onea
f1f738e653 Replace source based system server stubs
Use the ones obtained by merging the ones in non-updatable code, and the
ones exposed in mainline module jars that are part of
SYSTEMSERVERCLASSPATH.

Test: builds
Bug: 177640454
Change-Id: Ib4719d7eb8b76a0e5af79c1270e2f1e344682423
2021-03-17 13:37:19 +00:00
Anton Hansson
404adeefdc Merge "Fix api_fingerprint.txt generation" 2021-02-15 11:31:37 +00:00
Anton Hansson
973d31c757 Fix api_fingerprint.txt generation
Use the generated txt files as input instead
of globbing for now non-existent files.

Also add the system server api txt to the hash.

Bug: 179807354
Test: verify out/soong/api_fingerprint.txt changes now
Test: verify the command in the ninja file looks ok:
      grep api_fingerprint out/soong/build.ninja
Change-Id: I7a49ca134eb93b22537a4f3054285eb15f8c4256
2021-02-11 09:17:10 +00:00
Colin Cross
dcf71b299c Convert java.Dependency to JavaInfo provider
Export information about java dependencies through a Provider
instead of accessing the module directly.

Test: java_test.go
Test: no changes to build.ninja
Change-Id: Ifc5d566bf6f6ebc0ad399e948effaa1ef6a22876
2021-02-09 15:36:25 -08:00
Jaewoong Jung
18aefc1977 Remove unnecessary snake case variables.
Test: m nothing + TreeHugger
Change-Id: I99f7162944daa6c57c6ae4763261e108bb5cb6b1
2020-12-22 12:38:35 -08:00
Colin Cross
f1a035e6be Pass pctx and ctx to NewRuleBuilder
Enable the RuleBuilder and RuleBuilderCommand methods to access
the BuilderContext by passing it to NewRuleBuilder instead of
RuleBuilder.Build.

Test: genrule_test.go
Test: rule_builder_test.go
Test: m checkbuild
Change-Id: I63e6597e19167393876dc2259d6f521363b7dabc
2020-12-01 16:22:16 -08:00
Paul Duffin
fb6ae5bc73 Add support for using sdk_version: "module_<ver>"
Fixes a panic when trying to build something with
sdk_version: "module_30".

Test: m nothing
Change-Id: I9b2ce50957f59e2bead335ffa58888e15cda1f78
2020-09-30 16:20:17 +01:00
Dan Albert
4f378d75aa Convert more versions in config to ApiLevel.
The test case I removed is invalid. The codename has had its int
assigned, but the config claims it is not final.

If this ever does need to be supported it's just a matter of making
sure the Q -> 29 mapping (or whatever) in the finalized codenames map
in android/api_levels.go.

Test: treehugger
Bug: http://b/154667674
Change-Id: I4f42ec2fd4a37750519ee3937938a1c65b6bb1e8
2020-09-22 16:01:56 -07:00
Dan Albert
0b176c8038 Replace FutureApiLevel with an ApiLevel.
Keeping the int constant around for now as FutureApiLevelInt because
it's still useful in places that haven't adopted ApiLevel yet for
testing if their non-ApiLevel API level is current or not.

Test: treehugger
Bug: http://b/154667674
Change-Id: I47a7012703f41fdeb56f91edf9c83afa93042deb
2020-09-22 15:04:48 -07:00
Dan Albert
c8060536e8 Replace ApiStrToNum uses with ApiLevel.
Test: treehugger
Bug: http://b/154667674
Change-Id: I2954bb21c1cfdeb305f25cfb6c8711c930f6ed50
2020-09-22 15:04:48 -07:00
Jeongik Cha
219141c6bb Introduce BOARD_CURRENT_API_LEVEL_FOR_VENDOR_MODULES
If BOARD_CURRENT_API_LEVEL_FOR_VENDOR_MODULES has a numeric value,
it replaces "current" or "system_current" with the version which
the flag indicates.

Bug: 163009188
Test: BOARD_CURRENT_API_LEVEL_FOR_VENDOR_MODULES=29 m, and then check if every vendor
java module's sdk_version is 29 if its sdk_version was current.

Change-Id: I17b49b8e02caf2d1bc57b91648d4420f3ad9fcb9
2020-08-27 23:09:54 +09:00
Dan Willemsen
9f43597ff7 Remove obsolete PDK build functionality
This hasn't worked for a couple years, and continues to bitrot. Just
remove it.

Adds a bpfix rule so that we can eventually remove the
product_variables.pdk definition, which is now always a no-op.

Test: treehugger
Change-Id: I830b54d419b59f6db1d4617b45e61a78234f57a7
Merged-In: I830b54d419b59f6db1d4617b45e61a78234f57a7
2020-08-11 01:11:44 +00:00
Jeongik Cha
816a23a50d Introduce AlwaysUsePrebuiltSdks
Instead of UnbundledBuild, use AlwaysUsePrebuiltSdks
to determine if java modules needs to be built against prebuilt sdks.
And rename UnbundledBuildUsePrebuiltSdks to AlwaysUsePrebuiltSdks to
express its behavior more correctly.(It can be orthgonal to "Unbundled")

Bug: 160390776
Test: TARGET_BUILD_UNBUNDLED_IMAGE=true m vendorimage

Change-Id: I0be7265c1959d8774c295372cd7a9250169f6df9
2020-08-07 12:15:52 +09:00
Patrice Arruda
b481b87104 Inclusive fix: Replace sanity to validation in java/sdk.go
Bug: b/161896447
Test: "lunch 1" and "m nothing"
Change-Id: I8cbdec02dccd961a8cac92c6e957aeebd1e263c1
2020-07-28 18:30:44 +00:00
Pete Gillin
84c3807b99 Enforce the legacy core/platform API restriction.
This change silently decides whether modules which depend on either
sdkPrivate or sdkCorePlatform get the legacy or the stable version of
the core/platform API, based on whether the module's name is on a
hard-coded list or not.

Test: m java
Test: make a target from the list when its entry is commented out, which correctly fails
Bug: 157640067
Change-Id: I15e5a6c2f07e73718803501d705de0d7ab9bec90
Merged-In: I15e5a6c2f07e73718803501d705de0d7ab9bec90
Merged-In: Iaa97ddaa015e8079fcb3426585c5101c7ec9e22a
(cherry picked from commit c0f4373106)
2020-07-23 09:06:52 +01:00
Treehugger Robot
759ccc13af Merge "Tidy up the definition of sdkCore." 2020-07-06 12:48:44 +00:00
Treehugger Robot
756aa63aee Merge "Rename module -> module-lib" 2020-07-01 23:58:07 +00:00
Pete Gillin
880f964f16 Tidy up the definition of sdkCore.
Compared to the effect of the toModule function, this (a) sets
noFrameworkLibs to true, which seems logically correct, and (b) stops
setting java9Classpath, which was redundant since the same stubs are
available in 1.9+ builds via the systemModules setting. (This brings
sdkCore into line with sdkCorePlatform in both cases.)

Test: m art.module.api.annotations
Change-Id: I61a868bc363a3bf174725b78c0a1657021cfe08a
2020-07-01 13:17:16 +01:00
Pete Gillin
7b0bdce69e Remove frameworkResModule from sdkCorePlatform.
This seems unnecessary: modules using `sdk_version: "core_platform"`
don't expect frameworks dependencies.

Test: m framework
Change-Id: Ib7ad715d1f4b4934c3c4a84839f4ead85a5abb29
2020-07-01 13:05:32 +01:00
Pete Gillin
e3d44b245b Remove the concept of useDefaultLibs from Soong.
This field in the java/sdk structure was used in two of the many
possible configurations, so it wasn't really a "default". It also
meant that, to understand those configurations, the reader had to know
what was considered the default, which was only possibly by reading
the code in java.go and droiddoc.go which implemented special code
paths when useDefaultLibs was true. By eliminating that setting and
explicitly setting the required values, the code is simpler and easier
to understand.

This change is a straight refactoring, in the sense that the output of
the build should be unchanged.

Regarding the changes to the proguardRaiseTag dependency in java.go:
- This is a noop for anything which had sdkDep.useModule = true prior
  to this change, because they all had the same value for
  hasFrameworkLibs() and hasStandardLibs().
- This is a noop for anything which had sdkDep.useDefaultLibs = true
  prior to this change, because they do not use proguard settings.
- Therefore, it is a noop overall.
- Nevertheless, it is required to make sdkCorePlatform work. Without
  this change, such modules would pick up a dependency on framework
  libs via the (unused) proguardRaiseTag, which creates a circular
  dependency, because this is the sdk_version used when building
  framework libs themselves.

Bug: 157640067
Test: m java docs droid
Change-Id: I3a83b5edc1bd48c16b55f6f77e3e710fc8fbd8fa
2020-07-01 12:40:58 +01:00
Jiyong Park
49062c89f9 Rename module -> module-lib
When sdk_version: "module_current" in an unbundled build, previously
prebuilt/sdk/current/module/android.jar was used. But that path is
wrong; we have prebuilt/sdk/current/module-lib/android.jar for the
module APIs. Fix the mismatch by changing the string name of the module
API from "module" to "module-lib".

Also, "system_server" for system server API is changed to
"system-server" to match with the path
/prebuilts/sdk/current/system-server/.

Bug: N/A
Test: https://android-build.googleplex.com/builds/forrest/run/L86100000601675127
Change-Id: I23e9befbd9592d3b22989c2c8fbf34d434273288
2020-06-17 21:31:00 +09:00
Anton Hansson
3f07ab2d63 Create a framework.aidl for non-updatable platform
framework.aidl includes the parcelables from the non-updatable
part of the framework as well as the modules. This causes a
dependency cycle when building module stubs:
module_stub -> module_sdk -> framework.aidl -> public_sdk -> module_stub

The module_sdk only includes the the stubs for the non-updatable
part of the framework, so it should also only contain the non-updatable
parcelables. This change creates a framework_non_updatable.aidl with
those parcelables, and updates module_current to use that.

Bug: 144149403
Test: m
Test: m && diff out/soong/framework{,_non_updatable}.aidl
      (the diff contains just TestApi + module parcelables)
Change-Id: I224117a0ff695c22d4a4317a51a9b775ed73066b
Merged-In: I224117a0ff695c22d4a4317a51a9b775ed73066b
(cherry picked from commit 85c151c3f3)
2020-06-05 15:16:16 +01:00
Colin Cross
17dec171b4 Use system modules for prebuilt SDKs >=30
Prebuilt SDKs >=30 now contain core-for-system-modules.jar,
convert them to system modules and use them when compiling against
the SDK to allow using javac -source 1.9 -target 1.9.

Bug: 117069453
Test: TestClasspath
Change-Id: Iebadad5980b952ed91c3ffd56cff1ce1827d3247
Merged-In: Iebadad5980b952ed91c3ffd56cff1ce1827d3247
2020-05-15 17:21:58 -07:00
Artur Satayev
8cf899afcc Ensure APEX's Java deps use stable SDKs.
Test: m
Bug: 153333044
Change-Id: Ib1acf3073e96fe23c92d292ec0b1a91e2cd408db
2020-04-27 19:16:40 +01:00
Artur Satayev
2db1c3f1c4 Check updatable APKs compile against managed SDKs.
As a follow up, this property will be set to APKs participating in mainline program.

Bug: 153333044
Test: m
Change-Id: I6ea2f3c1d26992259e4e9e6a6d8cecf091d39c43
2020-04-15 12:06:13 +01:00
Anton Hansson
ba6ab2e791 Make system_server stubs consistent with other stubs
Include the module_api stubs in system_server one instead of
putting both of these jars on the classpath. Also rename it
to be in line with the other stubs.

Bug: 149293194
Test: m
Exempt-From-Owner-Approval: approved internally
Change-Id: Iead5af4152a49cd59a4fd7afc0312c2f0c872c1e
Merged-In: Iead5af4152a49cd59a4fd7afc0312c2f0c872c1e
(cherry picked from commit bbd78556da)
2020-03-25 14:18:02 +00:00
Nikita Ioffe
643953d170 Merge "Always use "${codename}.${sha}" if UNBUNDLED_BUILD_TARGET_SDK_WITH_API_FINGERPRINT=true" 2020-03-11 17:45:34 +00:00
Nikita Ioffe
1f4f345156 Always use "${codename}.${sha}" if UNBUNDLED_BUILD_TARGET_SDK_WITH_API_FINGERPRINT=true
Test: m checkbuild
Bug: 149733822
Change-Id: I3e1beeb721f7e87bc6adda61861fa962ec892360
Merged-In: I3e1beeb721f7e87bc6adda61861fa962ec892360
(cherry picked from commit 934c4f2acc)
2020-03-04 22:05:11 +00:00
Jeongik Cha
7c7083163a Apply sdk version check to product apps
1. Check if system sdk version of apps from product partition is one of BOARD_SYSTEMSDK_VERSIONS.
2. The logic that check system sdk version doesn't work after aosp/1212908 because it always returns early

Bug: 147711383
Test: m
Test: set sdk_version the version that doesn't exist in BOARD_SYSTEMSDK_VERSIONS, and check if it cannot build
Change-Id: I923477cffbcd9c763ee2deb5e7cce29aa005c715
2020-03-03 18:57:54 +09:00
Jiyong Park
aae9bd11c2 sdk_version: "system_server_current"
The new sdk version "system_server_current" is for system server
components that needs to use all public APIs, system APIs, module APIs,
and the system server APIs.

Bug: 146757305
Test: m
Change-Id: I24fd5af010532a110393676607dc90889f2ec17e
2020-02-12 04:36:43 +09:00
Jiyong Park
50146e9c8e sdk_version: "module_current" is supported
module_* is a new API surface for OS modules (e.g. APEXes). It has
slightly bigger API surface than the system_* SDK. Specifically, APIs
with @SystemApi(client=MODULE_LIBRARIES) are added there.

Bug: 146757305
Test: m

Change-Id: I8980e50c0e3a4cd843048e0de1f638e854384f46
2020-01-31 23:01:34 +09:00
Jiyong Park
6a927c4e6a Abstract sdk_version string using sdkSpec type
The value format that sdk_version (and min_sdk_version, etc.) can have
has consistently evolved and is quite complicated. Furthermore, with the
Mainline module effort, we are expected to have more sdk_versions like
'module-app-current', 'module-lib-current', etc.

The goal of this change is to abstract the various sdk versions, which
are currently represented in string and is parsed in various places,
into a type called sdkSpec, so that adding new sdk veresions becomes
easier than before.

The sdk_version string is now parsed in only one place 'SdkSpecFrom', in
which it is converted into the sdkSpec struct. The struct type provides
several methods that again converts sdkSpec into context-specific
information such as the effective version number, etc.

Bug: 146757305
Bug: 147879031
Test: m
Change-Id: I252f3706544f00ea71c61c23460f07561dd28ab0
2020-01-25 21:56:43 +09:00
Baligh Uddin
cf0bba7a30 Merge "Revert "Revert "Allow codename.fingerprint format for minSdkVersion""" 2020-01-25 04:12:30 +00:00
Baligh Uddin
f62013736a Revert "Revert "Allow codename.fingerprint format for minSdkVersion""
This reverts commit 230e241f58.

Reason for revert: This is a revert of a revert.  Downstream problem has been fixed and have been validated locally and via Forrest build.

Change-Id: I89c51d25b3adb818ea44a983d0ac681a88790d8c
2020-01-24 23:15:44 +00:00
Jaewoong Jung
9befb0c326 Add runtime_resource_overlay.
Fixes: 119811120
Test: app_test.go
Test: Converted and built IconPackFilledSystemUIOverlay
Change-Id: I71841148c25f820ba829f751a201d2c771c8bd20
2020-01-23 14:24:41 -08:00
Pete Gillin
230e241f58 Revert "Allow codename.fingerprint format for minSdkVersion"
This reverts commit 014a85712d.

Reason for revert: Caused vendor/google/build/build_mainline_modules.sh to fail with `Error: minSdkVersion (10000) is greater than maxSdkVersion (30)`.

Bug: 130541924
Change-Id: Ifa233bf40a674481d21b61ee816c5fdde8201080
2020-01-21 13:44:50 +00:00
Baligh Uddin
014a85712d Allow codename.fingerprint format for minSdkVersion
Use codename.fingerprint format for minSdkVersion if it is unset
in the manifest and
UNBUNDLED_BUILD_TARGET_SDK_WITH_API_FINGERPRINT=true.

Using a utility function in sdk.go to check whether to apply
api.fingerprint.

BUG: 130541924
Change-Id: I748a25c419033bf54b63171d334644fcd0ecc78f
2020-01-20 07:46:12 -08:00
Colin Cross
07c8856904 Delay invalid sdk_version error for AllowMissingDependencies
Unbundled builds set AllowMissingDependencies and attempt to use
prebuilts for some jars.  Delay the errors for missing jars for
modules with invalid sdk_version values in unbundled builds so
that they only block the build if those modules are built.

Also fix some error messages to show the original sdk_version
value.

Bug: 146513037
Test: m TARGET_BUILD_APPS=Camera2
Change-Id: I1812ef6dc80895f7a2162a8bdbf2c5067755e9a0
2020-01-07 20:14:14 +00:00
Colin Cross
1184b647d5 Add EarlyModuleContext for LoadHookContext
Make LoadHookContext embed a new EarlyModuleContext instead of
BaseModuleContext to reduce its API surface in preparation for
moving it to run during parsing instead of mutators.

Test: m checkbuild
Change-Id: I1cd3ff3b636e7e24991a9184d7521903473e505a
2020-01-04 08:40:55 -08:00
Colin Cross
6cef481ee7 Use java language 1.9 for sdk_version: "current"
Java language was set for 1.8 for anything building against the
current SDK because the stubs were not built in Soong, so the
system modules could not be built.  The stubs have been built
in Soong since Iabd32b30954b3f4a6d9a779fde52a032b684807e, but
I5e870c34dd0ebc8ae3f888ec627da590c846a76f missed updating
this TODO.

Use 1.9 when building against the stubs, but continue using
1.8 for unbundled builds until we have prebuilt system modules.

Always use the core-current-stubs-system-modules to avoid
splitting android.* packages between the system modules and
the classpath, which would cause new classes in android.*
packages in classpath jars to be ignored.  Add a new
java9Classpath field to sdkDep that will contain the stubs jar
when targeting Java language level 1.9, and plumb it through
to javac and turbine.  Rename the modules field to bootclasspath.

Bug: 142896162
Test: m checkbuild
Change-Id: Icfd32d0a863b2303a997c7cf03cb3708aade4724
2019-10-29 13:35:39 -07:00
Colin Cross
6d8d8c6a91 Remove special case for sdk_version: "none"
The dependency handling for sdk_version: "none" (which propagated to
!sdkDep.hasStandardLibs()) was very similar to the normal useModule
case.  Combine the cases by making decodeSdkDep set useModule: true
and put the system modules in modules for the sdk_version: "none"
case.

Test: TestClasspath
Change-Id: Icc9ff4d43a38da25cc0e3628be95951d61773ad5
2019-10-29 13:35:39 -07:00
Paul Duffin
e25c644f1e Add system_modules to droidstubs
This allows droidstubs to use the same system modules to create the
stubs that will be used to compile them. It improves consistency and
avoids droidstubs having to duplicate the libraries that make up the
system modules on its libs property.

Adds systemModules() to the sdkContext which allows consistent error
checking behavior between droidstubs and java_library.

Bug: 142534789
Test: m checkbuild
Change-Id: Ib2006906d9528a900f16851f50b62152ffb51a1b
2019-10-11 16:38:14 +01:00
Elliott Hughes
34b49d1402 Use toybox md5sum instead of md5 on macOS.
Test: builds locally
Change-Id: I04aa52de96e0525a44331cb0c057a8e18a95891d
2019-09-06 14:42:24 -07:00
Colin Cross
ee94d6ab14 Add RuleBuilder helper functions for built and prebuilt tools
Replace the common pattern of:
cmd.Tool(ctx.Config().HostToolPath(ctx, "tool"))
with:
cmd.BuiltTool("tool")

And similarly for PrebuiltBuildTool.

Test: m checkbuild
Change-Id: I7d63188505362c7df6a3b3e7330b4a2cca5a2409
2019-07-11 13:05:19 -07:00