We use the timestamps in builds to determine a downgrade, which might
not be always the truth. For examples, two builds cut from different
branches may carry timestamps in a reverse order. An incremental package
won't be able to be pushed nor applied, based on the timestamp
comparison.
We used to handle such a case with manual work, by setting the
post-timestamp to (pre-timestamp + 1) in the package metadata. This CL
automates the process by adding a new flag --override_timestamp.
Note that it doesn't change anything in the installed image, but only
affects the assertions for pushing / installing the package.
With the change in this CL:
- If it's a downgrade without any extra flag, fail the package
generation (we only print warnings prior to this CL);
- If it's a downgrade with --downgrade flag, generate a downgrade
package with forced data wipe (same as before);
- If it's a downgrade with --override_timestamp, generate a normal
incremental with hacked timestamp (pre-timestamp + 1) (new in this CL
to avoid the manual change);
- If it's not a downgrade but with any of the above two flags specified,
fail the package generation.
Bug: 33744169
Test: Generate an incremental from builds with reversed timestamps.
Change-Id: I8b187d32708b4a7c3e20f8c6adb8f9527b73b965
1) Updates the LOCAL_COMPATIBILITY_SUITE line to allow for a
testcase to belong to multiple suites.
2) Building testcases no longer fails if
COMPATIBILITY_TESTCASES_OUT_<suite> is not defined. This
testcase will just not output to that directory.
This will be utilized by the device-tests and general-tests
suites that don't require any extra output besides the common
testcase directory.
Bug: 35394351
Test: 1) Added multiple *TS testcases to cts & vts and verified they
ended up in the common directory and each suite's testcase
directory. Specifically tested CtsSplitApp to ensure the
split usecase still works as well.
2) Added a CTS testcase to the device-tests suite, built
device-tests and verified the cts/android-cts/testcases
copy was not produced.
Change-Id: Ic4c4e87e62be4fc0c5e394d88cc359518346dffa
Set USE_SOONG_UI=false to disable the new soong_ui wrapper.
Test: m -j blueprint_tools
Test: USE_SOONG_UI=false m -j blueprint_tools
Change-Id: I0b217939a29fd51e69ef594f824b14e25a4c8c4d
We introduced META/misc_info.txt to hold the misc info since Gingerbread
(commit 37974731fc). Remove the backwards
compatibility support for building pre-G TF zips.
Test: `m dist` works.
Change-Id: Ibff7aaf69cc7e460634c049d11a004f7196f8f73
The SELinux policy compiler needs to be available on-device to compile
SELinux policy at boot on some devices. For now, we're including this
unconditionally, but, going forward, we will include this binary only
on devices which need it.
Test: Device boots -- secilc is not yet used anyway
Bug: 31363362
Change-Id: I7712b70d7c5b4c57bec2cdb44519b42d26758d09
Specify list of paths to exclude from coverage instrumentation.
Test: make NATIVE_COVERAGE=true COVERAGE_PATHS=hardware/interfaces
COVERAGE_EXCLUDE_PATHS=hardware/interfaces/graphics
Bug: 35769817
Change-Id: I3747fdddb381101b9ebf51909b9686e820148aad
Otherwise the comparison is inconsistent between ReviseStashSize() and
WriteTransfers().
Bug: 35775675
Test: Successfully generate a previously failed incremental.
Change-Id: I554a51a210bf322cb5c79e28cf85607a417b094a
Stop blaming base_rules.mk, put the module makefile at the beginning
of the line. That way scripts like warn.py can parse this better.
Test: multiproduct_kati, grep logs
Change-Id: I5e6844097b44e47b8126159c1760b7524134310d
host_dalvik_java_library.mk shares less of the implementation
with the device side than I thought, port the desugar changes
from java.mk.
Test: m -j ANDROID_FORCE_JACK_ENABLED=disabled core-oj-hostdex
Change-Id: I5e09673c5eae6bf95d63540e7b626c67da87e9f1
Include the calling makefile and package name instead of just a warning
pointing to package-modules.mk
Test: multiproduct_kati, grep logs
Change-Id: I4fa35540b9695b44eea6c23463e137ec37d1d2c9
all_named_products was giving the file path instead of product name for
the name:path format of PRODUCT_MAKEFILES.
kati_all_products has been replaced with multiproduct_kati
Test: get_build_var all_named_products
Test: multiproduct_kati
Change-Id: I24015ef0778ac7cd45201aa55c1737b0553f09fe
Statically link libwinpthread.a, so that we don't have to ship a
libwinpthread dll with all of our windows exectuables.
Bug: http://b/31665213
Test: wine adb.exe
Test: wine fastboot.exe
Change-Id: I96414e980d8894f8f6a58c1c7bbd9dc5e0f5169c
Otherwise it may fail the targets that don't have
external/vboot_reference in the manifest.
Bug: 35467608
Test: lunch aosp_dragon-userdebug; m otatools-package
Test: lunch aosp_bullhead-userdebug; m otatools-package
Change-Id: I35700a4d7d0723992badcfb4a03be85d07a4b9a9
This is a partial revert to the last makeparallel change, since now we
were being killed by the signal before waiting for our child to exit. So
instead of not installing the handlers, only pass the signal along if
it's a SIGTERM.
Bug: 35214134
Test: Ensure that we're still only getting one signal for SIGINT
Test: Ctrl-C, ensure that all the soong_ui lines are before the make
error line.
Change-Id: I26fff9483a3abfd79ceb5a9ea47e3f7572d9e923
ifeq with quotes is technically correct make syntax, but rarely used.
Replace it with the simpler ifndef when comparing against empty, or
with parenthesis for consistency.
Test: builds
Change-Id: Idcbe0586c4626c67d560694596b0bd9f5f93484a
Adds the device-tests and general-tests makefiles and
added include lines to main.mk so that individual test
modules can be built properly.
These targets do not build any special tradefed wrapper
that normal *TS's usually employ. They are not necessary
here.
Bug: 35350788
Test: Updated several CTS test modules to belong to these
suites instead and ran:
`rm -rf out; make clean; make device-tests general-tests -j`
then verified the output.
Change-Id: I40e3958375f7b39cb56508f2812ce9760d403f6f
Don't print the config in Kati when we've already printed it in soong_ui
or the make wrapper.
Test: m -j
Test: USE_SOONG_UI=true m -j
Change-Id: Ic6dce0ab29d584d85c6642c597f6e9dcd6ba9dca
As part of the soong_ui effort, we noticed that we'd sometimes see
multiple SIGINTs from a single ctrl-c from a user. ctrl-c sends a SIGINT
to the entire process group, so make, makeparallel, soong_ui, and all of
its children would get a signal. Since makeparallel was passing it along
to it's child, soong_ui would get two signals.
So instead, follow what Make does and only pass along SIGTERM. Assume
that all other signals went to the entire process group.
Bug: 35214134
Test: Send SIGINT to process group, check in makeparallel's child for
more than one signal received.
Change-Id: I5b2a77ad0fcebbaa5087439948e71bf3b541061a
