When I added the --force-remote-binary flag, I needed to find a way to
get emerge to install binary packages whose use flags were not fully up
to date. To do this, I removed the --newuse flag from the call to the
emerge depgraph calculation code. This seemed safe at the time because
parallel_emerge already checks the use flags of every package it
installs. Unfortunately, things are a little more complex. Sometimes,
emerge decides to install a different version of a package because of a
use flag change. To implement this, Portage needs to know about the
--newuse flag.
To fix this bug, I've updated parallel_emerge to force in the remote
binary packages using a less intrusive method, that only impacts the
package that is being forced. This method should have less side effects
and allow us to install packages correctly.
TEST=./setup_board --board=tegra2_dev-board --force &&
./parallel_emerge --board=tegra2_dev-board -uDNvg opengles &&
./parallel_emerge --board=tegra2_dev-board -puDNvg board-devices
./parallel_emerge --board=tegra2_dev-board -uDNvg chromeos
--force-remote-binary=chromeos-chrome
BUG=chromium-os:10423
Change-Id: I2057d2992abcb08f184cfaa6c099bd6eb8c21a99
Review URL: http://codereview.chromium.org/6033001
I'm changing the boot-complete init script to start on
login-prompt-visible instead of login-prompt-ready in
http://codereview.chromium.org/5868004/.
This also adds a missing description and fixes an outdated
email address.
BUG=chromium-os:10349
TEST=ran mod_image_for_test.sh with --factory and with --factory_install and checked that resulting image contained boot-complete.conf change
Review URL: http://codereview.chromium.org/5975002
Change-Id: I709bc2b4eaaccd222b31388633522c32514206ac
We had a major issue because the latest dev channel release pre-dates the virtio change. Because of this, ctest was connecting to update that channel using virtio (which is unsupported for that). This change drops a file telling the vm lib to use e1000 as before.
Also, fixed stateful change logic.
Change-Id: I394a8cece71bdd0d55efa21ba5b6d24804432c6b
BUG=chromium-os:10434
TEST=Ran it with ctest and cros_au_test_harness on images with/without the file.
Review URL: http://codereview.chromium.org/5928005
Since the package.keywords change, it's already built as part of build_packages.
With this change we don't have to needlessly build chrome twice as part of the
pfq.
Change-Id: I8dae745496a5d6f6213663b034ffe8d279d4efd3
BUG=chromium-os:8693
TEST=Ran unittests
Review URL: http://codereview.chromium.org/5973003
This reverts the main change of
84e72fab23
and undoes the breakage of scripts that are symlinks to
...chromeos/scripts..., but at the cost of re-breaking scripts under
scripts/bin. The correct fix is a combination of removing all the
symlinks to ...chromeos/scripts... and a more robust way of finding
the chroot path of the executable. But that will have to wait until
someone else does it next week or I return from vacation
TEST=Ran all of these outside chroot. None seemed more broken than when run in the chroot. /setup_board --board=x86-generic, ./cros_workon list --all --board=x86-mario, ./start_devserver , ./build_image --help, ./build_packages --help, ./cros_upgrade_chroot --help, ./enter_chroot.sh , ./customize_rootfs --help, ./run_remote_tests.sh , ./verify_rootfs_chksum.sh , ./set_shared_user_password.sh
BUG=n0ne
Change-Id: Ic2b7484e53e6a977ed6ae675fe99e109b385020d
Review URL: http://codereview.chromium.org/5946003
Change-Id: I5366fb930ca76f596f49d0312228aa5beeccbc31
BUG=chromium-os:8693
TEST=Ran it with and without dryrun and chrome uprev.
Review URL: http://codereview.chromium.org/5846005
Change-Id: I42a6110d41e79bb0ad50817e2e76b06f7973f6c2
BUG=chromium-os:10434
TEST=Ran it with --test_prefix for delay payload test and it succeeded.
Review URL: http://codereview.chromium.org/5877007
BUG=10425
TEST=Tested before and after.
Change-Id: Ic3e66805945f379a79562338b16845bb5dbd674f
Before: Machine hangs till power-cycle.
After: Machine reboots and we get a crash dump.
Review URL: http://codereview.chromium.org/5893005
Examples of how people might be using enter_chroot:
1. ./enter_chroot [chroot_flags] VAR1=val1 VAR2=val2 -- cmd arg1 arg2
Set env vars and run cmd w/ args
2. ./enter_chroot [chroot_flags] VAR1=val1 VAR2=val2
Set env vars and run shell
3. ./enter_chroot [chroot_flags] -- cmd arg1 arg2
Run cmd w/ args
4. ./enter_chroot [chroot_flags] VAR1=val1 VAR2=val2 cmd arg1 arg2
Like #1 _if_ args aren't flags (if they are, enter_chroot will claim them)
5. ./enter_chroot [chroot_flags] cmd arg1 arg2
Like #3 _if_ args aren't flags (if they are, enter_chroot will claim them)
I also updated the help to indicate that whole-command quoting is no longer
supported. If you really need whole-command quoting (maybe you want in-
chroot redirection), you'll need to use sh -c like:
./enter_chroot.sh -- sh -c "echo \"Save me\" > /tmp/save.txt"
You should avoid single quotes in the command and arguments. This isn't
a new limitation: it's shflags related.
Change-Id: I0452a8730ac9b8197834edc753b9eece69896135
BUG=chromium-os:7072
TEST=Tried a whole bunch of these commands.
Review URL: http://codereview.chromium.org/5840003
Also add some more information output when we detect that there
is nothing to do.
Change-Id: Ia3dcf0d3c276e10d3368d6bd508f3a0d72be79cd
BUG=chromium-os:8693
TEST=Ran it with all 3 paramaters and current overlay.
Review URL: http://codereview.chromium.org/5884004
We also print to /dev/stderr directly to avoid flushing issues which have been occassionally given us broken manifests printed out
Change-Id: I97819bbc31142432aa5ef926f99f2c4f90fd50ef
BUG=chromium-os:9714
TEST=Ran cbuildbot through sync
Review URL: http://codereview.chromium.org/5888003
Change-Id: I2760e33e1473ce04f7adeda39d0db5725eaa81ae
BUG=chromium-os:7072
TEST=Replaced ./build_packages with bash and ran command; valided env var set right.
Review URL: http://codereview.chromium.org/5930002
This change will not work properly unless the chroot sudo is
updated to handle escaping properly:
http://codereview.chromium.org/5923001
...and the chroot sudo will break image_to_live unless this
change goes in. Thus, the two changes are intertwined.
Change-Id: Ib1602ee47178fc5d32c286e9a914111c1140252b
BUG=chromium-os:7072
TEST=Valided that the dev server seemed to be getting proper args.
Review URL: http://codereview.chromium.org/5929001
This change is in preparation for the new sudo, which corrects how
parameters are passed to subcommands.
Change-Id: I34d09d2716fbef137d02748a2934459a8a10d693
BUG=chromium-os:7072
TEST=Ran modified "update_kernel.sh" w/ old sudo and new sudo.
Review URL: http://codereview.chromium.org/5859002
Change-Id: I32c17e7353b420d7a895acffab4579c9b861095c
BUG=chromium-os:8693
TEST=Ran it with chrome_rev=tot, latest_release and stable_release
Review URL: http://codereview.chromium.org/5854001
This specifically gets the pfq one step closer to being completely ready to rev. We add --noprebuilts for the chrome pfq as its not ready to host prebuilts and instead we push unstable ebuilds. Anush noticed an Anush today where we are revving even when we don't change, and I've addressed this also in this CL.
Specifically:
1) Sort in cbuildbot long options by long format
2) Add --noprebuilts so that the Chrome PFQ can use this for now.
3) Fixed a unittest that had broken in a previous CL for cros_mark_chrome_as_stable.
4) Add ability for chrome to only rev if something really has changed:
-- This means, if the new ebuild either:
-- doesn't have the exact same chrome_version
-- or diff -Bu (actual diff) between last ebuild and new ebuild isn't the same (i.e. detecting a 9999 change or different CROS_SVN_COMMIT)
Change-Id: I9b289c2168d6868299573f5d7c7a676380ee497e
BUG=chromium-os:8693
TEST=Ran unittests (cbuildbot, cros_mark_*) and ran cbuildbot with new chrome pfq items.
Review URL: http://codereview.chromium.org/5783001
Change-Id: I1ac3ed12df00424ba5ba2d1c61fd6fe9716e13b9
BUG=chromium-os:9714
TEST=Ran new unittests as well as ran cbuildbot on local b/cbuild
Review URL: http://codereview.chromium.org/5689003
Change-Id: Ib9fcc078d8e0223cd1b1376db91d8e71ae8359a6
BUG=cp file to itself fails in case of nostatefuldev
TEST=ran build_image with and without statefuldev
Review URL: http://codereview.chromium.org/4959001
Patch from Trevor Bourget <tbourget@codeaurora.org>.
Change-Id: Id4ed6af96e0b621f4ad36966fad45c887b991373
BUG=None
TEST=Ran a fresh build_packages and build_image for x86-generic.
Review URL: http://codereview.chromium.org/5689004
On ARM platform read kernel from device passed in from u-boot
instead of hardcoding it to 0
BUG=none
TEST=emerged on seaboard, booted from SD card and emmc
Change-Id: Ia4506ed9f85d94eb37a9ac57430e1490d106c403
Review URL: http://codereview.chromium.org/5612008
Patch from Allen Martin <amartin@nvidia.com>.
ability to filter or manipulate content as it's passed through.
Create two example tests using this utility that cause
updates to close early to test resumes, and to add delays
to make sure the update still completes correctly.
What other tests should be created for this?
BUG=chromium-os:8207
TEST=Run by hand
Review URL: http://codereview.chromium.org/5632002
Change-Id: Iefb8c8e223fb2ba6bad2c551f7c4403a0bec6ecf
Use general purpose wait in case your connecting to a vm
that you didn't start.
Change-Id: Ie67e2d3235882d372d04e4a44d0781ddfc8d4645
BUG=chromium-os:10121
TEST=Ran it, caused some crashes. Still ran test after retries.
Review URL: http://codereview.chromium.org/5757001
This allows us to avoid deleting the board and still ensure our packages are
in sync with the full buildbot. This allows us to detect conflicts between
old and new packages.
BUG=chromium-os:10127
TEST=Run cbuildbot.py with emptytree feature forced to on and verify that
build_packages rebuilds all packages. Also run unit tests.
Change-Id: I83e186937e28eef44285724697895a4583910b58
Review URL: http://codereview.chromium.org/5652007
Adds verbose flag to print output while updates are running rather than on error.
Adds test_prefix to specify a prefix of tests you want to run
Flushes stderr more reliably for builders and debugging purposes
Dies when stateful_updates fail
Change-Id: I9e4289deec3688c52b92763eee9334a8868f948e
BUG=chromium-os:6910
TEST=Ran with test_prefix and verbose on and off
Review URL: http://codereview.chromium.org/5573007
archive_build.sh has code to archive host prebuilts, but the code isn't working because of a typo in archive_build.sh.
The typo: $FLAGS_board is misspelled as $FLAGS_BOARD.
BUG=chromium-os:10148
TEST=Ran prebuilt.py with --sync-host option
Change-Id: I50129b1fdfb931b1c076a064b970d7d943f556ac
Review URL: http://codereview.chromium.org/5724001
into mod_image_for_test.sh rather than slightly different versions
of the same in image_to_usb.sh and image_to_vm.sh
Added a function to get a test image into common.sh
Added --inplace option to mod_image_for_test, which is the default,
and preserves the original behaviour. But using --noinplace it will
now do the copy for you.
Found that chromiumos_image.bin appears throughout the scripts, so added it and the test variant to common.sh
BUG=chromiumos-10126
TEST=run mod_image_for_test.sh with and without --noinplace
run image_to_usb.sh and image_to_vm.sh with both options
test on Seaboard that correct image is provided
Really we should have automated testing for these scripts
Change-Id: I5cfa91792c7fded35e7f4ca8f8f27c6b270817fb
Review URL: http://codereview.chromium.org/5271010
Change-Id: I0dda2cfb6e11cb69a6a1a6a4bae478cab78b8947
BUG=chromium-os:8693
TEST=Ran it with cros_mark_chrome with stable_release
TBR=Config change to get chrome builder up.
Change-Id: I1d0e546f6ea837928ba426e4690007ef2649f68e
BUG=10102
TEST=test KVM with -net virtio and ensure it boots and can copy files in etc
Review URL: http://codereview.chromium.org/5543006
This server requires that unsafe characters are quoted in URLs, so we do that
here.
Change-Id: I6b7c430671b02a8da11777029b0fdb59a87dcae1
BUG=chromium-os:9902
TEST=Ran unit tests. Ran example buildbot run and diff'd new version and old
version of Packages file.
Review URL: http://codereview.chromium.org/5542002
I started to use /tmp, but it doesn't exist both inside and outside of the chroot. Using a consistent name like this has the advantage that the files will be cached and not redownloaded every time. Well, except on the build servers that wipe the related directories between builds.
Change-Id: I58b42d9f1bd18cc4eb44043ed2436d4ffa482767
BUG=chromium-os:10012
TEST=
Review URL: http://codereview.chromium.org/5597004
