TBR=Needed to fix preflight build.
BUG=chromium-os:5311
TEST=Remove SHA1 from one of the packages in the Packages file, make sure
that prebuilt.py treats that as a duplicated package. Also run new
unit tests. Also ran buildbot.py
Change-Id: Ie35a7c818fd659bcc2296605ace850aee2a88448
Review URL: http://codereview.chromium.org/5370005
urlparse.urljoin('scheme:foo/bar/', 'baz') == 'baz', so it doesn't work the
way I expected. Instead, let's use regular string joins.
BUG=chromium-os:5311
TEST=Ran buildbot in test mode.
Change-Id: Iea4205573db602e0cc20af24a77d255d17e79221
Review URL: http://codereview.chromium.org/5357012
The --fast support for build_packages is stable now, and should significantly
improve the speed of our official builders. We should turn it on so that
builders can finish their builds in less than 8 hours.
BUG=chromium-os:6706
TEST=Run ./enter_chroot.sh --official_build and verify that --fast is on by
default now
Change-Id: I6ad126b9b6ce16ffc9887a7af22c2e3f85afbf42
Review URL: http://codereview.chromium.org/3418001
Preflight prebuilts reference the last full build, so they should always be complete as long as the last full build is still there.
Also add host prebuilts to the preflight prebuilts.
BUG=chromium-os:5311
TEST=Run unit tests for cbuildbot and prebuilt.py. Test runs of cbuildbot.py with --dryrun.
Review URL: http://codereview.chromium.org/5344002
Change-Id: Id95f94c02cc2f6cbd70a029d4f8b94617b7cc071
This change updates cbuildbot.py to upload prebuilts and reference them via PREFLIGHT_BINHOST.
Also ensure make.conf ends in a newline, as it looks nicer.
BUG=chromium-os:5311
TEST=Test that prebuilt.py updates PREFLIGHT_BINHOST in right make.conf
file. Test that cbuildbot.py runs prebuilt.py with right
arguments. Run unit tests for cbuildbot and prebuilt.py. Test runs of cbuildbot.py with --dryrun. Test upload of Packages file to local web server and check that emerge works with them.
Change-Id: Iad03c6c469e05b9ee1cceff69cbe2bdd51225e25
Review URL: http://codereview.chromium.org/4969003
The default DB is an all-pass DB. If no DB found, the hardware_Components test
falls back to the all-pass one and skip the real check. In order to prevent
skipping the check, remove the default DB in the boards already have DB
directories.
TEST=run "image_to_usb.sh --factory" and check the default DB removed
BUG=chrome-os-partner:1737
Change-Id: Ia25f2551a5eba3a95fe736e86bfe52c340f329a8
Review URL: http://codereview.chromium.org/5339011
BUG=9717
TEST=run cros_workon_make --scrub and ensure that it only scrubs if the user types y or yes (case insensitive)
Change-Id: Ic1f8020fcca8bd57924391fcbf075d3e2adc5e4a
Review URL: http://codereview.chromium.org/5265009
What I did:
1. enter_chroot once and only once if necessary. The previous version would enter/exit chroot 3 times in worst case. And the entire logic happens inside chroot only, this has greatly simplified the overall workflow, and reduced the number of variables used. Shell script variables could be tricky.
2. Change variable type to test_type since type is a bash builtin.
3. get rid of {$TMP}/run_test.sh script, since it is not necessary any more.
4. get rid of rsync step from third_party/autotest/files to ${BUILD_DIR}, since it is not neccessary either.
overall, reduced ~40 lines of code.
All the change should be transparent to end users and there should be no regression changes at all.
w/wo emerge autotest
w/wo cros_workon
in/outside of chroot.
and all its combinations.
Change-Id: I9c1532e9cb6cc0e724d4b6d870723df3e2a147ec
BUG=9291
TEST=Run storageFio test since it need prebuild test and deps.
Review URL: http://codereview.chromium.org/5176009
It's like RunCommand, but you can run the command, capture it's error code and output together.
This is to support tests that need to capture output of image_to_live, and to know if it succeeded.
Change-Id: If674b6d79697c0f0b5c96be9fc83adbed9b9893e
BUG=chromium-os:9502
TEST=Ran by hand
Review URL: http://codereview.chromium.org/5339006
cros_workon_make provides a make-style interface to incrementally building
cros_workon-enabled packages, appropriately setting up the toolchain for
the target board beforehand.
To incrementally build your package:
cros_workon_make --board <board> <package>
To re-run your package's configure steps and compile:
cros_workon_make --board <board> <package> --reconf
To incrementally build your package and tests, and run the tests:
cros_workon_make --board <board> <package> --tests
To incrementally build and install your package:
cros_workon_make --board <board> <package> --install
To remove ALL files in your package repo not known by git:
cros_workon_make --board <board> <package> --scrub
With the exception of --scrub, these flags all compose
BUG=6843
TEST=build workon packages in place
Change-Id: Ie219074127549e1e29adad6d6a03142ab74e0172
Review URL: http://codereview.chromium.org/5244001
This also removes the rw_checkout var as it's not longer changes the affect of whether we use the ssh url or not.
Note that this replaces the other CL from last night because the other CL didn't work. The other CL still pulled in the same manifest that has http: url's for each git repo. I reverted that change last night after I noticed this.
Change-Id: I3d4ad2be6887ac2cf4ed2009bad9cae6dfdf5bbf
BUG=chromium-os:9509
TEST=Ran with cbuildbot --clobber and incremental
Review URL: http://codereview.chromium.org/5278010
Other things:
Removing commit|clean|push command from cros_mark_chrome because it's unnecessary ... I'm just gonna use the same stabilizing branch as the other stable marker, so I can use the same clean / push from the other script. This makes this change fit in much more nicely to cbuildbot.
Other changes:
Add ability for cros_mark_chrome_as_stable to communicate to calling script through stdout.
Make STABLE_BRANCH in cros_mark_as_stable public for above reason.
Removed push_options from cros_mark and accompanying change to cbuildbot as it isn't used anymore.
Made it easier to debug with cbuildbot (I ran cbuildbot A LOT of times to test this change so I felt the pain)...this includes:
- only build when syncing
- allow one to explicitly disable tests
- use dryrun is options.debug is set
Change-Id: I413a2e81a99cdde2e4d9139561cd518245b9b346
BUG=chromium-os:8693
TEST=Ran cbuildbot with --notest --debug --nosync x86-pre-flight and
saw it go through correctly and dryrun push. Ran cbuildbot with same and --chrome_rev=tot. Ran with syncing and running tests too.
Review URL: http://codereview.chromium.org/5154008
This is a complement to the proposed fix for http://crosbug.com/9447, and I am getting it out first since it is a relatively simple change.
This change ensures that the rootfs on the image output by cros_make_image_bootable wasn't corrupted accidentally (for example, no boot.desc was provided and the defaults were incorrect for the image). This should catch rootfs errors that we end up discovering at delta AU time.
BUG=chromium-os:9578
TEST= Ran cros_make_image_bootable manually on an existing image with and without boot.desc. As expected, without boot.desc the output image rootfs had errors.
Also ran ./build_image - which worked fine too.
Change-Id: I4ed3a56634f37ab7d5ff2dc052ad607740356a40
Review URL: http://codereview.chromium.org/5216003
- Conditionally patch chrontel.conf
- Add a patch for nvrm to start on factory, so that X can start.
BUG=crosbug.com/p/1721
TEST=Built and run factory image for tegra2_seaboard
Change-Id: Ied87b6d959f940ef8e99e124ab419fdeea48aa26
Review URL: http://codereview.chromium.org/5126005
Change-Id: I680e28bcafbfeb4f19e2404e2b28761e26111f7f
BUG=chromium-os:9509
TEST=Ran cbuildbot on my local builder checkout with both clobber and without.
It sync'd and built.
Review URL: http://codereview.chromium.org/5312002
added --test option
removed superfluous output (e.g. from dd) unless --verbose given
added --debug option for when you really want that output
added use of pv for a nice progress bar
TEST=(run script several times)
BUG=chromium-os:9563
Change-Id: Ib9d89ad5649b6cb6d5db35f66f6a3fda1ada0c53
Review URL: http://codereview.chromium.org/5176002
Change-Id: I791fa01fcd2feb0c7b4c7e3d63cebc88d896c8ac
BUG=chromium-os:9459
BUG=chromium-os:9460
TEST=run the script, and check results with showbootdata or by manually inspecting the output directory
TEST=run "bootperf invalid-hostname", and see that the error message is useful
Review URL: http://codereview.chromium.org/5156008
Change-Id: I6744aac09fb075f91f710f16343fab45b8dc8af9
BUG=chromium-os:7283
TEST=Ran it with cbuildbot through the unit tests phase and looked
at the file and saw it get parsed correctly.
Review URL: http://codereview.chromium.org/5124006
This will also make it so that AU's unit tests run on buildbot.
BUG=9435
TEST=new chroot, build_packages, ./cros_run_unit_tests
Change-Id: I4dbbcf1db4c1b1f3634202261b08b8f10a9a0ac6
Review URL: http://codereview.chromium.org/5145008
Have subsequent review for cbuildbot that calls this for the pfq.
This module has 3 modes ... TOT, latest_release, sticky_release that describe the three different types of revs for chrome we will be supporting.
For tot, we grab the latest svn revision # from the chrome src tree, and rev to the corresponding ebuild i.e. 9.0.553.0_alpha-r1 with CROS_SVN_COMMIT=svn revision number.
For latest_release, we grab the latest release from the releases page, and create an ebuild (or rev a previously existing on with same version) and use the release candidate suffix _rc.
For sticky_release, we use much of the same logic as latest_release, however, we specifically look to rev _rc's in the sticky branch.
This change depends on http://codereview.chromium.org/5172003 for the cros_mark_as_stable changes.
Change-Id: Idc1f4cd905cc1dcb4c44422f6104760077a24b8d
BUG=chromium-os:8693
TEST=Ran cros_mark_as_stable with --packages="chromeos-base/chromeos-chrome" --chrome_version=0.8.65.0 --chrome_revision=67890 commit. Re-ran unit-tests as well (though thinking about adding another unit test for chrome revving)
Review URL: http://codereview.chromium.org/4798001
Change-Id: Ieaa2553612aefd6161785185350d495cbf0dad81
BUG=9291
TEST=Run storageFio test since it need prebuild test and deps.
w/wo emerge autotest
w/wo cros_workon
in/outside of chroot.
w/wo --used_emerged
w/wo --results_dir_root
and all its combinations.
Review URL: http://codereview.chromium.org/5212003
Besides changing a lot of _ to not _, I added lots of docstrings, moved things around,
factored out the Marking component of the EBuildStableMarker (for use with other CL),
and fixed the version logic.
The versioning logic was too hacktastic. So I took a note from David's changes to
take the information from pkgsplit rather than doing the ugly parsing I was doing before.
This is much more robust in the case of no _r*'s and _rc or _alpha*, etc.
This CL also adds the ability to test Push changes in 2 ways. One using --dryrun which allows someone to call it by not actually push to the remote server and two, a unit test for the Push method.
Used in http://codereview.chromium.org/4798001/
BUG=chromiumos:8693
TEST=Ran unit tests and chrome mark as stable. Will run some more tests.
Committed: http://chrome-svn/viewvc/chromeos?view=rev&revision=be485ce
Review URL: http://codereview.chromium.org/5172003
Change-Id: I6d1231c6f46d8cc2e5fb57d04f2d3417cfbfb4f8
Besides changing a lot of _ to not _, I added lots of docstrings, moved things around,
factored out the Marking component of the EBuildStableMarker (for use with other CL),
and fixed the version logic.
The versioning logic was too hacktastic. So I took a note from David's changes to
take the information from pkgsplit rather than doing the ugly parsing I was doing before.
This is much more robust in the case of no _r*'s and _rc or _alpha*, etc.
This CL also adds the ability to test Push changes in 2 ways. One using --dryrun which allows someone to call it by not actually push to the remote server and two, a unit test for the Push method.
Used in http://codereview.chromium.org/4798001/
BUG=chromiumos:8693
TEST=Ran unit tests and chrome mark as stable. Will run some more tests.
Review URL: http://codereview.chromium.org/5172003
Change-Id: I4319dcd3a4235474ecfd61c680a5242bab5bd00b
This would not have been a problem if "--no_graphics" is used.
Change-Id: I3645391d25d4bb8618109ff5be0ca2fbe8e96a60
BUG=chromium-os:9373
TEST=Ran by hand
Review URL: http://codereview.chromium.org/5104006
Besides changing a lot of _ to not _, I added lots of docstrings, moved things around,
factored out the Marking component of the EBuildStableMarker (for use with other CL),
and fixed the version logic.
The versioning logic was too hacktastic. So I took a note from David's changes to
take the information from pkgsplit rather than doing the ugly parsing I was doing before.
This is much more robust in the case of no _r*'s and _rc or _alpha*, etc.
Used in http://codereview.chromium.org/4798001/
BUG=chromiumos:8693
TEST=Ran unit tests and chrome mark as stable. Will run some more tests.
Review URL: http://codereview.chromium.org/5172003
Change-Id: Id6ef9a480591d1a28352f360c6d076564b43e318
Change-Id: I1032932b696929fccf97f911c555ec5a24f589d4
BUG=9291
TEST=Run storageFio test since it need prebuild test and deps.
w/wo emerge autotest
w/wo cros_workon
in/outside of chroot.
and all its combinations.
Review URL: http://codereview.chromium.org/5188002
The update_bootloaders needs to unmount file system first for syslinux to
manipulate the raw device file, but the real "umount+losetup -d" in
cleanup stage would cause failure.
This CL remounts partition again so that cleanup can be executed
successfully.
BUG=chromium-os:9278
TEST=Manually executed ./build_image;
Before fix: seeing "umount: /tmp/esp.2q8fSn: not mounted" in last page.
After fix: no such error.
The output image is verified to be bootable successfully.
Change-Id: I19e2e062ad814e7bcd54777b5c3cee31f2b92e81
Review URL: http://codereview.chromium.org/5165001
Major changes are:
- prefix functions in cros_image_common with 'image' to avoid naming conflicts
- use awk to replace grep+sed+cut
- use case to handle board name list
- factory_setup.sh should return error if applying patch failed
- refine for shell script coding style guide:
* replace `` with $()
* quote variables
* replace "! -z" by "-n"
* no space between redirection symbol (>) and target name (eg, >>filename)
* 1> should be simply >
* no trailing \ for && and ||
BUG=chrome-os-partner:1583
TEST=manually verified:
./image_to_usb.sh --factory # calls mod_image_for_test --factory, and works fine
./make_factory_package.sh --factory PATH_TO_IMAGE --release PATH_TO_IMAGE ...
# factory payloads were created successfully
Change-Id: I6bb10bdfb12cbdb14e9816b3ad72dfe4b7b0472f
Review URL: http://codereview.chromium.org/5168001
