|
|
Created:
9 years, 10 months ago by zbehan Modified:
9 years, 7 months ago CC:
chromium-os-reviews_chromium.org, Mandeep Singh Baines, anush, sosa Visibility:
Public. |
Descriptioncros_generate_update_payload: kill old style update image generation
This disables an unwanted dependency.
Change-Id: I02752fa38cf5f64af366f1831a351b1b903563d0
BUG=n0ne
TEST=
Committed: http://chrome-svn/viewvc/chromeos?view=rev&revision=fad7fec
Patch Set 1 #
Messages
Total messages: 10 (0 generated)
LGTM On Wed, Feb 2, 2011 at 4:24 PM, <zbehan@chromium.org> wrote: > Reviewers: adlr, Nick Sanders, anush, > > Description: > cros_generate_update_payload: kill old style update image generation > > This disables an unwanted dependency. > > Change-Id: I02752fa38cf5f64af366f1831a351b1b903563d0 > > BUG=n0ne > TEST= > > Please review this at http://codereview.chromium.org/6349060/ > > SVN Base: ssh://git@gitrw.chromium.org:9222/crosutils.git@master > > Affected files: > M cros_generate_update_payload > > > Index: cros_generate_update_payload > diff --git a/cros_generate_update_payload b/cros_generate_update_payload > index > b55f438526109e2db548938d8d5b588a5ede0ec9..586b2f8f2396640e54f93e7755e3407b848c38ab > 100755 > --- a/cros_generate_update_payload > +++ b/cros_generate_update_payload > @@ -247,6 +247,11 @@ if [ "$DELTA" -eq "$FLAGS_TRUE" -o "$FLAGS_old_style" > -eq "$FLAGS_FALSE" ]; then > else > echo "Generating old-style full update" > > + # TODO(zbehan): Remove this branch entirely and disable the cmdline > option. > + echo "This is a deprecated and unused behaviour. It also introduces > unwanted" > + echo "dependency to a third script (mk_memento_images) which has become > broken." > + exit 1 > + > trap cleanup INT TERM EXIT > DST_KERNEL=$(extract_partition_to_temp_file "$FLAGS_image" 2) > if [ "$FLAGS_patch_kernel" -eq "$FLAGS_TRUE" ]; then > @@ -254,7 +259,7 @@ else > fi > DST_ROOT=$(extract_partition_to_temp_file "$FLAGS_image" 3) > > - GENERATOR="${SCRIPTS_DIR}/mk_memento_images.sh" > +# GENERATOR="${SCRIPTS_DIR}/mk_memento_images.sh" > > CROS_GENERATE_UPDATE_PAYLOAD_CALLED=1 "$GENERATOR" "$DST_KERNEL" > "$DST_ROOT" > mv "$(dirname "$DST_KERNEL")/update.gz" "$FLAGS_output" > > >
I think the network test lab still uses this flow. Checkout bin/cros_image_to_target.py.
Yes we depend on it heavily. On Feb 2, 2011 9:38 PM, <petkov@chromium.org> wrote: > I think the network test lab still uses this flow. Checkout > bin/cros_image_to_target.py. > > > http://codereview.chromium.org/6349060/
OK. That does not make me happy, but I'll revert this. On Thu, Feb 3, 2011 at 4:34 PM, Sam Leffler <sleffler@google.com> wrote: > Yes we depend on it heavily. > On Feb 2, 2011 9:38 PM, <petkov@chromium.org> wrote: > > I think the network test lab still uses this flow. Checkout > > bin/cros_image_to_target.py. > > > > > > http://codereview.chromium.org/6349060/ >
We're happy to switch over but every time we tried we failed. If someone is willing to help we're happy to move but w/o something operational all our testing is stopped. -Sam On Thu, Feb 3, 2011 at 9:13 AM, Zdenek Behan <zbehan@chromium.org> wrote: > OK. That does not make me happy, but I'll revert this. > > On Thu, Feb 3, 2011 at 4:34 PM, Sam Leffler <sleffler@google.com> wrote: >> >> Yes we depend on it heavily. >> >> On Feb 2, 2011 9:38 PM, <petkov@chromium.org> wrote: >> > I think the network test lab still uses this flow. Checkout >> > bin/cros_image_to_target.py. >> > >> > >> > http://codereview.chromium.org/6349060/ > >
cros_image_to_target.py should be switched to use new style payloads. This should be relatively easy with the exception that cros_image_to_target needs to run outside chroot while cros_generate_update_payload needs to run inside chroot. On 2011/02/03 17:40:16, sleffler wrote: > We're happy to switch over but every time we tried we failed. If > someone is willing to help we're happy to move but w/o something > operational all our testing is stopped. > > -Sam > > On Thu, Feb 3, 2011 at 9:13 AM, Zdenek Behan <mailto:zbehan@chromium.org> wrote: > > OK. That does not make me happy, but I'll revert this. > > > > On Thu, Feb 3, 2011 at 4:34 PM, Sam Leffler <mailto:sleffler@google.com> wrote: > >> > >> Yes we depend on it heavily. > >> > >> On Feb 2, 2011 9:38 PM, <mailto:petkov@chromium.org> wrote: > >> > I think the network test lab still uses this flow. Checkout > >> > bin/cros_image_to_target.py. > >> > > >> > > >> > http://codereview.chromium.org/6349060/ > > > >
I'm really curious why does the script have to be ran outside the chroot. This would be yet another complication for the effort of moving away from monolithic src/scripts. Nonetheless, i sent out a CL to revert this. I'll see first if there can be anything done to make it not break. On Thu, Feb 3, 2011 at 6:54 PM, <petkov@chromium.org> wrote: > cros_image_to_target.py should be switched to use new style payloads. This > should be relatively easy with the exception that cros_image_to_target > needs to > run outside chroot while cros_generate_update_payload needs to run inside > chroot. > > > On 2011/02/03 17:40:16, sleffler wrote: > >> We're happy to switch over but every time we tried we failed. If >> someone is willing to help we're happy to move but w/o something >> operational all our testing is stopped. >> > > -Sam >> > > On Thu, Feb 3, 2011 at 9:13 AM, Zdenek Behan <mailto:zbehan@chromium.org> >> > wrote: > >> > OK. That does not make me happy, but I'll revert this. >> > >> > On Thu, Feb 3, 2011 at 4:34 PM, Sam Leffler <mailto:sleffler@google.com >> > >> > wrote: > >> >> >> >> Yes we depend on it heavily. >> >> >> >> On Feb 2, 2011 9:38 PM, <mailto:petkov@chromium.org> wrote: >> >> > I think the network test lab still uses this flow. Checkout >> >> > bin/cros_image_to_target.py. >> >> > >> >> > >> >> > http://codereview.chromium.org/6349060/ >> > >> > >> > > > > http://codereview.chromium.org/6349060/ >
This has broken the Autotest pipeline. I assume the revert is pending the opening of the tree? If there's a new way we should transition to instead, let me know. On 2011/02/03 19:34:45, zbehan wrote: > I'm really curious why does the script have to be ran outside the chroot. > This would be yet another complication for the effort of moving away from > monolithic src/scripts. > > Nonetheless, i sent out a CL to revert this. I'll see first if there can be > anything done to make it not break. > > On Thu, Feb 3, 2011 at 6:54 PM, <mailto:petkov@chromium.org> wrote: > > > cros_image_to_target.py should be switched to use new style payloads. This > > should be relatively easy with the exception that cros_image_to_target > > needs to > > run outside chroot while cros_generate_update_payload needs to run inside > > chroot. > > > > > > On 2011/02/03 17:40:16, sleffler wrote: > > > >> We're happy to switch over but every time we tried we failed. If > >> someone is willing to help we're happy to move but w/o something > >> operational all our testing is stopped. > >> > > > > -Sam > >> > > > > On Thu, Feb 3, 2011 at 9:13 AM, Zdenek Behan <mailto:zbehan@chromium.org> > >> > > wrote: > > > >> > OK. That does not make me happy, but I'll revert this. > >> > > >> > On Thu, Feb 3, 2011 at 4:34 PM, Sam Leffler <mailto:sleffler@google.com > >> > > >> > > wrote: > > > >> >> > >> >> Yes we depend on it heavily. > >> >> > >> >> On Feb 2, 2011 9:38 PM, <mailto:petkov@chromium.org> wrote: > >> >> > I think the network test lab still uses this flow. Checkout > >> >> > bin/cros_image_to_target.py. > >> >> > > >> >> > > >> >> > http://codereview.chromium.org/6349060/ > >> > > >> > > >> > > > > > > > > http://codereview.chromium.org/6349060/ > >
Reverted. On Thu, Feb 3, 2011 at 11:30 PM, <dalecurtis@chromium.org> wrote: > This has broken the Autotest pipeline. I assume the revert is pending the > opening of the tree? If there's a new way we should transition to instead, > let > me know. > > > On 2011/02/03 19:34:45, zbehan wrote: > >> I'm really curious why does the script have to be ran outside the chroot. >> This would be yet another complication for the effort of moving away from >> monolithic src/scripts. >> > > Nonetheless, i sent out a CL to revert this. I'll see first if there can >> be >> anything done to make it not break. >> > > On Thu, Feb 3, 2011 at 6:54 PM, <mailto:petkov@chromium.org> wrote: >> > > > cros_image_to_target.py should be switched to use new style payloads. >> This >> > should be relatively easy with the exception that cros_image_to_target >> > needs to >> > run outside chroot while cros_generate_update_payload needs to run >> inside >> > chroot. >> > >> > >> > On 2011/02/03 17:40:16, sleffler wrote: >> > >> >> We're happy to switch over but every time we tried we failed. If >> >> someone is willing to help we're happy to move but w/o something >> >> operational all our testing is stopped. >> >> >> > >> > -Sam >> >> >> > >> > On Thu, Feb 3, 2011 at 9:13 AM, Zdenek Behan <mailto: >> zbehan@chromium.org> >> >> >> > wrote: >> > >> >> > OK. That does not make me happy, but I'll revert this. >> >> > >> >> > On Thu, Feb 3, 2011 at 4:34 PM, Sam Leffler <mailto: >> sleffler@google.com >> >> > >> >> >> > wrote: >> > >> >> >> >> >> >> Yes we depend on it heavily. >> >> >> >> >> >> On Feb 2, 2011 9:38 PM, <mailto:petkov@chromium.org> wrote: >> >> >> > I think the network test lab still uses this flow. Checkout >> >> >> > bin/cros_image_to_target.py. >> >> >> > >> >> >> > >> >> >> > http://codereview.chromium.org/6349060/ >> >> > >> >> > >> >> >> > >> > >> > >> > http://codereview.chromium.org/6349060/ >> > >> > > > > http://codereview.chromium.org/6349060/ > |