Add docker-based update system #52
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
We get a Dockerfile and associated .dockerignore to create an Ubuntu 16.04 LTS container that can be used for generating updates. A Makefile is used with a few targets to make invocation of docker related tasks easier. The README is updated to reflect the docker implementation. Several scripts are updated to: - support `set -x` debugging when the DEBUG env var is set - be quieter when running The .gitignore gets some changes to ignore artefacts generated during the docker-update process. A new update-wrapper.sh script is used as the docker entry point. The get-release-image.py script will now only validate if it's downloaded, and will exit 1 if the validation failed. The install-su.sh script forces the copy of keys, which would otherwise break the `set -e` mode. The strip-cpio tool is inherently noisy. The easy workaround here is to simply send all stderr output from it to /dev/null. Add usage output to the update.sh script, along with a trap handler to inform the user of an error if the script unexpectedly exits. There's also a new `--scheduled` argument to change the behaviour when the new signed image is ready: without `--scheduled` the behaviour is the same as before; with the flag set instructions are provided for what to do next. The script will also clone the helper repos if they're not already in place.
With the make_keys.sh script in the keys folder it becomes harder to manage our keys in the folder. With only keys in that folder we can handle those secrets as we please, for example as a git repository with blackbox GPG encryption.
Contributor
Author
|
The changes here also benefit from the changes in mikeperry-tor/super-bootimg#1 |
The conditionals requred the DEBUG var to always be set.
It's no longer available
A series of new vdex files are in the latest images. Missing out these files during the installation of gapps is causing the builds to fail, so we need to have these in both the gapps-removelist and the docker-sudoers config.
When generating signed updates we set the file name to match the factory name so it's easier to be clear about which update you're applying. We also generate a hash of the file for validation, especially useful if we're moving the update around and want to validate it's integrity before using the new update image.
Contributor
Author
|
Have made some changes to accommodate the 2017-09 CopperheadOS images that contain some new vdex files that were causing problems for applying gapps-delta. Older commits that are added here have had commit author corrected and signed. |
CopperheadOS have changed the URLs from which the updates are retrieved. The new logic implemented here reflects what the javascript on their site is doing.
CopperheadOS have chaned the update mechanism. The new method is now used by the get-release-image.py script. Additionally created images were not being ignored.
The previous update-binary was not successfully getting the properties,
and therefore wasn't detecting that the device was an angler device.
This binary has been taken from here:
https://legacy.copperhead.co/angler-ota_update-2017.09.29.01.zip
The end of the updater-script is causing some odd output when applied,
so following the example from the stock one the last command is a
set_progress(), and the bang (!) is taken out of the ui_print() in case
that was causing any issues too.
Contributor
Author
|
Latest updates handle changes to how CopperheadOS make update metadata available, and fix regressions in angler updates. I don't have any other device to test with, but it's likely that the update-binary for other devices will need to be updated to successfully apply updates of android 8. |
When the metadata can't be retrieved we should exit with an error and prevent the rest of the process continuing.
We don't need intermediates to be kept and they use up space.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
We get a Dockerfile and associated .dockerignore to create an Ubuntu
16.04 LTS container that can be used for generating updates. A Makefile
is used with a few targets to make invocation of docker related tasks
easier.
The README is updated to reflect the docker implementation.
Several scripts are updated to:
set -xdebugging when the DEBUG env var is setThe .gitignore gets some changes to ignore artefacts generated during
the docker-update process.
A new update-wrapper.sh script is used as the docker entry point.
The get-release-image.py script will now only validate if it's downloaded,
and will exit 1 if the validation failed.
The install-su.sh script forces the copy of keys, which would otherwise
break the
set -emode.The strip-cpio tool is inherently noisy. The easy workaround here is to
simply send all stderr output from it to /dev/null.
Add usage output to the update.sh script, along with a trap handler to
inform the user of an error if the script unexpectedly exits. There's
also a new
--scheduledargument to change the behaviour when thenew signed image is ready: without
--scheduledthe behaviour is thesame as before; with the flag set instructions are provided for what to
do next. The script will also clone the helper repos if they're not
already in place.
The make_keys.sh script is also moved as it makes it hard to
manage our keys in the folder. With only keys in that folder we can
handle those secrets as we please, for example as a git repository with
blackbox GPG encryption.