Previously jobs were running on a large multi-core server where 10 jobs
as default made sense (or may even have been on the low side).
Using hosted runners with fixed (and lower) numbers of vCPUs we should
adapt compilation to match the number of cpus we have dynamically.
This is cross-platform compatible with macos and linux only.
To remove multiple occurances of the respository name, against which we
compare `${{ github.repository }}` to check if we should use Cirrus
Runners, introduce a helper job which can check a single environment
variable and output this as an input to subsequent jobs.
Forks can maintain a trivial patch of their repo name against the
`REPO_USE_CIRRUS_RUNNERS` variable in ci.yml if they have Cirrus Runners
of their own, which will then enable cache actions and docker build
cache to use Cirrus Cache.
It's not possible to use `${{ env.USE_CIRRUS_RUNNERS }}` in the
`runs-on:` directive as the context is not supported by GitHub.
If it was, this job would no longer be necessary.
Whilst the action cirruslabs/actions/cache will automatically set this
host, the docker `gha` build cache backend will not be aware of it.
Set the value here, which will later be used in the docker build args to
enable docker build cache on the cirrus cache.
Another action to reduce boilerplate in the main ci.yml file.
This action will set up a docker builder compatible with caching build
layers to a container registry using the `gha` build driver.
It will then configure the docker build cache args.
If set, Cirrus runners will be used on pushes to, and pull requests
against, this repository.
Forks can set this if they have their own cirrus runners.
Add "Restore" and "Save" caching actions.
These actions reduce boilerplate in the main ci.yml configuration file.
These actions are implemented so that caches will be saved on `push`
only.
When a pull request is opened it will cache hit on the caches from the
lastest push, or in the case of depends will hit on any matching depends
hash, falling back to partial matches.
Depends caches are hashed using
`$(git ls-tree HEAD depends "ci/test/$FILE_ENV" | sha256sum | cut -d' ' -f1)`
and this hash is passed in as an input to the actions. This means we
direct cache hit in cases where depends would not be re-built, otherwise
falling back to a partial match.
Previous releases cache is hashed similarly to depends, but using the
test/get_previous_releases.py file.
The cirruslabs cache action will fallback transparently to GitHub's
cache in the case that the job is not being run on a Cirrus Runner,
making these compatible with running on forks (on free GH hardware).
941b8f54c0 ci: run get_previous_releases as part of test cross win job (Max Edwards)
5e2182140b test: increment mocked time for migrating wallet backups (Max Edwards)
5174565802 ci: disable feature_unsupported_utxo_db functional test (Max Edwards)
3dc90d69a6 test: remove mempool.dat before copying (Max Edwards)
67a6b20d50 test: add windows support to get previous releases script (Max Edwards)
1a1b478ca3 scripted-diff: rename tarball to archive (Max Edwards)
4f06dc8484 test: remove building from source from get prev releases script (Max Edwards)
Pull request description:
This PR updates the `test/get_previous_releases.py` script to also work on Windows by changing to be pure python rather than using unix tools such as `curl` and `tar`.
This enables additional functional tests to run such as `wallet_migration.py`, `mempool_compatability.py` and `wallet_backwards_compatibility.py`.
Unfortunately `feature_unsupported_utxo_db.py` _could_ run but this test requires Bitcoin `v0.14.3` which will not run under windows with emojis in the data directory (as the functional test runner has by default) . This test could be run as it's own step in the ci workflow file and would pass but as it's quite an old version / feature I have assumed it's not worth worrying about and best just to exclude.
Two tests needed to be slightly modified to run under windows. Both were issues with trying to overwrite a file that already exists which windows seems to be more strict on than the unix based systems.
Finally, building from source has been dropped from the `get_previous_releases.py` script. This had not been updated after the move to cmake and so it was assumed that nobody could have been using that feature.
ACKs for top commit:
maflcko:
re-ACK 941b8f54c0🍪
achow101:
ACK 941b8f54c0
hodlinator:
re-ACK 941b8f54c0
Tree-SHA512: 22933d0ec278b9b0ffcd2a8e90026e1a3631b00186e7f78bd65be925049021e319367d488c36a82ab526a07b264bac18c2777f87ca1174b231ed49fed56d11cb
This test required v0.14.3 which doesn't support special characters like
emojis in the datapath on windows. Functional test runner includes a few
emojis in it's default data directory.
* Run git config earlier and only once
* Run git merge in the yaml, before calling the bash script
* Run git reset in the yaml as well, for symmetry
* Replace "git merge --abort" with "git reset --hard", because it does
not fail when already up to date and no merge was started.
faa807bdf8 ci: Merge master in test-each-commit task (MarcoFalke)
Pull request description:
The `test-each-commit` task will often fail, when the CI config yaml is updated along with code changes.
This is because, GitHub seems to be merging the CI config on a fresh pull with the current target branch (`master`). However, the code changes are not.
A tedious workaround would be for every developer to rebase on every intermittent (https://github.com/bitcoin/bitcoin/issues/31946#issuecomment-2740911853) and non-intermittent CI issue.
However, fix this instead by merging with `master`.
ACKs for top commit:
laanwj:
ACK faa807bdf8
hebasto:
ACK faa807bdf8.
Tree-SHA512: 4849bd558dc6cdc7d86b95164ccee32ab7c08c9b7d31cf8ec5c8e9a2251fc819630f8fa9b929ed39e8e033c67bb006f0beb33e0de216e1224680be88c5fa0161
