As the repo publishes per-SDK/component release tags i.e.
`sdk/typescript/v0.21.9`, `helm/chart/v0.21.9`), we need to specify
`extractVersion` to extract the version as we need.
Co-Authored-By: Claude Sonnet 5 <jamie.tanna+claude-code@mend.io>
As Maven tags releases as maven-X.Y.Z rather than plain SemVer, we need to set a custom extractVersion.
As ranges can be used, we can treat it as Node-package-registry-style versioning.
Co-Authored-By: Claude Sonnet 5 <jamie.tanna+claude-code@mend.io>
As the Action only uses the last `toolchain` as the active one, we can
only update that value.
Co-Authored-By: Claude Sonnet 5 <jamie.tanna+claude-code@mend.io>
setup-beam is a bit more complex, as it has multiple inputs, and
`erlang/otp` releases as `OTP-x.y.z`, so we need to specify a custom
`extractVersion`.
Co-Authored-By: Claude Sonnet 5 <jamie.tanna+claude-code@mend.io>
* test(manager): close azure-pipelines, bundler and cocoapods gaps
Adds a pipeline whose step task is not versioned, so nothing is
extracted, and a gem lock line carrying no version, which is skipped.
Marks four guards that cannot be false: `extractContainer()` always
returns a dep, `getLockFilePath()` always returns a path, and the
cocoapods dep name, spec name and git account all derive from values
their enclosing check already proved present.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* test(manager): close cargo, composer and cpanfile gaps
Adds a composer repository map that disables something other than
packagist, which is dropped rather than recorded.
Marks five guards: the cargo warning only fires for a crate dep and its
retry only runs when a dep was already current, a gitlab rule without a
token never passes the auth check, and both cpanfile handlers run only
once their match produced a value.
Two implicit elses stay counted, in the cargo version resolver and the
composer repository mapper, since a coverage-ignore hint does not attach
to an `else if`.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* test(manager): close deps-edn, docker-compose and dockerfile gaps
Adds three extractions: a deps.edn alias whose value is not a map, a
compose extension carrying no image, and a dockerfile escape directive
naming a character other than a backtick, which leaves the default in
place.
Marks the deps-edn repository name check, since object keys are always
strings, and notes the repository spec chain, where a coverage-ignore
hint does not attach to an `else if`.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* test(manager/flux): close the flux coverage gaps
Adds two GitRepository resources whose url is ssh rather than http, one
pinned to a tag and one to a commit. Neither yields a source url, since
only http urls can be turned into one.
Marks the Kustomization image guard: the schema rejects an image without
a name, so `extractImage()` cannot return null there.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* test(manager): mark github-actions, gitlabci and glasskube guards
None of these six needs a test. A parsed action ref is either a docker
or a repository ref, `getDep()` always returns a dep for a job container
and a service, and the remaining three need a listed file that cannot be
read, a listed file that yields no deps, or a manifest whose
repositories match no name.
The glasskube resource kind chain stays counted, since a coverage-ignore
hint does not attach to an `else if`.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* test(manager): mark gomod and gradle-wrapper guards
Four guards that need setups no test provides: a removed dep that is not
added back, a go.mod the replace massaging leaves unchanged, and a
daemon-jvm file carrying no toolchain version. A parsed go directive
always yields a version.
The additional-dependency count chain stays counted, since a
coverage-ignore hint does not attach to an `else if`.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
---------
Co-authored-by: Claude Opus 5 <michael.kriese+claude-code@mend.io>
test(manager/terraform): close the terraform family coverage gaps
Adds two terraform extractions that no test covered: a lock file that
cannot be read, and one that holds no locks. Both leave the deps without
a locked version.
Marks three guards. The lockfile update path needs a change that also
satisfies the recorded constraints, `extractTFLintPlugin()` never
returns null, and only terragrunt's already-ignored default case leaves
its result unset.
Two implicit elses stay counted, in the terraform module extractor and
the tflint key parser: a coverage-ignore hint does not attach to an
`else if`, so they carry a plain note.
Co-authored-by: Claude Opus 5 <michael.kriese+claude-code@mend.io>
test(manager/sbt,swift): close the sbt and swift coverage gaps
sbt is now fully covered by tests: a build definition referencing
symbols that were never defined, a resolver whose location is not a url,
a package file that yields nothing during a multi-file extract, and two
scala version strings that normalization leaves alone - a range and a
version with no patch part.
swift keeps two marks. `artifacts.ts` falls back to an empty depName
that extract always sets, and the null state in `extract.ts` is only
re-entered at a deps boundary. Its two `.package(` transitions stay
counted, since a coverage-ignore hint does not attach to an `else if`.
Co-authored-by: Claude Opus 5 <michael.kriese+claude-code@mend.io>
test(manager/bazel): close the bazel branch coverage gaps
Adds two extractions that no test covered: a `go_repository` carrying
both a tag and a commit, which must not be marked digest-only, and a
`maven_install` artifact whose coordinate is not group:artifact:version.
Marks the rest. `parser.ts` guards are type narrowing over the parsed
fragment tree, where the else only happens if the parser hands back a
shape it never produces. In `artifacts.ts` only http rules reach
updateArtifacts, and the empty url list already returned earlier.
The digest branch in `rules/http.ts` stays counted: a coverage-ignore
hint does not attach to an `else if`, so it carries a plain note.
Co-authored-by: Claude Opus 5 <michael.kriese+claude-code@mend.io>
Previously, we were caching private Go modules' information, even when
`cachePrivatePackages=false`.
This leads to the module path, versions, and possibly the `go`
directive being cached unintentionally.
Instead of doing so, we should make sure that we're intentionally
caching packages.
To do this, we can use a mix of the `GOPRIVATE`/`GONOPROXY` to determine
which packages shouldn't be fetched in the public Go proxy, given
they're private, following how the Go toolchain itself does this.
When interacting with the public Go proxy, we treat these packages as
public (as the public Go proxy has no means to authenticate) and so
anything outside of this should be treated as private.
When `GOPROXY` contains `direct`, we only treat a package as public only
when the public proxy is in the list too, the way Go's own default is
written. This is used to imply that `direct` is the fallback for
whatever the public proxy doesn't have.
When thre is no Go proxy set, `GOPROXY=direct` means everything comes
from source, which tells us nothing about who is allowed to read it.
This doesn't catch everything: with the default `GOPROXY`, a private
package the public proxy doesn't have falls through to `direct` and is
still cached by Renovate, unless `GOPRIVATE` says otherwise. We only
learn that no proxy served it after we've fetched it, which is after
`cacheable` is determined.
When using Renovate with a private Go proxy mirror, such as Artifactory,
this will result in no cached packages until `cachePrivatePackages:
true` is set.
This is a similar change to #45348.
Co-authored-by: Claude Opus 5 <jamie.tanna+claude-code@mend.io>
refactor(maven-wrapper): avoid matching each line twice
`extractLineInfo` ran `line.match(regex)` only to re-run `regex.exec(line)`
on the same line immediately afterwards. None of the regexes are global,
so `exec` alone is equivalent and halves the regex work per line.
Co-authored-by: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* feat(manager/github-actions): support aws-actions/amazon-ecs-render-task-definition's `image`
Per #43038, and following the existing pattern for Docker image
extraction.
Co-Authored-By: Claude Sonnet 5 <jamie.tanna+claude-code@mend.io>
* fixup! feat(manager/github-actions): support aws-actions/amazon-ecs-render-task-definition's `image`
Co-Authored-By: Claude Sonnet 5 <jamie.tanna+claude-code@mend.io>
---------
Co-authored-by: Claude Sonnet 5 <jamie.tanna+claude-code@mend.io>
* test(versioning/aws-machine-image): cover the empty version guard
`_parse()` returns null for a falsy version before it tries the AMI id
regex, and no test passed one. Checking that an empty string is not
valid closes the only branch the module was missing.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* test(datasource/buildpacks-registry): cover a response without a homepage
`latest.homepage` is optional in the registry response, but every test
supplied one, so the release result was never built without it. A
buildpack whose latest entry carries no homepage now checks that the
releases still come back and no source url is derived.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* test(datasource): cover optional-field branches in four datasources
Each of these datasources builds its release result by attaching fields
only when the registry supplies them, and the "not supplied" path was
never exercised:
- cdnjs: a library with neither homepage nor repository
- cpan: a search result where no release is marked latest, so no tag
- dart: a latest release carrying no pubspec
- ruby-version: a release table row whose first column is not a version
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* test(datasource): cover missing-metadata branches in hex and pypi
Two more datasources only attached fields when the registry supplied
them, and the absent case was untested: a hex package with neither an
html_url nor a github link in its meta, and a pypi response carrying no
releases at all.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* test(datasource): cover galaxy and repology fallbacks, mark unreachable guards
Adds a galaxy role with empty github fields, so no source url is
derived, and a repology lookup where the api fallback after a 403 finds
no matching package.
Marks four guards that are not worth a test: cpan and dart both check a
variable that is only null when the request threw, and
`handleGenericErrors()` always rethrows; conan's registry url is always
supplied by the datasource layer; and repology's resolver either returns
a body or throws.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* test(datasource): cover pod CDN branches, mark the sbt listing guards
Adds two pod lookups: an empty CDN body, and a CDN index that lists no
matching pod.
Marks the guards whose tests would need multi-level directory listing
fixtures: both sbt datasources walk nested index pages, and pod's github
helper checks a body that an empty json response never produces because
parsing throws first.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* test(datasource): cover maven metadata without a latest tag
Adds maven metadata carrying a release but no latest, so respectLatest
is left alone. Marks the three deb guards, which would each need a
crafted package index: a block missing a required key, a repeated
package in the final block, and two component indexes whose meta agrees.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* test(datasource): cover npmrc registry fallback, mark cache-bound guards
Adds a resolveRegistryUrl lookup for a package no npmrc rule matches.
Marks the guards whose tests would need a primed cache or a specific
blob shape: nuget's services index, deno's release cache, the docker hub
cache window, docker's manifest and config handling, go's pseudo-version
fallback, and both rubygems cache layers.
Three implicit elses are left counted: a coverage-ignore hint does not
attach to an `else if`, so the registry strategy chain, the npmrc reset
and docker's digest-header branch carry a plain note instead.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* fix(datasource/repology): pass the mock body as a string
`ResponseMock.body` is typed as a string, and every other call in the
spec passes `'[]'`. The new api fallback test passed an array literal,
which tsc rejected.
Co-Authored-By: Claude Opus 5 <michael.kriese+claude-code@mend.io>
---------
Co-authored-by: Claude Opus 5 <michael.kriese+claude-code@mend.io>
* feat(manager): pass extracted constraints to updateArtifacts
`extractedConstraints` were collected during extraction but only ever read
by `fetch.ts` for the lookup. At artifact time every manager had to derive
the same tool constraints a second time, and when the updated package file
no longer carried the constraint there was nothing to fall back to.
`generateBranchConfig()` now merges the `extractedConstraints` of all
upgrades of a branch onto the branch config, which reaches `updateArtifacts()`
through `patchConfigForArtifactsUpdate()`. `UpdateArtifactsConfig` declares
the field and documents what it contains.
The new `resolveToolConstraint()` helper implements the ladder once:
user `constraints` > an optional artifact-time derivation > the extracted
value. The derivation is ranked above the extracted value on purpose: it
reads the package files as the branch changed them, so poetry (python is an
updatable dependency there), composer, pipenv (`pipenv` can be a dev
package) and pep621 (`requires-python` is an updatable dependency) would
otherwise regress to a stale constraint. pipenv, pep621 uv/pdm, poetry and
composer are migrated to the helper and keep their derivation.
Behaviour change: these managers now use the extract-time constraint as a
last resort instead of running unconstrained, for example on lock file
maintenance where pipenv deletes `Pipfile.lock` before deriving.
`execute-post-upgrade-commands.ts` is unaffected: it reads
`upgrade.constraints` per upgrade, and the extracted values are kept in a
separate field rather than folded into `constraints`, so post-upgrade task
tool constraints still come from user config only.
Not migrated, because their constraints are never extracted, so the
fallback would be dead code: bundler (`ruby`/`bundler`), pixi (`pixi`),
pep621 uv (`required-version`) and npm's node ladder, whose last step
already reads `package.json` engines.
* refactor(manager): resolve tool constraints through resolveToolConstraint()
Every manager now obtains the constraint of a tool it runs through
`resolveToolConstraint()` instead of reading `config.constraints` itself, so
the precedence "user config, then a value derived from the updated files, then
the constraints collected during extraction" applies everywhere. The helper
treats an empty string as not set at every step and accepts the
`PostUpdateConfig` shape, whose `constraints` may be `null`.
Behaviour changes that go beyond the new fallback to extracted constraints:
- `mix` applies its hardcoded erlang default only after the extracted
constraint, so an erlang version collected at extraction can win over it.
- `kustomize` decides `HELM_EXPERIMENTAL_OCI` from the fully resolved helm
constraint; `generateHelmEnvs()` takes that value instead of the config.
- `npm` (yarn): an upgrade of yarn without a new value now consults the user
constraint before deriving one from package.json or the lock file.
- The "Using <tool> constraint from config" debug lines in bundler, composer,
copier and pip-compile are gone, as the helper is the only reader of the
config now.
Helpers that read the config synchronously became async: bundler
`getBundlerConstraint()`, copier `getPythonVersionConstraint()` and
`getCopierVersionConstraint()`, pip-compile `get*VersionConstraint()` and
`getToolVersionConstraint()`, gomod `deriveGoToolchainConstraints()` and
`getUpdateImportPathCmds()`, mise `getMiseLockToolConstraints()`. composer
`getPhpConstraint()` is removed.
Each converted manager has a test showing the extracted constraint reaching
the executed tool, and, where a derive callback exists, that the derived value
still wins over it.
* refactor(linters): add prefer-resolve-tool-constraint rule
Reports member reads and destructuring of `constraints` and
`extractedConstraints` from a manager config in `lib/modules/manager/**`, so
that tool constraints are resolved through `resolveToolConstraint()` in
`lib/modules/manager/util.ts` only. Writes and the same properties on package
files, which extraction has to populate, are not reported. The helper itself
and spec files are exempt in `.oxlintrc.json`.
* refactor(manager): inline thin resolveToolConstraint wrappers
Drop copier/utils.ts and the pip-compile getPipToolsVersionConstraint /
getUvVersionConstraint helpers, which only forwarded to
resolveToolConstraint(). Callers now call it directly.
* docs(development): note lock file maintenance in tool constraints section
* refactor(manager): address review on tool constraint resolution
composer only runs extractConstraints() when a user constraint is missing,
via Lazy. UpdateArtifactsConfig no longer re-declares constraints and
extractedConstraints, which ToolConstraintsConfig already provides.
* refactor(platform): share one implementation between gitea and forgejo
The `forgejo` platform was copied from `gitea` in #37075 and the two have
drifted since: fixes landed on one side only (zod validation #43900, empty
email #44068, null issues #44048, team reviewers #39533, merge-style fallback
#37768 vs #38040), and lint sweeps had to be applied to both copies by hand.
Both platforms now build on `lib/modules/platform/gitea-common/`, which holds
the zod schemas, API helpers, utils, PR cache and the platform object itself.
`createPlatform()` closes over its own state, so `gitea` and `forgejo` keep
separate config, caches and host types. `gitea/index.ts` and `forgejo/index.ts`
are thin wrappers that supply the platform id, default endpoint, http client
and the version gates that genuinely differ.
Reconciliations, forgejo's behaviour adopted for gitea:
- Responses are validated with zod instead of `getJsonUnchecked` casts. Gitea
therefore tolerates empty emails, null PRs in list responses and lenient
repo/issue payloads the same way Forgejo already did.
- `getRepoContents` returns a discriminated union, so `getRawFile` returns null
for dir/symlink/submodule entries instead of reading a missing field. The
gitea preset resolver now throws PRESET_INVALID for non-file entries, like
the forgejo one.
- An unknown `default_merge_style` (e.g. `manually-merged`) is caught by the
schema and skipped, instead of aborting the repository with
REPOSITORY_BLOCKED from `isAllowed()`.
- `initRepo()` resolves whether the repo owner is an org and only fetches org
labels when it is, adding one `/orgs/{owner}` request and an `orgs` entry in
the gitea repository cache.
- Reviewers prefixed with `team:` are sent as `team_reviewers`. Gitea keeps its
`v1.14.0` guard for requesting reviewers at all.
- `getBranchStatusCheck()` maps unknown commit status values through the
schema's `unknown` fallback rather than warning and returning yellow.
Gitea's behaviour is kept where it is deliberately different: the Forgejo
instance detection and its deprecation notice, the `1.24.0` (or `10.0.0` for a
detected Forgejo) native automerge gate, and the reviewer version guard are
passed in as options.
Two shared log messages became static so they can be grouped in metrics:
`{platform}-native automerge: fail` is now `Platform-native automerge: fail`
and `{Platform} getRepos() error` is now `Error auto-discovering repositories`,
both with the platform id in the metadata. The dead `defaults.hostType` field
is gone.
* refactor(platform): keep the shared gitea implementation under gitea/
Drop the `gitea-common/` folder. The zod schemas, API helpers, utils, PR
cache and `createPlatform()` live in `lib/modules/platform/gitea/` under
their original file names, and `forgejo/` imports them from there.
The preset resolvers pass `giteaHttp` and `forgejoHttp` to
`getRepoContents()` directly, which removes the two bound wrappers. The
gitea changelog source uses the shared `RepoContents` type like the
forgejo one already does.
* refactor(http): derive ForgejoHttp from GiteaHttp
The two clients were identical apart from the default `hostType`.
`ForgejoHttp` now extends `GiteaHttp` and only keeps its own base url,
so the pagination logic exists once. The platform helpers accept
`GiteaHttp` directly instead of the `HttpBase` alias.
Also covers the `updatePr()` branch where the updated pull request
cannot be converted and therefore is not cached.
* refactor(platform): create the gitea and forgejo http clients in index.ts
`forgejo-helper.ts` only held the `ForgejoHttp` instance, so it is gone
and `forgejo/index.ts` creates the client next to the platform. Gitea
does the same for symmetry, which leaves `gitea-helper.ts` with only the
shared API helpers.
* refactor(platform): fetch gitea labels and issues with await
`getLabelList()` and `getIssueList()` still memoize the in-flight
promise, but build it from small async functions instead of `.then()`
and `.catch()` chains.
* refactor(platform): keep declaration order of gitea/index.ts
Order the module the way it was before `createPlatform()` so the diff
against main aligns: platform id and http client after the repo config
interface, pure helpers before the factory, and `maxBodyLength()` after
the platform object.
* refactor(platform): move gitea label fetching to labels.ts
`fetchLabelList()` takes the http client and the repository details
instead of closing over the platform state, and gets its own spec. The
factory only keeps the memoizing `getLabelList()`.
* refactor(platform): move gitea label lookup to labels.ts
* refactor(platform): create the gitea pr cache per platform
`GiteaPrCache` is now created once per platform with its http client and
platform id, and `initRepo()` records the repository. The per-repository
state moves to a private `RepoPrCache` which is bound to the repository
cache on first use and reused until the next `initRepo()`, instead of
being rebuilt on every lookup.
Where we are able to support Java distributions that follow a
straightforward release cadence, we opt these distributions in, and skip
others as `unsupported`.
As the Temurin/Adopt distributions can be reliably tracked by the
`java-version` Datasource, backed by Adoptium, we can support them.
We can also use the `java-package` key to look up whether we want a JRE
or JDK.
Co-authored-by: Claude Sonnet 5 <jamie.tanna+claude-code@mend.io>
In the case a multi-line `dotnet-version` is specified to install
multiple SDKs at once, we skip it as `invalid-version`, as we don't
currently have a way to update multiple values in the field.
Co-authored-by: Claude Sonnet 5 <jamie.tanna+claude-code@mend.io>
Similar to how we already generate a list of the community actions we
support, we should also generate which first-party Actions from GitHub
are supported, and simplify some of the wording in the docs.
Instead of duplicating this, we can instead move them into a single
section in the documentation.
Co-authored-by: Claude Sonnet 5 <jamie.tanna+claude-code@mend.io>
* refactor(manager/github-actions): don't duplicate parsing logic for first-party Actions
Instead of having a separate set of Action parsing and their
configuration between first-party GitHub-owned Actions and those of
third-party Actions, we can merge them into a single collection.
This also leads to `https://<host>/` handling noted in #35777 now being
handled out-of-the-box, as it was already supported on third-party
actions.
Co-Authored-By: Claude Sonnet 5 <jamie.tanna+claude-code@mend.io>
* fixup! refactor(manager/github-actions): don't duplicate parsing logic for first-party Actions
---------
Co-authored-by: Claude Sonnet 5 <jamie.tanna+claude-code@mend.io>
* fix(managers/pep621): use lowest locked version from uv
* reduce diff
* add file extension to import
* use correct file path in import
Co-authored-by: Michael Kriese <michael.kriese@gmx.de>
---------
Co-authored-by: Michael Kriese <michael.kriese@gmx.de>
Jamie Tanna