Add regression tests for the fix that allows a null or missing
home_page in PyPI JSON responses (#43814). Covers both the schema
parser and the datasource, asserting that homepage is left unset and
sourceUrl/changelogUrl fall back to project_urls when home_page is null.
* fix(datasource/pypi): allow null home_page in PyPI JSON response
PyPI returns "home_page": null for projects that declare only
project_urls. The strict z.string().optional() schema threw a ZodError
in getDependency, aborting the JSON-API lookup and resolving such
packages to no-result. Use .nullish() to accept null, matching how
requires_python already handles PyPI's nulls.
* fix(datasource/pypi): allow missing home_page in PyPI JSON response
Apply the suggestion to chain `.optional()` and `.nullish()`.
* fix(datasource/npm): tolerate non-string `time` entries in packument
JFrog Artifactory npm registries emit `"unpublished": null` under the
packument `time` field. Since the schema declared `time` values as
`z.string()`, the `null` failed validation and Zod rejected the entire
packument, so every package lookup returned no-result.
Strip non-string `time` entries via a shared `PackumentTime` transform so
a stray metadata key cannot invalidate the whole packument, while keeping
the real version->timestamp entries the datasource relies on.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* refactor(datasource/npm): use LooseRecord for packument time
Per PR review.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(datasource): replace getJsonUnchecked with validated getJson in npm/Node/PyPI/Go
Replace all getJsonUnchecked calls in npm, Node.js, PyPI, and Go datasources
with Zod-validated getJson calls:
- datasource/npm: add NpmResponseSchema to schema.ts (shared with presets)
- config/presets/npm: migrate to NpmResponseSchema with passthrough version fields
- datasource/node-version: add NodeReleasesSchema using LooseArray
- datasource/pypi: add PypiJSONSchema (nullable+catch for lenient empty-body handling)
- datasource/go: add VersionInfoSchema for both versionInfo and getLatestVersion calls
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* some manual cleaning up
* removing schema level catch
* removing some more types
* refactor: rename schema
* fix linter errors
* remove schema suffix which hasn't been catched by linter
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(datasource): replace getJsonUnchecked with validated getJson in misc+merge-confidence group
Replaces all getJsonUnchecked calls with Zod-validated getJson calls in:
- crate: adds CrateMetadataSchema to existing schema.ts
- conan: uses existing ConanJSON schema directly
- puppet-forge: new schema.ts with PuppetModuleSchema
- conda: new schema.ts with CondaPackageSchema
- repology: new schema.ts with RepologyPackagesSchema
- jenkins-plugins: new schema.ts, parameterizes helper with ZodType<T>
- custom/formats/json: uses z.unknown() for user-defined responses
- merge-confidence: adds inline MergeConfidenceResponseSchema
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: correct nullable/optional schema fields for crate, jenkins-plugins, puppet-forge
- crate: remove .optional() from all CrateMetadata fields to match string|null type
- jenkins-plugins: use .default({}) instead of .optional() for plugins fields so schemas match required Record types; widen ZodType input to unknown; update JenkinsPluginVersion.releaseTimestamp to Timestamp|null to match MaybeTimestamp output
- puppet-forge: change releases from .optional() to .default([]) to match Release[] type
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* chore: migrate zod/v3 imports to zod/v4 in misc datasources and merge-confidence
Update remaining zod/v3 imports in files added/modified by the getJsonUnchecked
migration to use zod/v4 as required after the upstream migration.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* chore minor improvements
* remove unnecessary types and reintroduce nullable where necessary
* refactor: move logic into transform
* test: add tests to improve coverage
* chore removed some unused types
* remove type declaration
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
* Add support for Azure DevOps in dependency dashboards feature
related #9592
* Add missing test for got wrapper
* Apply all current CR suggestions
* fix: resolve eslint and tsx issues in Azure platform code
* test: add coverage for issue.number undefined cases
Add test cases to ensure proper handling when:
- Trying to reopen an issue without a number
- Trying to update an issue without a number
This improves code coverage for the null checks added in previous commit.
* test: fix and add tests of azure/issue.ts to up coverage
there was mismatch in mock resulting in testing of different code branch
* refactor: address review feedback
* fix: type-check
* refactor: address review feedback round 2
* apply suggestions
* Apply suggestions from code review
Co-authored-by: Sebastian Poxhofer <secustor@users.noreply.github.com>
* refactor: use partial helper and remove host-rules mock in issue.spec.ts
* more changes from suggestions
* Update lib/modules/platform/azure/issue.spec.ts
Co-authored-by: Michael Kriese <michael.kriese@visualon.de>
---------
Co-authored-by: Viktorie Novotná <viktorie.novotna@drmax.eu>
Co-authored-by: Sebastian Poxhofer <secustor@users.noreply.github.com>
Co-authored-by: Michael Kriese <michael.kriese@visualon.de>
* fix(nuget): replace getJsonUnchecked with validated getJson
Migrate NuGet v3 datasource from unvalidated HTTP responses to Zod-schema
validated ones. Creates schema.ts with ServicesIndexRaw, CatalogPage, and
PackageRegistration schemas derived from the existing TypeScript interfaces.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(nuget): migrate zod import from v3 to v4
Update schema.ts to use 'zod/v4' instead of 'zod/v3' following the
upstream migration in renovatebot/renovate PR #43691.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* chore: some manual improvements
* refactor: rename schema
* fix linter errors
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(terraform-provider): replace getJsonUnchecked with validated getJson
Convert all five TypeScript interfaces (TerraformProviderVersions,
TerraformProviderReleaseBackend, TerraformRegistryVersions,
TerraformRegistryBuildResponse, VersionDetailResponse) to co-located
Zod schemas in schema.ts and update all five call sites in index.ts to
use the typed http.getJson() overload. Types are now derived with
z.infer<> to prevent interface/schema drift.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix linting errors
* remove types
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
* docs: skip validation on JSON files that aren't Renovate config
As part of future changes, we'll enable validation on all JSON code
snippets to run Config Validation.
Before we do that, we should opt-out JSON snippets that aren't anything
related to Renovate config.
* docs: don't validate invalid Renovate config
As it's intentionally referencing a manager that doesn't exist.
* docs: don't validate partial Renovate config
As part of future changes, we'll enable validation on all JSON code
snippets to run Config Validation.
There are several code snippets we have that are part of a Renovate JSON,
but it doesn't make sense to convert these to a fully passing config, as
it'd reduce clarity on what we're trying to explain.
* docs: convert `baseBranchPatterns` snippets to code blocks
So we can validate them.
* docs: make sure that `datasourceTemplate` is set
So this passes validation.
* docs: clarify that `force` is global config
As it may not be clear, so we can make it extra clear with a `.js` file,
and note that it's global self-hosted config.
* docs: make sure `packageRules` is an array
* docs: `matchBaseBranches` needs `baseBranchPatterns` to be set
* docs: correct typo for `prBodyColumns`
* docs: mark global config snippets as such
As part of future changes, we'll use this to detect whether it's global
self-hosted configuration.
* docs: make sure `packageRules` wraps its children
* docs: add `allowedHeaders`
* docs: ignore deprecation warnings
As part of future changes, we'll use this to note that we don't care
about the validation warnings.
* docs: make sure `customManagers` wraps its children
* docs: make sure `fleet` wraps its children
* docs: validate JSON snippets are valid Renovate config
To make sure that we're always presenting valid JSON configuration to
our users, we can validate them as part of `check-fenced-code`.
As well as configuration errors we can warn for warnings, but allow code
blocks to opt-out, for instance when there's a known deprecation
affecting the code block.
We need to make sure to set `allowedEnv` for various code snippets that
reference environment variables.
* docs(development): document fenced code block checks
* docs: wrap `configType` and `ignoreConfigWarnings` in `{ ... }`
As otherwise we see build failures i.e.
WARNING - Doc file 'modules/platform/gitlab/index.md' contains a link
'../../../self-hosted-configuration.md#autodiscoverrepoorder', but the doc
'self-hosted-configuration.md' does not contain an anchor '#autodiscoverrepoorder'.
This needs to be wrapped alongside all other options like `title`.
* fixup! docs: wrap `configType` and `ignoreConfigWarnings` in `{ ... }`
* chore: migrate from zod/v3 to zod/v4
This commit contains the logic changes
* chore: migrate imports from zod/v3 to zod/v4
This contains import updates and adds explicit record keys as zod/v4 requires these now
* ci: enforce usage of zod/v4
* ci: add local mkdocs-admonitions plugin for prettier to allow admonition preservation
* docs: remove uses of `prettier-ignore` for admonitions
* test: rather use await expect
* chore: use named group exclusively
* fix(maven): determing properly closed ranges
A Maven range like [1,2) contains various (usually pre-release) versions of the 2.x line.
By Maven definition the smallest possible 2.x version is 2-alpha, so a properly closed range that only allows 1.x versions
would be [1,2-alpha). This is now detected and properly advanced.
* Review Feedback Vol. 1
* Do not coerce as that will be fixed by #38895
* Ignore uncoverable if condition coverage
* Extract constant
* Review Feedback Vol. 2
* docs(manager/circleci): document YAML merge key parsing limitation
Signed-off-by: Ethan Setnik <esetnik@gmail.com>
* fix(manager/circleci): parse config files as YAML 1.1
CircleCI processes config files as YAML 1.1, but Renovate parsed them
as YAML 1.2. Configs that use YAML 1.1 features failed to parse and
the entire file was silently skipped, dropping every `docker` and `orb`
dep in it.
The most common trigger is a mapping with more than one `<<` merge key,
which YAML 1.2 rejects as a duplicate key. CircleCI's own processor
allows it, since merge keys are a YAML 1.1 feature.
Pass `{ version: '1.1' }` to `parseSingleYaml` so the parser matches
CircleCI's behavior, and add a regression test that covers the
duplicate-merge-key case.
Replaces the workaround note added in the previous commit on this
branch (per review feedback) — the config-side `%YAML 1.1` directive is
no longer needed.
---------
Signed-off-by: Ethan Setnik <esetnik@gmail.com>
* docs: ensure 2 spaces used for generated admonitions
As a follow-up to 4b0146ab7b and
1844916259b6dc17f532e491eb93f6116327c324., this correctly renders
admonitions with 2 spaces (1 indent).
* fixup! docs: use 2-space indentation for Markdown (#43537)
Plus using 3-space for mixed-style `<ul>` and `<ol>`s.
* fixup! docs: use 2-space indentation for Markdown (#43537)
As otherwise it breaks the rendering of the table.
* feat(datasource/github-runners): stabilize arm64
Mark the Ubuntu ARM runners as stable.
* feat(datasource/github-runners): add test
Add test for upgrading ubuntu-22.04-arm runners.
* feat(datasource/github-runners): add windows-11-arm label
Add support for Windows 11 ARM runners in the GitHub Runners datasource.
* feat(datasource/github-runners): add test
Add test for extracting windows-11-arm runners.
* feat(datasource/github-runners): fix test
Update expected length.
* feat(pip-compile): support --constraint alias and --overrides for uv
uv accepts both --constraint and --constraints (clap aliases) and
records the exact form used into lockfile headers. Only --constraints
was in the uv allowlist, so --constraint headers were rejected.
Similarly, uv supports --override/--overrides for version overrides,
but neither form was recognized by Renovate at all.
This commit:
- adds --constraint to uvOptionsWithArguments (value handler already
handles both forms)
- adds --override and --overrides to uvOptionsWithArguments with a
value handler that parses overridesFiles
- tracks override files as dependencies in the file graph (like
constraints) so lockfiles are recompiled when overrides change
- uses dotted mermaid edges for overrides (same as constraints)
Ref: https://github.com/renovatebot/renovate/discussions/39488
Co-authored-by: Cursor <cursoragent@cursor.com>
* test(pip-compile): add coverage for overrides dependency tracking
Add a test case in extract.spec.ts that exercises the overridesFiles
loop in extractAllPackageFiles, fixing the codecov/patch failure.
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(datasource/rust-version): skip blank lines in manifests.txt
`https://static.rust-lang.org/manifests.txt` ends with a trailing
newline, so `response.body.split('\n')` always yields an empty final
element. `parseManifestUrl('')` returns `null`, which triggered the
`Failed to parse manifest URL` warning on every run. The warning then
surfaces as a `Repository Problems` entry on the Dependency Dashboard
for any repo with a rust-version-tracked dep (`rust-toolchain.toml`,
or a custom-regex manager pointing at the `rust-version` datasource).
Skip blank/whitespace-only lines silently; keep warning on genuinely
malformed non-empty lines (covered by the existing
"ignores unexpected URLs" test).
* docs(datasource): condense comment
Co-authored-by: RahulGautamSingh <rahultesnik@gmail.com>
* Remove accidental newline
---------
Co-authored-by: RahulGautamSingh <rahultesnik@gmail.com>
Co-authored-by: Jamie Tanna <github@jamietanna.co.uk>
* feat(rust-toolchain): Implement `rust-toolchain` manager
* feat(rust-toolchain): Add support for legacy filename
* feat(rust-toolchain): Add support for legacy file format
* refactor(rust-toolchain): Reject empty `channel` strings at parse stage
* tests(rust-toolchain): Cover valid TOML missing `[toolchain]` or `channel`
* refactor(rust-toolchain): Log legacy format fallback at `trace` level
* refactor(rust-toolchain): Reject empty and multi-line legacy files explicitly
* tests(rust-toolchain): Cover unparseable channel value
* tests(rust-toolchain): Use `codeBlock` for multi-line fixtures
* docs(rust-toolchain): Scope pin example to `rust-toolchain` manager
Co-authored-by: Sergei Zharinov <1239644+zharinov@users.noreply.github.com>
* tests(rust-toolchain): Use `codeBlock` for schema fixtures too
---------
Co-authored-by: Sergei Zharinov <1239644+zharinov@users.noreply.github.com>
* feat(mise): add lock file support
- Add lockfile.ts with getLockFileName() for deriving lock file paths
- Add artifacts.ts with updateArtifacts() and updateLockedDependency()
- Add MiseLockFile Zod schema for parsing lock files
- Update extract.ts to be async and extract lockedVersion from lock files
- Export supportsLockFileMaintenance, lockFileNames, updateArtifacts, updateLockedDependency
* refactor(mise): remove redundant comments
Remove comments that simply restate what the code already clearly
communicates. Keep valuable comments that explain non-obvious behavior
(e.g., why temporary errors are rethrown) and move helpful examples
into JSDoc comments.
* fix(mise): improve lock file version lookup and add tests
- Fix lock file key lookup for non-registry tools (aqua:, ubi:, etc.)
Registry tools use short names (node), non-registry use full names
(aqua:cli/cli). New getLockedVersion() tries both patterns.
- Add injection prevention test for shell quoting
- Add update-failed test for error handling
- Add comprehensive tests for getLockedVersion()
- Add null check for depName in updateLockedDependency
- Add JSDoc to updateArtifacts and extractPackageFile
- Add lock file support documentation to README
- Remove code duplication by consolidating into lockfile.ts
* fix(mise): pass --local flag and MISE_ENV for env/local config files
When updating mise.local.toml, `mise lock` updates mise.lock, not
mise.local.lock. Add --local flag to target the correct lock file.
For env-specific configs (mise.test.toml), set MISE_ENV=test so mise
reads and locks the correct configuration.
Extracted getConfigType() from getLockFileName() to share the parsing
logic between lock file name derivation and command flag generation.
Builds on #42099 by altendky.
* test(mise): cover tool absent from lock file
* refactor(mise): remove dead branches in extractPackageFile
tools is required by MiseFile schema so if(tools) is always true.
depName is always set by createDependency so if(dep.depName) is always true.
Add test for tool absent from lock file to cover the lockedVersion
assignment branch.
* fix(mise): address review feedback on lock file support
* fix(mise): remove unsupported tool constraint
* fix(mise): add containerbase tool definition
* fix(mise): add typed exec constraint support
* fix(config): restore requireConfig typing
* test(mise): await registry extraction cases
* Update lib/modules/manager/mise/artifacts.ts
Co-authored-by: Michael Kriese <michael.kriese@gmx.de>
* fix: revert unrelated types cleanup
* fix: drop missing VersioningName import
* fix(mise): use npm versioning string after import removal
* test(mise): set binarySource=global in artifacts spec
GlobalConfig.get now returns defaults; without explicit binarySource the
test fell into containerbase dynamic install and tried to resolve the
mise version via the github-releases datasource, causing missing HTTP
mock errors. Force binarySource=global so the spec only exercises the
mocked exec path.
---------
Co-authored-by: Kyle Altendorf <sda@fstab.net>
Co-authored-by: Michael Kriese <michael.kriese@gmx.de>
Co-authored-by: Jamie Tanna <jamie.tanna@mend.io>
As noted in #43445, there are cases where:
- an infrequently updated repository prepares a GitHub Release by
creating a draft release
- some time passes, and the maintainers publish the Release
- no new commits are pushed to the release branch (i.e. `main`) in that
time
In this case, Renovate would previously use the `created_at` of the Git
tag (according to the GitHub API) rather than the `published_at` time
from the GitHub Release.
This is correct when using a plain Git tag, as we only have metadata to
indicate when the commit the tag points to was created.
However, when we have GitHub Release information, we should use that to
take precedence.
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
In 01bcd4fd43 we introduced knowledge of
the Maven Central mirror, but haven't taken it into account when looking
at rate limit errors.
To prevent this in the future, we can introduce a new helper function,
`isMavenCentral`, for other usage.
As a way of indicating how fresh our list of supported tools in the
registry is, we can make sure we populate the version into a note above
the table.
We can extract a type to make it easier to reference the parsed JSON
file.
This requires a small tweak to handle `content`'s table insertion.
The git-submodules extractor was setting sourceUrl to the raw git HTTP URL,
including the .git suffix (e.g. https://github.com/owner/repo.git).
This caused the 'githubDigestChangelogs' preset's changelogUrl template
'{{sourceUrl}}/compare/{{currentDigest}}..{{newDigest}}' to produce
invalid URLs like:
https://github.com/owner/repo.git/compare/abc123..def456
which 404s because GitHub compare URLs don't support the .git suffix.
Other managers (bundler, deps-edn, flux) already strip .git from sourceUrl.
This fix follows the same pattern.
* feat(manager/swift): support SSH URLs in Package.swift extraction and artifact matching
Swift packages can use SSH URLs (e.g. `git@github.com:owner/repo.git`) for
dependencies, but the Swift manager only supported HTTPS URLs. This caused
SSH-based dependencies to be silently ignored during extraction, and
Package.resolved pin matching to fail when protocols differed.
- Convert SCP-style SSH and `ssh://` URLs to HTTPS in `parseUrl()` using
the existing `getHttpUrl()` utility before platform detection
- Enhance `normalizeUrl()` in artifacts to use `parseGitUrl()` for
protocol-agnostic URL comparison (falls back to basic normalization
when parsed components are incomplete)
* test(manager/swift): add edge case tests to improve coverage
- Add test for unparseable SSH URL in extraction
- Add test for host-only URL fallback in artifact matching
* fix(manager/swift): remove redundant .git stripping in normalizeUrl
parseGitUrl() already strips the .git suffix from full_name,
so the additional regex replacement is unnecessary.
* test(manager/swift): add ssh:// protocol test for artifact matching
Cover the ssh://git@github.com/... URL format in addition to the
existing git@ SCP-style test for Package.resolved pin matching.
* refactor(manager/swift): use trimTrailingSlash helper in normalizeUrl
* refactor(manager/swift): always normalize URLs in parseUrl
* fix(datasource/maven): cache 404s on `maven-metadata.xml` for 24h
As part of efforts to reduce the general overload on Maven Central[0]
from Renovate, we can look to reduce unnecessary HTTP lookups.
The Maven Datasource will use the `registryStrategy=merge`, which leads
to looking up a package in each of our registries, and merging them if
present.
However, this means that when looking up a package that /will only ever
be in a given registry/ (such as in
01bcd4fd43) we're hitting other registries
with packages they are unlikely to have.
This is unnecessary, significantly increases traffic to Maven Central,
and leads to us re-fetching the same 404s which count towards rate
limits, and increased, unnecessary, server load on Maven Central.
We can instead introduce a new cache for 404s for the
`maven-metadata.xml`, for 24 hours, which should reduce traffic to Maven
Central (and other configured registries) if they only have a subset of
dependencies.
[0]: https://www.sonatype.com/blog/maven-central-and-the-tragedy-of-the-commons
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* fixup! fix(datasource/maven): cache 404s on `maven-metadata.xml` for 24h
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* cache-down
* perf(datasource/maven): add jitter
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* sq
---------
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
Sebastian Poxhofer