fix(datasource/maven): cache `.pom` URLs for 7 days
As part of efforts to reduce the general overload on Maven Central[0]
from Renovate, we can look to reduce unnecessary HTTP lookups.
For instance, when a `.pom` URL is retrieved, it is re-checked every 15
minutes in case it has been updated.
This is unnecessary, significantly increases traffic to Maven Central,
and leads to us re-checking constantly for artifacts that are not
updating.
Each time we perform the lookup, we'll use the ETag result to do an
`If-None-Match`, but this still introduces an HTTP call upstream that
counts towards rate limits, and server load on Maven Central.
Instead, we can wrap the lookups for immutable pom.xml results with a 28
day soft cache. In a rare case where a `.pom` URL is updated, we'll then
re-fetch it after 28 days, but until then, it'll be safely returned from
cache.
[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>
* fix(config/global): return default values for `GlobalConfig.get`
As noted in #39949, it can be awkward working with in `GlobalConfig.get`, where
we would previously need to specify the default value when calling `.get`,
which could lead to drift with the defaults for a given config option.
To reduce this risk, we can rely on a given `default` for a configuration
option, providing that as a default.
Because this now definitely returns a value that matches the config option, it
now allows us to remove `!` assertions.
* fixup! fix(config/global): return default values for `GlobalConfig.get`
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
* fixup! fix(config/global): return default values for `GlobalConfig.get`
* test(presets): remove impossible test
As it's now not possible to set a value as `undefined` (as
`GlobalConfig` will then use the default), this test no longer validates
anything.
* test(workers/repository): add mock
Now `platform` defaults to `github`, we need to mock out the call to
fetch config.
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* test: coverage
* test: coverage
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* test: ensure we set `binarySource=global`
As part of future changes, we'll correctly return the default value from
`GlobalConfig.get`, instead of `undefined` if no default is provided in
the `get` call.
As `binarySource=install` is the default, this then leads to these tests
failing, as they're intended to run with `binarySource=global`.
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* fixup! test: ensure we set `binarySource=global`
* fixup! fixup! test: ensure we set `binarySource=global`
---------
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* docs: add inline JSDoc for `RegistryStrategy`
Noticed while making other changes, that it wasn't clear looking at the
type what these values meant.
We can use a `@see` annotation to move the inline documentation from the
`DatasourceApi` to use our source-of-truth.
* docs: update `RegistryStrategy` docs in line with implementation
With help from Claude Sonnet to derive more information from the
implementation.
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
---------
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* fix(datasource/maven): don't look up potential Gradle Plugins in Maven Central
As part of efforts to reduce the general overload on Maven Central[0]
from Renovate, we can look to avoid lookups that we know are likely to
never resolve.
For instance, via #43146, we can skip any packages that appear to be a
Gradle Plugin, as they will only ever be sourced from the Gradle Plugin
Portal.
[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>
* Update lib/modules/datasource/maven/readme.md
Co-authored-by: Michael Kriese <michael.kriese@visualon.de>
---------
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
Co-authored-by: Michael Kriese <michael.kriese@visualon.de>
* test(noNewURL): add new rule enforcing parseUrl() from 'lib/util/url.ts'
* replace newURL with parseUrl
* remove non-null assertions marks if appropiate
* test: add additional tests
* test: add additional tests and ignore where it makes sense
* implement suggestions
* test: further improve coverage
* apply suggestion
* refactor(docs): don't use an `end` for autogenerated docs
As part of #42910, we wanted to avoid needing a "begin" and "end" to
each of our autogenerated docs replacements.
While looking into it, we're not actually using the "end" for anything,
so we can instead refactor this to be a single-line replacement.
To simplify the code and docs, we can remove this. At the point we need
this again, we can re-add the functionality.
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* fixup! Merge remote-tracking branch 'origin/HEAD' into chore/no-replaceStop
---------
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* fix(github): set commit message explicitly for platform automerge
When GitHub's "Use PR title and body as commit message" setting is
enabled, the verbose Renovate PR description gets used as the squash
commit body. Pass `commitHeadline` and `commitBody` to the
`enablePullRequestAutoMerge` GraphQL mutation so the actual commit
message is used instead.
* add REBASE test
* pr feedback
* feat(managers/maven): support `pom.template.xml` files
For users who are using templating for their POMs, we can parse them by
default, rather than requiring users explicitly add this to the
`managerFilePatterns`.
We can also add a test to validate the functionality.
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* fixup! feat(managers/maven): support `pom.template.xml` files
---------
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* central function for creating SimpleGit instance
* Updated tests.
* Set the LANG and LC_ALL environment variables for all Git operations instead of just initRepo.
* Moved GIT_SSH_COMMAND to `createSimpleGit` function.
* Removed unnecessary custom env vars.
When a GitLab project has merge trains enabled, automerged MRs are now added to the merge train via POST /projects/:id/merge_trains/merge_requests/:iid with auto_merge:true (GitLab 17.11+). Non-merge-train projects keep using the existing /merge endpoint, preserving Renovate's pipeline-success guarantee for free-tier users.
* fix(platform): preserve explicit local dryRun
* fix: preserve explicit dryRun in local platform
* fix(platform/local): only preserve dryRun=extract; fall back to lookup otherwise
Signed-off-by: Gregor Zeitlinger <gregor.zeitlinger@grafana.com>
* docs(platform/local): document dryRun=extract override
---------
Signed-off-by: Gregor Zeitlinger <gregor.zeitlinger@grafana.com>
* feat(manager/renovate-config): extract `constraints`
As noted in #41571, it would be useful to detect and suggest updates
when specifying `constraints` in Renovate config, where it's a known
tool we provide versioning information for.
We can extend our existing `renovate-config` manager to add support for
`constraints` that are known Containerbase tools.
We can support top-level `constraints` as well as `constraints` inside
`packageRules`.
* refactor(renovate-config): remove nesting
When a Go dependency uses a pseudo-version (v0.0.0-timestamp-hash) and
Renovate detects a newer pseudo-version, updateType is set to 'digest'
via the gomod special-case in lookup/index.ts. The digest update code
path was then writing only the 12-character bare commit hash, producing
invalid go.mod syntax:
- knative.dev/pkg v0.0.0-20250312035536-b7bbf4be5dbd
+ knative.dev/pkg 4a022ed9999a
Historical cause:
In 2018, the Go module proxy did not yet exist. Renovate queried the
GitHub API for the latest commit SHA and had no way to obtain the correct
commit timestamp, so it fabricated a pseudo-version using the wall-clock
time when Renovate ran. The result was syntactically valid but semantically
wrong. Go 1.11 and 1.12 did not validate pseudo-version timestamps, and
`go get` would normalise the fabricated version to the correct one anyway,
so this worked in practice.
Go 1.13 (2019) introduced strict pseudo-version timestamp validation,
rejecting any pseudo-version whose timestamp does not match the actual
commit time. Renovate's fix (31ce47297) was to drop the fabricated
timestamp entirely and write just the bare commit hash, then rely on
`go get` to expand it into a canonical pseudo-version. However, when
the `go get` command or other subsequent operations fail, Renovate may
leave the go.mod file containing dependencies pinned to bare commit hashes.
Why the fix is now safe:
Since the 2024 goproxy datasource changes, both newValue and newDigest are
extracted from the same proxy version string via pseudoVersionToRelease().
They always reference the same commit, so writing newValue directly produces
correct, syntactically valid go.mod output without depending on `go get` to
clean it up afterward.
Signed-off-by: Qixiang Wan <qwan@outlook.com>
* fix(poetry): perform deepmerge to fully enrich projectDep with info from poetryDep
previously, just additional `managerData` from the poetryDep was merged into the projectDep specification
* test(poetry): verify 'invalid-dependency-specification' if constraints specified twice
When running Renovate in a shared environment, such as Mend-hosted
Renovate, if relying on the new functionality for `constraintsFiltering`
checks in the Go datasource, it's possible to get data without
the `constraints` embedded in them.
This is due to us using a cache key that doesn't take into account
whether we are `constraintsFiltering`, which controls whether we look up
the `go` directive for a given Go module.
We can fix this by introducing an additional entry into the cache key
for the `constraintsFiltering` value.
We can make sure that this doesn't modify existing
(non-`constraintsFiltering`) entries, and only add additional keys if
`constraintsFiltering` is set to something other than the default.
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* feat(managers): rename `renovate-config-presets` to `renovate-config`
As part of future changes, we'll extend this support to cover other
functionality. Instead of having multiple managers for this
functionality, we'll use a single, shared, manager for all Renovate
config management.
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* fixup! feat(managers): rename `renovate-config-presets` to `renovate-config`
---------
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* feat(datasource/go): conditionally expose `constraints`
As noted in #42567, some Go modules are constrained by their
compatibility with the source compatibility with the Go language as per
the `go` directive.
Bumping the `go` directive will result in all of a library's consumers
also needing to bump their own `go`, which can be something library
authors will attempt to avoid.
Previously, Renovate users would receive an update from any modules that
have a new release, regardless of whether they will bump the `go`
directive.
With this change, users who use `constraintsFiltering=strict` will have
the filtering applied, so they only receive updates for Go modules in
their current minor version.
Due to the additional API calls needed, this is an opt-in feature, and
only includes the API calls when the using
`constraintsFiltering=strict`.
We can also add an integration test to build our confidence in the
end-to-end flow.
* fix(datasource/go): correctly check Go directive for major version bumps
As we weren't passing around the `pkg` (which included the Semantic
Module Import Path), we were incorrectly lacking major version upgrades.
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* chore: remove `fallback`
As it's unnecessary.
* docs(gomod): rephrase
---------
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
Jamie Tanna