* feat(manager/mise): support gitlab backend
Maps the mise `gitlab` backend (e.g. `gitlab:gitlab-org/cli`) to the gitlab-releases datasource, mirroring the github backend including version_prefix support.
* refactor(manager/mise): use optional instead of Nullish for tool options
TOML has no null type, so MiseToolOptions values are only ever a string or absent. .optional() models that exactly; Nullish only adds value for sources that can emit explicit null.
* fix(manager/mise): use RegExp.escape in gitlab backend
Main switched `createGithubToolConfig` from the `escapeRegExp` util to
`RegExp.escape` and dropped the import, which left the newly added
`createGitlabToolConfig` referencing an undefined symbol after the merge.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Michael Kriese <michael.kriese@mend.io>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat(data): automatic update of static data
* test(manager/mise): drop stale cargo-backend registry test
The mise registry now lists a github mirror for magika alongside its
cargo backend, so a bare `magika` entry resolves via the
github-prioritised branch (already covered by the
'prioritising the github backend over others' test) instead of cargo.
No tool in the current registry data reaches the cargo case through
the unprefixed lookup path anymore, and the cargo backend itself
stays covered by the explicit `cargo:` prefix test.
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: secustor <sebastian@poxhofer.at>
The `rust` tool was resolved via the `github-tags` datasource on
`rust-lang/rust` with the default (semver) versioning. That meant the
release channel values rustup actually accepts were not understood.
Switch all three managers to the `rust-version` datasource, which
declares `rust-release-channel` as its `defaultVersioning`, matching the
`rust-toolchain` manager and `lib/util/exec/containerbase.ts`.
`packageName` becomes `rust` rather than the `rust-lang/rust` owner/repo
slug, which is meaningless for this datasource. It is set rather than
omitted because mise would otherwise fall back to the backend-prefixed
depName, yielding `core:rust` / `asdf:rust` as the package name.
No explicit `versioning` is set: the datasource default already applies,
and the convention in these files is to set `versioning` only to
override a datasource default.
Behaviour change: `nightly-<date>` pins were previously skipped as
`invalid-value` and now update. `stable`, `beta` and `nightly` are now
recognised as ranges rather than flagged invalid, but still only resolve
to a concrete version under `rangeStrategy: pin`.
* feat(manager/mise): set depType for tools and task tools
Allow packageRules to match mise deps via matchDepTypes:
- tools for top-level [tools]
- task-tools for tasks.*.tools
* test(manager/mise): override RUST_197 depType in task extract cases
* Make the task tool dynamic
* Restore formatting
* Restore another formatting
* 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>
* feat(mise): automagically support all tools through Mise Registry
As noted in #42250, we should try and auto-support tools from Mise where
possible.
Similar to other areas of the codebase, we should use the "static data"
pattern, and periodically sync the `mise registry --json` output with
`lib/data/mise-registry.json`, and then auto-detect the tools that we
support through the known backends the tools support.
This takes lower precedence than existing in-code tool definitions, so
we can still provide an "override" if we can support lookups or version
extraction more appropriately.
Because we now have many more tools we (mostly) support out-of-the-box,
we need to provide an improved documentation page, and:
- only link to tools we have an explicit tools URL for
- note where a backend may not be supported
To automate the updating of this data, we:
- make sure we have `mise` in GitHub Actions, before we run
- don't pin the `mise` version to make sure that we're always seeing the
latest supported tools
- make sure the commits generated are now `feat(...)`, as they introduce
functionality changes (in the case new tool(s) are supported)
This improves support from ~170 tool short names to just over 1000.
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* refactor: use Typescript
As we can now, we should.
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* refactor(tools/static-data): use Zod schemas for parsing
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* refactor(mise): use Zod for types
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
* chore(mise): prefer the GitHub backend over other sources
As we generally prefer it as a project, we should automagically use that
when looking at a mise registry entry.
With help from GPT-4.1 to write the `sortBackend` function.
Co-authored-by: gpt-4.1 <jamie.tanna+github-copilot@mend.io>
* test: ignore else coverage
* docs(mise): preference of `github` backend
* docs(mise): auto-link autogenerated tools
* perf(mise): sort registry only once
* refactor(mise): transform when parsing
Co-authored-by: Sebastian Poxhofer <secustor@users.noreply.github.com>
* refactor: use `execa`
* perf(mise): read the Mise registry once
* perf(mise): pre-transform
---------
Co-authored-by: Claude Sonnet 4.6 <jamie.tanna+claude-code@mend.io>
Co-authored-by: gpt-4.1 <jamie.tanna+github-copilot@mend.io>
Co-authored-by: Sebastian Poxhofer <secustor@users.noreply.github.com>
* fix: preserve vendor prefixes in mise manager version updates
* test: cover kafka fallback path for 100% coverage
* test(mise): add failing tests for short java version versioning
Bare major versions like java = "21" or temurin-21 should use
semver-partial versioning to prevent invalid update proposals.
Full versions like "21.0.2" should keep the default versioning.
* fix(mise): use semver-partial versioning for short java versions
Short java versions (bare major like "21" or major.minor like "21.0")
now use semver-partial versioning, which treats them as ranges and
preserves precision in updates (e.g. "21" → "22" on major bump).
This replaces the replaceString/autoReplaceStringTemplate approach
with a versioning-based fix as suggested in discussion #41159.
Full versions (e.g. "21.0.2") keep the default semver-coerced
versioning unchanged.
* test(mise): remove unrelated kafka test
* refactor(mise): inline regex in shortJavaVersioning
* refactor(mise): return empty object instead of undefined
* refactor: use it.each to reduce duplication
* refactor: hardcode semver-partial in it.each test
* feat(mise): add support for github backend
* test(mise): remove redundant github backend test
* refactor(mise): flatten nested conditionals in github backend
* refactor(mise): only set extractVersion for custom version_prefix
* refactor(mise): use escapeRegExp and simplify extractVersion spread
Pieter Willekens