.golangci.yml

---
linters-settings:
  gocritic:
    enabled-tags:
      - diagnostic
      - opinionated
      - performance
      - style
    disabled-checks:
      - ifElseChain
      - unnamedResult
  gocyclo:
    min-complexity: 15
  godot:
    exclude:
      - ^\s*\+
  goheader:
    template: |-
      SPDX-License-Identifier: Apache-2.0

      Copyright Contributors to the Submariner project.

      Licensed under the Apache License, Version 2.0 (the "License");
      you may not use this file except in compliance with the License.
      You may obtain a copy of the License at

      http://www.apache.org/licenses/LICENSE-2.0

      Unless required by applicable law or agreed to in writing, software
      distributed under the License is distributed on an "AS IS" BASIS,
      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
      See the License for the specific language governing permissions and
      limitations under the License.
  govet:
    enable:
      - fieldalignment
  lll:
    line-length: 140
  wrapcheck:
    ignoreSigs:
      - .Error(
      - .Errorf(
      - errors.New(
      - errors.Unwrap(
      - .Wrap(
      - .Wrapf(
      - .WithMessage(
      - .WithMessagef(
      - .WithStack(
  wsl:
    # Separating explicit var declarations by blank lines seems excessive.
    allow-cuddle-declarations: true
linters:
  disable-all: true
  enable:
    - asasalint
    - asciicheck
    - bidichk
    - bodyclose
    # - canonicalheader # This is a slow linter and we don't use the net/http.Header API
    - containedctx
    - contextcheck
    - copyloopvar
    # - cyclop # This is equivalent to gocyclo
    - decorder
    # - depguard # depguard now denies by default, it should only be enabled if we actually use it
    - dogsled
    - dupl
    - durationcheck
    - errcheck
    - errchkjson
    - errorlint
    - errname
    - exhaustive
    # - exhaustruct # This is too cumbersome as it requires all string, int, pointer et al fields to be initialized even when the
    #   type's default suffices, which is most of the time
    - fatcontext
    # - forbidigo # We don't forbid any statements
    # - forcetypeassert # There are many unchecked type assertions that would be the result of a programming error so the
    #                     reasonable recourse would be to panic anyway if checked so this doesn't seem useful
    # - funlen # gocyclo is enabled which is generally a better metric than simply LOC.
    - gci
    - ginkgolinter
    - gocheckcompilerdirectives
    # - gochecknoglobals # We don't want to forbid global variable constants
    # - gochecknoinits # We use init functions for valid reasons
    # - gochecksumtype # The usefulness is very narrow
    - gocognit
    - goconst
    - gocritic
    - gocyclo
    - err113
    - godot
    # - godox #  Let's not forbid inline TODOs, FIXMEs et al
    - gofmt
    - gofumpt
    - goheader
    - goimports
    # - gomoddirectives # We don't want to forbid the 'replace' directive
    # - gomodguard # We don't block any modules
    # - goprintffuncname # This doesn't seem useful at all
    # - gosmopolitan # This is related to internationalization which is not a concern for us
    - gosec
    - gosimple
    - govet
    - grouper
    - iface
    - importas
    - inamedparam
    - ineffassign
    # - interfacebloat # We track complexity elsewhere
    - intrange
    # - ireturn # The argument to always "Return Concrete Types" doesn't seem compelling. It is perfectly valid to return
    #   an interface to avoid exposing the entire underlying struct
    - lll
    - loggercheck
    - maintidx
    - makezero
    - mirror
    - misspell
    # - mnd # It doesn't seem useful in general to enforce constants for all numeric values
    - musttag
    - nakedret
    # - nestif # This calculates cognitive complexity but we're doing that elsewhere
    - nilerr
    - nilnil
    # - nlreturn # This is reasonable with a block-size of 2 but setting it above isn't honored
    # - noctx # We don't send HTTP requests
    - nolintlint
    - nonamedreturns
    # - nosprintfhostport # The use of this is very narrow
    # - paralleltest # Not relevant for Ginkgo UTs
    - perfsprint
    - prealloc
    - predeclared
    - promlinter
    # - protogetter # We don't use protobuf
    - reassign
    - recvcheck
    - revive
    # - rowserrcheck # We don't use SQL
    # - sloglint # We don't use log/slog
    # - spancheck # We don't use OpenTelemetry/OpenCensus
    # - sqlclosecheck # We don't use SQL
    - staticcheck
    - stylecheck
    - tagalign
    # - tagliatelle # Inconsistent with stylecheck and not as good
    # - tenv # Not relevant for our Ginkgo UTs
    # - testableexamples # We don't need this
    # - testifylint # We don't use testify
    - testpackage
    # - thelper # Not relevant for our Ginkgo UTs
    # - tparallel # Not relevant for our Ginkgo UTs
    - typecheck
    - unconvert
    - unparam
    - unused
    - usestdlibvars
    # - varnamelen # It doesn't seem necessary to enforce a minimum variable name length
    - wastedassign
    - whitespace
    - wrapcheck
    - wsl
    # - zerologlint # We use zerolog indirectly so this isn't needed
issues:
  exclude-rules:
    # Allow dot-imports for Gomega BDD directives per idiomatic Gomega
    - linters:
        - revive
        - stylecheck
      text: "dot imports"
      source: "gomega"

    # Allow dot-imports for Ginkgo BDD directives per idiomatic Ginkgo
    - linters:
        - revive
        - stylecheck
      text: "dot imports"
      source: "ginkgo"

    # Ignore long line and variable name non-compliance warnings in auto-generated file
    - linters:
        - lll
        - stylecheck
        - revive
      path: "pkg/embeddedyamls/yamls.go"

    # BrokerK8sApiServer parameter is used by other projects, like ACM,
    # so not changing it to BrokerK8sAPIServer as suggested by stylecheck
    - linters:
        - revive
        - stylecheck
      text: "struct field BrokerK8sApiServer"

    # Ignore pointer bytes in struct alignment tests (this is a very
    # minor optimisation)
    - linters:
        - govet
      text: "pointer bytes could be"

    # Full text of the error is "do not define dynamic errors, use wrapped static errors instead". See
    # https://github.com/Djarvur/go-err113/issues/10 for an interesting discussion of this error. While there are cases
    # where wrapped sentinel errors are useful, it seems a bit pedantic to force that pattern in all cases.
    - linters:
        - err113
      text: "do not define dynamic errors"

    # Ignore certain linters for test files
    - path: _test\.go|test/|fake/
      linters:
        - errname
        - gochecknoinits
        - err113
        - wrapcheck

    # Ignore header linting for files copied from the Operator SDK
    - path: pkg/metrics/service-monitor.go
      linters:
        - goheader

    # Ignore duplicate strings in generated code
    - path: pkg/embeddedyamls/yamls.go
      linters:
        - goconst