Documentation
¶
Overview ¶
Package branchfilter matches branch names against glob patterns.
Patterns follow path.Match semantics. Notably, `*` matches any run of non-separator characters and therefore never crosses `/` — on every OS (branch names are `/`-layered and OS-independent, so the matcher is the path-flavored primitive, not filepath.Match whose separator is `\` on Windows; issue #134):
"release/*" matches "release/v1" but not "release/a/b" "release-*" matches "release-1.0" but not "release/1.0"
There is no `**`; to match additional path levels, spell them out (e.g. "feature/*/*"). Malformed patterns (such as an unclosed "[") are rejected by both constructors with an error wrapping path.ErrBadPattern — a bad pattern never silently matches nothing. The pre-flight covers patterns malformed AFTER a star too (e.g. "release/*["): the matcher validates the whole pattern, not just the chunks a failing name happens to reach.
Pattern lists come in two shapes:
- NewPatterns([]string): each element IS one glob pattern, verbatim — commas are ordinary pattern characters. This is the only way to name a branch whose ref name contains a comma ("release,1.0" is a valid git ref name; git check-ref-format accepts it).
- New("main,release-*"): the convenience string form splits on commas, trims whitespace, and skips empty items. The comma is ALWAYS a separator in this form — there is no escape — so a ref name that contains a comma cannot be expressed here; use NewPatterns.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type BranchFilter ¶
type BranchFilter struct {
// contains filtered or unexported fields
}
BranchFilter matches branch names against a set of comma-separated glob patterns. An empty filter matches all branches.
func New ¶
func New(filterStr string) (*BranchFilter, error)
New creates a BranchFilter from a comma-separated list of glob patterns (e.g. "main,release-*,feature/*"). An empty string matches everything.
The comma is ALWAYS a separator in this form (whitespace around each pattern is trimmed, empty items skipped) — there is no escape mechanism, so a branch name containing a comma can never be expressed here; use NewPatterns for that (issue #54).
Each pattern is validated up front with path.Match; a malformed pattern yields an error wrapping path.ErrBadPattern instead of a filter that silently never matches.
func NewPatterns ¶ added in v0.125.0
func NewPatterns(patterns []string) (*BranchFilter, error)
NewPatterns creates a BranchFilter from explicit patterns: each element IS one glob pattern, verbatim — no comma splitting, no trimming, no escaping rules to remember (issue #54: "release,1.0" is a valid git ref name and is matched literally here). An empty slice matches everything.
Each pattern is validated up front with path.Match; a malformed pattern yields an error wrapping path.ErrBadPattern instead of a filter that silently never matches.
func (*BranchFilter) FilterBranches ¶
func (f *BranchFilter) FilterBranches(branches []string) []string
FilterBranches returns only the branches that match the filter.
func (*BranchFilter) IsEmpty ¶
func (f *BranchFilter) IsEmpty() bool
IsEmpty reports whether the filter has no patterns (matches everything).
func (*BranchFilter) Match ¶
func (f *BranchFilter) Match(branchName string) bool
Match reports whether branchName matches any of the filter's patterns. Returns true (matches everything) when the filter has no patterns.
func (*BranchFilter) Patterns ¶
func (f *BranchFilter) Patterns() []string
Patterns returns the raw glob patterns that were used to create this filter.