ConfigKit/source.go

package configkit

import (
	"bytes"
	"errors"
	"io"
	"io/fs"
	"os"
	"path/filepath"
	"reflect"
	"sync/atomic"
)

const (
	maxSourceBytes = 8 << 20
	maxTotalBytes  = 64 << 20
)

// Source is an opaque configuration input. Construct sources with File, Bytes,
// Reader, or NoSource; the zero value is invalid.
type Source struct {
	kind      sourceKind
	path      string
	label     string
	data      []byte
	reader    *readerSource
	createErr sourceCreationError
}

type sourceKind uint8

const (
	sourceInvalid sourceKind = iota
	sourceFile
	sourceBytes
	sourceReader
	sourceNone
)

type sourceCreationError uint8

const (
	creationValid sourceCreationError = iota
	creationInvalidPath
	creationInvalidLabel
	creationInvalidReader
	creationTooLarge
)

type readerSource struct {
	input io.Reader
	used  atomic.Bool
}

// File returns a source that reopens path for every load attempt.
func File(path string) Source {
	if path == "" {
		return Source{kind: sourceFile, createErr: creationInvalidPath}
	}
	return Source{kind: sourceFile, path: path, label: path}
}

// Bytes snapshots data immediately. Label must be nonempty and safe to display;
// constructor problems are reported when the source is consumed.
func Bytes(label string, data []byte) Source {
	source := Source{kind: sourceBytes, label: label}
	if label == "" {
		source.createErr = creationInvalidLabel
		return source
	}
	if len(data) > maxSourceBytes {
		source.createErr = creationTooLarge
		return source
	}
	source.data = bytes.Clone(data)
	return source
}

// Reader returns a single-use source. ConfigKit consumes input but never closes it.
// Copies of the returned Source share the same single-use claim.
func Reader(label string, input io.Reader) Source {
	source := Source{kind: sourceReader, label: label}
	if label == "" {
		source.createErr = creationInvalidLabel
		return source
	}
	if nilReader(input) {
		source.createErr = creationInvalidReader
		return source
	}
	source.reader = &readerSource{input: input}
	return source
}

// NoSource explicitly selects a defaults-only input. It creates no source record.
func NoSource() Source {
	return Source{kind: sourceNone}
}

type sourceLimits struct {
	perSource int64
	remaining int64
}

func defaultSourceLimits() sourceLimits {
	return sourceLimits{perSource: maxSourceBytes, remaining: maxTotalBytes}
}

type sourceRead struct {
	data []byte
	info *SourceInfo
}

func readSource(source Source, sourceID string, limits sourceLimits) (sourceRead, []Diagnostic) {
	if diagnostic := sourceConstructorDiagnostic(source); diagnostic != nil {
		return sourceRead{}, []Diagnostic{*diagnostic}
	}
	if source.kind == sourceNone {
		return sourceRead{}, nil
	}

	limit := min(limits.perSource, limits.remaining)
	if limit < 0 {
		limit = 0
	}
	info := &SourceInfo{ID: sourceID, Label: source.label, Layer: LayerBase}

	var (
		input io.Reader
		close func() error
	)
	switch source.kind {
	case sourceFile:
		absolute, err := filepath.Abs(source.path)
		if err != nil {
			return sourceRead{info: info}, []Diagnostic{sourceDiagnostic(CodeSourceRead, sourceID, "source path could not be resolved")}
		}
		info.Path = filepath.Clean(absolute)
		file, err := os.Open(source.path)
		if err != nil {
			return sourceRead{info: info}, []Diagnostic{classifySourceError(err, sourceID)}
		}
		close = file.Close
		stat, err := file.Stat()
		if err != nil {
			_ = file.Close()
			return sourceRead{info: info}, []Diagnostic{classifySourceError(err, sourceID)}
		}
		if !stat.Mode().IsRegular() {
			_ = file.Close()
			return sourceRead{info: info}, []Diagnostic{sourceDiagnostic(CodeSourceNotRegular, sourceID, "source is not a regular file")}
		}
		canonical, err := filepath.EvalSymlinks(info.Path)
		if err != nil {
			_ = file.Close()
			return sourceRead{info: info}, []Diagnostic{classifySourceError(err, sourceID)}
		}
		info.Path = canonical
		if stat.Size() > limit {
			_ = file.Close()
			return sourceRead{info: info}, []Diagnostic{sourceDiagnostic(CodeSourceLimitExceeded, sourceID, "source exceeds the configured byte limit")}
		}
		input = file
	case sourceBytes:
		input = bytes.NewReader(source.data)
	case sourceReader:
		if !source.reader.used.CompareAndSwap(false, true) {
			return sourceRead{info: info}, []Diagnostic{sourceDiagnostic(CodeSourceReaderUsed, sourceID, "reader source has already been consumed")}
		}
		input = source.reader.input
	default:
		return sourceRead{info: info}, []Diagnostic{sourceDiagnostic(CodeSourceInvalid, sourceID, "source is invalid")}
	}
	if close != nil {
		defer func() { _ = close() }()
	}

	data, err := io.ReadAll(io.LimitReader(input, limit+1))
	if err != nil {
		return sourceRead{info: info}, []Diagnostic{classifySourceError(err, sourceID)}
	}
	if int64(len(data)) > limit {
		return sourceRead{info: info}, []Diagnostic{sourceDiagnostic(CodeSourceLimitExceeded, sourceID, "source exceeds the configured byte limit")}
	}
	return sourceRead{data: bytes.Clone(data), info: info}, nil
}

func sourceConstructorDiagnostic(source Source) *Diagnostic {
	var message string
	code := CodeSourceInvalid
	switch source.createErr {
	case creationValid:
		if source.kind != sourceInvalid {
			return nil
		}
		message = "source is invalid"
	case creationInvalidPath:
		message = "source path must not be empty"
	case creationInvalidLabel:
		message = "source label must not be empty"
	case creationInvalidReader:
		message = "source reader must not be nil"
	case creationTooLarge:
		code = CodeSourceLimitExceeded
		message = "source exceeds the configured byte limit"
	}
	diagnostic := sourceDiagnostic(code, "", message)
	return &diagnostic
}

func classifySourceError(err error, sourceID string) Diagnostic {
	switch {
	case errors.Is(err, fs.ErrNotExist):
		return sourceDiagnostic(CodeSourceMissing, sourceID, "source does not exist")
	case errors.Is(err, fs.ErrPermission):
		return sourceDiagnostic(CodeSourcePermission, sourceID, "source cannot be accessed")
	default:
		return sourceDiagnostic(CodeSourceRead, sourceID, "source could not be read")
	}
}

func sourceDiagnostic(code Code, sourceID, message string) Diagnostic {
	diagnostic := Diagnostic{Code: code, Severity: SeverityError, Message: message}
	if sourceID != "" {
		diagnostic.Primary = &Location{SourceID: sourceID}
	}
	return diagnostic
}

func nilReader(input io.Reader) bool {
	if input == nil {
		return true
	}
	value := reflect.ValueOf(input)
	switch value.Kind() {
	case reflect.Chan, reflect.Func, reflect.Interface, reflect.Map, reflect.Pointer, reflect.Slice:
		return value.IsNil()
	default:
		return false
	}
}