package mailparser

import (
	"bufio"
	"bytes"
	"fmt"
)

var (
	mboxSeparator = []byte("From ")
	mboxHeaderTag = []byte("From: ")
	mboxQuoted    = []byte(">From ")
)

// SplitMbox splits a raw mbox file into individual RFC 2822 message bytes.
// Each message starts with a "From " separator line which is skipped.
//
// Scan errors (e.g. a single line larger than the 10 MB scanner limit) are
// ignored here for backwards compatibility — use SplitMboxErr to detect a
// truncated result instead of silently importing partial data.
func SplitMbox(data []byte) [][]byte {
	messages, _ := SplitMboxErr(data)
	return messages
}

// SplitMboxErr behaves like SplitMbox but also reports scanner failures.
// On error the already-collected messages are still returned so the caller can
// decide between aborting and importing a partial file.
func SplitMboxErr(data []byte) ([][]byte, error) {
	var messages [][]byte
	var current bytes.Buffer

	scanner := bufio.NewScanner(bytes.NewReader(data))
	scanner.Buffer(make([]byte, 1024*1024), 10*1024*1024)

	inMessage := false
	for scanner.Scan() {
		// Bytes() avoids a string allocation per line — mbox files routinely
		// have millions of lines.
		line := scanner.Bytes()
		// mbox separator: line starts with "From " but not "From:" header
		if bytes.HasPrefix(line, mboxSeparator) && !bytes.HasPrefix(line, mboxHeaderTag) {
			if inMessage && current.Len() > 0 {
				// Copy: current.Bytes() aliases the buffer, which is reused
				// after Reset() and would corrupt already-collected messages.
				msg := bytes.TrimSpace(current.Bytes())
				messages = append(messages, append([]byte(nil), msg...))
				current.Reset()
			}
			inMessage = true
			continue
		}
		if inMessage {
			// unescape ">From " lines (mbox quoting)
			if bytes.HasPrefix(line, mboxQuoted) {
				line = line[1:]
			}
			current.Write(line)
			current.WriteByte('\n')
		}
	}
	if inMessage && current.Len() > 0 {
		msg := bytes.TrimSpace(current.Bytes())
		messages = append(messages, append([]byte(nil), msg...))
	}

	if err := scanner.Err(); err != nil {
		return messages, fmt.Errorf("mailparser: mbox scan aborted after %d messages: %w", len(messages), err)
	}
	return messages, nil
}