goget/internal/protocol/gemini/gemini.go

//go:build linux || freebsd
// +build linux freebsd

package gemini

import (
	"bufio"
	"context"
	"crypto/sha256"
	"crypto/tls"
	"crypto/x509"
	"encoding/hex"
	"fmt"
	"io"
	"net"
	"net/url"
	"os"
	"strconv"
	"strings"
	"time"

	"codeberg.org/petrbalvin/goget/internal/core"
	"codeberg.org/petrbalvin/goget/internal/protocol"
)

// Protocol implements gemini:// downloads.
type Protocol struct {
	*protocol.BaseProtocol
	TLSInsecure bool // allow self-signed certs (testing only)
}

// maxGeminiRedirects limits redirect chain depth to prevent stack overflow.
const maxGeminiRedirects = 10

// NewProtocol creates a new Gemini protocol handler.
func NewProtocol() *Protocol {
	return &Protocol{
		BaseProtocol: protocol.NewBaseProtocol(protocol.ProtocolInfo{
			Name:        "GEMINI",
			Scheme:      "gemini",
			DefaultPort: 1965,
			Features:    []string{},
		}),
	}
}

// Download starts a Gemini download. It handles redirect limits internally.
func (p *Protocol) Download(ctx context.Context, req *core.DownloadRequest) (*core.DownloadResult, error) {
	return p.downloadWithRedirects(ctx, req, 0)
}

func (p *Protocol) downloadWithRedirects(ctx context.Context, req *core.DownloadRequest, redirectCount int) (*core.DownloadResult, error) {
	startTime := time.Now()

	host := req.URL.Hostname()
	port := req.URL.Port()
	if port == "" {
		port = "1965"
	}

	addr := net.JoinHostPort(host, port)

	tlsCfg := &tls.Config{
		MinVersion:         tls.VersionTLS12,
		ServerName:         host,
		InsecureSkipVerify: p.TLSInsecure,
	}

	// Certificate pinning: require the leaf certificate to match the
	// caller-provided SHA-256 hash. The check runs after the standard
	// verification chain so a pinned self-signed cert still needs
	// InsecureSkipVerify (or a matching CA) to clear the normal path.
	if req != nil && req.PinnedCertHash != "" {
		expected := strings.ToLower(strings.TrimSpace(req.PinnedCertHash))
		tlsCfg.VerifyPeerCertificate = func(rawCerts [][]byte, _ [][]*x509.Certificate) error {
			for _, rawCert := range rawCerts {
				cert, err := x509.ParseCertificate(rawCert)
				if err != nil {
					continue
				}
				hash := sha256.Sum256(cert.Raw)
				if hex.EncodeToString(hash[:]) == expected {
					return nil
				}
			}
			return fmt.Errorf("certificate pinning failed: no certificate matches sha-256 %s", expected)
		}
	}

	dialer := &net.Dialer{Timeout: 30 * time.Second}
	conn, err := tls.DialWithDialer(dialer, "tcp", addr, tlsCfg)
	if err != nil {
		return nil, core.NewNetworkError("failed to connect to gemini server", err, core.SafeURL(req.URL))
	}
	defer conn.Close()

	conn.SetDeadline(time.Now().Add(30 * time.Second))

	// Send request: URL + CRLF
	requestURL := req.URL.String()
	if _, err := fmt.Fprintf(conn, "%s\r\n", requestURL); err != nil {
		return nil, core.NewNetworkError("failed to send gemini request", err, core.SafeURL(req.URL))
	}

	// Read response header
	reader := bufio.NewReader(conn)
	header, err := reader.ReadString('\n')
	if err != nil {
		return nil, core.NewNetworkError("failed to read gemini response header", err, core.SafeURL(req.URL))
	}
	header = strings.TrimSpace(header)

	// Parse status code (first two chars)
	if len(header) < 2 {
		return nil, core.NewProtocolError("invalid gemini response header", nil, core.SafeURL(req.URL))
	}

	statusCode, _ := strconv.Atoi(header[:2])
	meta := ""
	if len(header) > 3 {
		meta = header[3:]
	}

	switch {
	case statusCode >= 10 && statusCode < 20:
		// 1x INPUT — server requests input (prompt user with meta)
		return nil, core.NewProtocolError(
			fmt.Sprintf("gemini input required: %s", meta), nil, core.SafeURL(req.URL))

	case statusCode >= 30 && statusCode < 40:
		// 3x REDIRECT — follow the redirect
		if meta == "" {
			return nil, core.NewProtocolError("gemini redirect with no target URL", nil, core.SafeURL(req.URL))
		}
		if redirectCount >= maxGeminiRedirects {
			return nil, core.NewProtocolError(
				fmt.Sprintf("gemini redirect chain exceeded %d", maxGeminiRedirects), nil, core.SafeURL(req.URL))
		}
		redirectURL, err := url.Parse(meta)
		if err != nil {
			return nil, core.NewProtocolError(fmt.Sprintf("invalid gemini redirect URL: %s", meta), err, core.SafeURL(req.URL))
		}
		// Resolve relative redirects
		resolved := req.URL.ResolveReference(redirectURL)
		if req.Verbose {
			fmt.Fprintf(os.Stderr, "[gemini] redirecting to: %s\n", resolved.String())
		}
		// Follow redirect with incremented depth
		redirectReq := &core.DownloadRequest{
			URL:            resolved,
			Output:         req.Output,
			Verbose:        req.Verbose,
			Writer:         req.Writer,
			PinnedCertHash: req.PinnedCertHash,
			Ctx:            ctx,
		}
		return p.downloadWithRedirects(ctx, redirectReq, redirectCount+1)

	case statusCode >= 40 && statusCode < 50:
		// 4x TEMPORARY FAILURE
		return nil, core.NewNetworkError(
			fmt.Sprintf("gemini temporary failure: %d %s", statusCode, meta),
			nil, core.SafeURL(req.URL))

	case statusCode >= 50 && statusCode < 60:
		// 5x PERMANENT FAILURE
		return nil, core.NewProtocolError(
			fmt.Sprintf("gemini permanent failure: %d %s", statusCode, meta), nil, core.SafeURL(req.URL))

	case statusCode >= 60 && statusCode < 70:
		// 6x CLIENT CERTIFICATE REQUIRED
		return nil, core.NewProtocolError(
			fmt.Sprintf("gemini client certificate required: %d %s", statusCode, meta), nil, core.SafeURL(req.URL))

	case statusCode < 20 || statusCode >= 30:
		return nil, core.NewProtocolError(
			fmt.Sprintf("gemini server returned %d %s", statusCode, meta), nil, core.SafeURL(req.URL))
	}

	// Category 2x: success — continue to read body
	var totalRead int64
	var writer io.Writer

	if req.Writer != nil {
		writer = req.Writer
	} else {
		writer = io.Discard
	}

	buf := make([]byte, 32*1024)
	for {
		select {
		case <-ctx.Done():
			return nil, ctx.Err()
		default:
		}
		n, err := reader.Read(buf)
		if n > 0 {
			if _, werr := writer.Write(buf[:n]); werr != nil {
				return nil, core.NewFileError("failed to write gemini data", werr)
			}
			totalRead += int64(n)
			if req.ProgressCallback != nil {
				speed := float64(totalRead) / time.Since(startTime).Seconds()
				req.ProgressCallback(totalRead, -1, speed)
			}
		}
		if err == io.EOF {
			break
		}
		if err != nil {
			return nil, core.NewNetworkError("failed to read gemini response body", err, core.SafeURL(req.URL))
		}
	}

	duration := time.Since(startTime)
	speed := float64(totalRead) / duration.Seconds()

	return &core.DownloadResult{
		BytesDownloaded: totalRead,
		TotalSize:       totalRead,
		Duration:        duration,
		Protocol:        "GEMINI",
		IPVersion:       4,
		Speed:           speed,
		OutputPath:      req.Output,
	}, nil
}