ReKoPub/internal/web/run.go

package web

import (
	"bytes"
	"context"
	"errors"
	"os/exec"
	"strings"
	"time"
)

// Exit codes of the rekopub command, as documented in its README.
const (
	exitFetch     = 1
	exitNoContent = 3
)

// runResult is the outcome of one run of the rekopub command.
type runResult struct {
	Path string // the saved book, from stdout
	Code int    // exit code; -1 if the command didn't finish
	Err  string // the command's last "error: " line, without the prefix
}

// runFunc runs the rekopub command with args. Tests replace it.
type runFunc func(ctx context.Context, args []string) runResult

// execRunner returns a runFunc that starts the command at cli. Each run is
// its own process, so runs can't affect each other and a stuck one can be
// killed. The command gets an empty environment: no proxy settings or
// anything else from the server leak into it.
func execRunner(cli string) runFunc {
	return func(ctx context.Context, args []string) runResult {
		cmd := exec.CommandContext(ctx, cli, args...)
		cmd.Env = []string{}
		cmd.WaitDelay = 5 * time.Second
		var stdout, stderr bytes.Buffer
		cmd.Stdout = &stdout
		cmd.Stderr = &stderr
		err := cmd.Run()
		res := runResult{Path: strings.TrimSpace(stdout.String()), Err: lastError(stderr.String())}
		var exit *exec.ExitError
		switch {
		case err == nil:
		case errors.As(err, &exit) && exit.Exited():
			res.Code = exit.ExitCode()
		default:
			res.Code = -1
			if res.Err == "" {
				res.Err = err.Error()
			}
		}
		return res
	}
}

// lastError returns the text of the last "error: " line in stderr.
func lastError(stderr string) string {
	var msg string
	for line := range strings.Lines(stderr) {
		if rest, ok := strings.CutPrefix(line, "error: "); ok {
			msg = strings.TrimSpace(rest)
		}
	}
	return msg
}

// explain turns a failed run into a message for the visitor. Only fetch
// errors pass on the command's own text, which describes the page (not
// found, too large, not HTML and so on); the rest are fixed messages.
func explain(ctx context.Context, res runResult) string {
	switch {
	case errors.Is(ctx.Err(), context.DeadlineExceeded):
		return "That page took too long to convert."
	case strings.Contains(res.Err, "local or private") || strings.Contains(res.Err, "-LOCAL"):
		return "That page leads to a local or private network address, which this site doesn't connect to."
	case res.Code == exitFetch && res.Err != "":
		msg := strings.TrimPrefix(res.Err, "couldn't fetch the page: ")
		if r := []rune(msg); len(r) > 300 {
			msg = string(r[:300]) + "…"
		}
		return "Couldn't download the page: " + msg
	case res.Code == exitFetch:
		return "Couldn't download the page."
	case res.Code == exitNoContent:
		return "No readable article was found on that page. Pages that build their content with JavaScript, paywalls and cookie walls can't be converted."
	default:
		return "Something went wrong while converting that page."
	}
}