// URLs implements a breadth-first search webcrawler based on the
// example given in section 8.6 of The Go Programming Language.
package main

import (
	"context"
	"errors"
	"fmt"
	"log"
	"net/url"
	"os"
	"runtime/pprof"
	"strings"
	"time"

	"github.com/urfave/cli/v3"
)

const shortcodeFilename = "urls.csv"

func main() {
	if os.Getenv("GOEXPERIMENT") != "goroutineleakprofile" {
		log.Fatal("Missing GOEXPERIMENT=goroutineleakprofile environment setting")
	}

	// Setting shorfile helps especially for when we log errors
	// without returning them.
	log.SetFlags(log.LstdFlags | log.Lshortfile)

	cmd := &cli.Command{
		Usage: "A configurable web crawler",
		Commands: []*cli.Command{
			{
				Name:  "get",
				Usage: "Crawl the target",
				MutuallyExclusiveFlags: []cli.MutuallyExclusiveFlags{
					{
						Required: true,
						Flags: [][]cli.Flag{
							{
								&cli.StringFlag{
									Name:  "shortcode",
									Usage: "Specify target using a shortcode",
								},
							},
							{
								&cli.StringFlag{
									Name:  "url",
									Usage: "Specify target using a URL",
								},
							},
						},
					},
				},
				Flags: []cli.Flag{
					&cli.IntFlag{
						Name:    "concurrency",
						Aliases: []string{"c"},
						Usage:   "Allowable number of concurrent URL fetches",

						// Default concurrency setting is 1.
						Value: 1,
					},
					&cli.IntFlag{
						Name:        "maxurls",
						Aliases:     []string{"m"},
						Usage:       "Maximum number of URLs to collect",
						DefaultText: "no limit",
					},
					&cli.IntFlag{
						Name:        "depth",
						Aliases:     []string{"d"},
						Usage:       "Maximum URL depth",
						DefaultText: "no limit",
					},
				},
				Action: func(ctx context.Context, cmd *cli.Command) error {
					if shortcode := cmd.String("shortcode"); shortcode != "" {
						rawURL, err := getURLFromShortcode(shortcodeFilename, shortcode)
						if err != nil {
							return err
						}

						u, err := convertToURL(rawURL)
						if err != nil {
							return err
						}

						fmt.Printf("URL: %s\n", &u)
						classic(u, cmd.Int("concurrency"), cmd.Int("maxurls"), cmd.Int("depth"))
						return nil
					}

					if rawURL := cmd.String("url"); rawURL != "" {
						u, err := convertToURL(rawURL)
						if err != nil {
							return err
						}

						fmt.Printf("URL: %s\n", &u)
						classic(u, cmd.Int("concurrency"), cmd.Int("maxurls"), cmd.Int("depth"))
						return nil
					}

					// We shouldn't reach this
					// code, but let's at least
					// document our intentions.
					return errors.New("url and shortcode arguments should be mutually exclusive")
				},
			},
			{
				Name:  "shortcode",
				Usage: "Configure shortcodes",
				Action: func(ctx context.Context, cmd *cli.Command) error {
					fmt.Println("shortcode!")
					return nil
				},
			},
		},
	}

	if err := cmd.Run(context.Background(), os.Args); err != nil {
		fmt.Printf("\n%v\n", err)
		os.Exit(1)
	}
}

// convertToURL parses the given rawURL into a [url.URL]. If the
// rawURL is missing a scheme, "https://" is prepended before parsing.
//
// Return the parsed URL, along with any error.
func convertToURL(rawURL string) (url.URL, error) {
	if !strings.HasPrefix(rawURL, "http://") && !strings.HasPrefix(rawURL, "https://") {
		rawURL = "https://" + rawURL
		fmt.Printf("start url: %s\n", rawURL)
	}

	u, err := url.Parse(rawURL)
	if err != nil {
		return url.URL{}, fmt.Errorf("can't parse %s: %w", rawURL, err)
	}

	return *u, nil
}

// getLeakProfile runs a leaky program snippet, extracts the goroutine leak profile,
// and writes it to stdout.
func getLeakProfile(leakySnippet func()) {
	prof := pprof.Lookup("goroutineleak")
	defer func() {
		time.Sleep(2 * time.Second)
		var content strings.Builder

		prof.WriteTo(&content, 2)
		// Ignore non leaked goroutines
		leaks := strings.SplitSeq(content.String(), "\n\n")
		for leak := range leaks {
			if strings.Contains(leak, "(leaked)") {
				fmt.Println(leak + "\n")
			}
		}
	}()

	leakySnippet()
}