upd/asset.go

package main

import (
	"fmt"
	"os"
	"os/exec"
	"regexp"
	"sort"
	"strings"
	"sync"
)

// Aliases commonly used in release asset names, including Rust target triples.
func osAlias(o string) []string {
	switch o {
	case "darwin":
		return []string{"darwin", "macos", "osx", "mac", "apple"}
	case "windows":
		return []string{"windows", "win"}
	}
	return []string{o}
}

func archAlias(a string) []string {
	switch a {
	case "amd64":
		return []string{"amd64", "x86_64", "x64", "64bit"}
	case "arm64":
		return []string{"arm64", "aarch64"}
	case "386":
		return []string{"386", "i386", "i686", "x86", "32bit"}
	case "arm":
		return []string{"arm", "armv7", "armv6", "armhf"}
	case "riscv64":
		return []string{"riscv64", "riscv64gc"}
	}
	return []string{a}
}

var (
	// Checksums, signatures and OS packages are not plain binaries.
	notBinaryRe = regexp.MustCompile(`\.(sha\d*|sha256sum|md5|asc|sig|pem|sbom|json|txt|deb|rpm|apk|dmg|pkg|msi|snap|flatpak|appimage)$`)
	// Cross-compile targets whose triples contain a host OS token:
	// aarch64-linux-android is Android, not Linux.
	foreignRe = regexp.MustCompile(`(^|[^a-z0-9])(android\w*|ios|wasi|wasm\w*|emscripten)([^a-z0-9]|$)`)
	archiveRe = regexp.MustCompile(`\.(tar\.gz|tgz|zip|tar\.xz|tar\.bz2|tar\.zst|tzst)$`)
	noiseRe   = regexp.MustCompile(`debug|symbols|static-pie|profile`)
	tarballRe = regexp.MustCompile(`\.(tar\.gz|tgz|tar\.bz2|tbz|tar\.xz|txz|tar\.zst|tzst|tar)$`)
	singleRe  = regexp.MustCompile(`\.(gz|bz2|xz|zst|lz4|lzma)$`)
	refuseRe  = regexp.MustCompile(`\.(rar|dmg|pkg|msi|deb|rpm)$`)
	widerArm  = regexp.MustCompile(`arm64|aarch64`)
	wider386  = regexp.MustCompile(`x86[_-]?64|amd64`)

	win64Re = regexp.MustCompile(`(^|[^a-z0-9])win64($|[^a-z0-9])`)
	win32Re = regexp.MustCompile(`(^|[^a-z0-9])win32($|[^a-z0-9])`)
)

// Compound tokens that name OS and architecture in one word. Rewritten before
// matching so the normal token rules apply. "win32" is Node-speak for Windows
// in general (pnpm-win32-x64.zip), so it contributes no architecture.
func normalizeTokens(n string) string {
	n = win64Re.ReplaceAllString(n, "${1}windows-amd64${2}")
	n = win32Re.ReplaceAllString(n, "${1}windows${2}")
	return n
}

// tokenRe matches any of the words as a token, i.e. not glued to letters or
// digits. RE2 has no look-around, so the boundaries are ordinary characters.
func tokenRe(words []string) *regexp.Regexp {
	quoted := make([]string, len(words))
	for i, w := range words {
		quoted[i] = regexp.QuoteMeta(w)
	}
	return regexp.MustCompile(`(^|[^a-z0-9])(` + strings.Join(quoted, "|") + `)([^a-z0-9]|$)`)
}

// Can this machine turn the asset into a binary? tar, zip, gzip and bzip2 are
// handled in-process; the rest needs a helper on $PATH.
func unpackable(n string) bool {
	n = strings.ToLower(n)
	switch {
	case regexp.MustCompile(`\.(tar\.gz|tgz|tar\.bz2|tbz|tar|zip)$`).MatchString(n):
		return true
	case regexp.MustCompile(`\.(tar\.xz|txz)$`).MatchString(n):
		return decompressor("xz") != nil
	case regexp.MustCompile(`\.(tar\.zst|tzst)$`).MatchString(n):
		return decompressor("zst") != nil
	case strings.HasSuffix(n, ".7z"):
		return sevenZip() != ""
	case singleRe.MatchString(n):
		return decompressor(singleRe.FindStringSubmatch(n)[1]) != nil
	case refuseRe.MatchString(n):
		return false
	}
	return true // bare binary or an unversioned name
}

// Glibc vs musl builds: pick what this system actually runs.
var muslOnce = sync.OnceValue(func() bool {
	if _, err := os.Stat("/etc/alpine-release"); err == nil {
		return true
	}
	ldd, err := exec.LookPath("ldd")
	if err != nil {
		return false
	}
	out, _ := exec.Command(ldd, "--version").CombinedOutput()
	return strings.Contains(strings.ToLower(string(out)), "musl")
})

type candidate struct {
	score  int
	pref   int
	length int
	asset  *Asset
}

// pickAsset mirrors the Perl scoring one to one - the corpus test in
// asset_test.go checks that the two keep agreeing.
func pickAsset(assets []Asset, bin, wantOS, wantArch string, spec *Spec) (*Asset, error) {
	if spec.Asset != "" {
		for i := range assets {
			if assets[i].Name == spec.Asset {
				return &assets[i], nil
			}
		}
		return nil, fmt.Errorf("asset %q is not part of the release", spec.Asset)
	}
	if spec.Pattern != "" {
		re, err := regexp.Compile(spec.Pattern)
		if err != nil {
			return nil, fmt.Errorf("invalid --pattern: %w", err)
		}
		for i := range assets {
			if re.MatchString(assets[i].Name) {
				return &assets[i], nil
			}
		}
		return nil, nil
	}

	arches := archAlias(wantArch)
	// Universal macOS builds serve both architectures.
	if wantOS == "darwin" {
		arches = append(arches, "universal", "universal2")
	}
	osRe, archRe := tokenRe(osAlias(wantOS)), tokenRe(arches)

	var cands []candidate
	var rejected []string
	for i := range assets {
		as := &assets[i]
		n := strings.ToLower(as.Name)
		m := normalizeTokens(n)

		switch {
		case notBinaryRe.MatchString(n):
			rejected = append(rejected, n+" (not a binary)")
			continue
		case foreignRe.MatchString(m):
			rejected = append(rejected, n+" (foreign target)")
			continue
		case !osRe.MatchString(m):
			rejected = append(rejected, n+" (os)")
			continue
		case !archRe.MatchString(m):
			rejected = append(rejected, n+" (arch)")
			continue
		// Keep 32-bit "arm" off "arm64" and "386" off "x86_64".
		case wantArch == "arm" && widerArm.MatchString(m),
			wantArch == "386" && wider386.MatchString(m),
			wantArch == "amd64" && widerArm.MatchString(m):
			rejected = append(rejected, n+" (wider arch)")
			continue
		}

		score := 0
		if bin != "" && strings.HasPrefix(n, strings.ToLower(bin)) {
			score += 10 // named after the binary
		}
		if archiveRe.MatchString(n) || !strings.Contains(n, ".") {
			score += 3
		}
		if noiseRe.MatchString(n) {
			score -= 5
		}
		// Only avoid formats this machine has no tool for - restic, for one,
		// ships nothing but .bz2, and that is handled in-process.
		if !unpackable(n) {
			score -= 8
		}

		// Toolchain preference is a tie-break, not a penalty: a project that
		// ships musl only (or mingw only) must not be downranked for it.
		pref := 0
		switch {
		case wantOS == "linux" && (strings.Contains(n, "musl") || strings.Contains(n, "gnu")):
			libc := "gnu"
			if muslOnce() {
				libc = "musl"
			}
			if !strings.Contains(n, libc) {
				pref = 1
			}
		case wantOS == "windows" && (strings.Contains(n, "msvc") || strings.Contains(n, "gnu")):
			if !strings.Contains(n, "msvc") { // msvc is the normal Windows build
				pref = 1
			}
		}
		cands = append(cands, candidate{score: score, pref: pref, length: len(n), asset: as})
	}

	// Best score, then preferred toolchain, then the shortest name (which
	// avoids special variants like -baseline or -static).
	sort.SliceStable(cands, func(i, j int) bool {
		a, b := cands[i], cands[j]
		if a.score != b.score {
			return a.score > b.score
		}
		if a.pref != b.pref {
			return a.pref < b.pref
		}
		return a.length < b.length
	})

	if opt.verbose {
		for _, c := range cands {
			verbose("candidate score=%-3d pref=%d %s", c.score, c.pref, c.asset.Name)
		}
		for _, r := range rejected {
			verbose("rejected: %s", r)
		}
	}
	if len(cands) == 0 {
		return nil, nil
	}
	return cands[0].asset, nil
}