add apt gestion

Signed-off-by: GnomeZworc <nicolas.boufidjeline@g3e.fr>
This commit is contained in:
GnomeZworc 2026-04-25 19:03:26 +02:00
commit c4776d81dd
Signed by: nicolas.boufideline
GPG key ID: 4406BBBF8845D632
22 changed files with 1465 additions and 163 deletions

View file

@ -0,0 +1,184 @@
package apt
import (
"compress/gzip"
"context"
"fmt"
"io"
"net/http"
"os"
"path/filepath"
"strings"
"time"
)
const (
upstreamInRelease = ".upstream-InRelease"
upstreamPackagesGz = ".upstream-Packages.gz"
)
// Cloner mirrors an APT repository into a local directory.
type Cloner struct {
SourceURL string
DestDir string
Config Config
HTTPClient *http.Client
OnProgress func(file string, bytes int64)
}
// New creates a Cloner with a 30-minute HTTP timeout.
func New(sourceURL, destDir string, cfg Config) *Cloner {
return &Cloner{
SourceURL: strings.TrimRight(sourceURL, "/"),
DestDir: destDir,
Config: cfg,
HTTPClient: &http.Client{Timeout: 30 * time.Minute},
}
}
// Clone performs a full mirror of the configured APT repository.
func (c *Cloner) Clone(ctx context.Context) error {
suite := c.Config.Suite
// 1. Create base directories.
if err := os.MkdirAll(filepath.Join(c.DestDir, "dists", suite), 0o755); err != nil {
return fmt.Errorf("create dists dir: %w", err)
}
if err := os.MkdirAll(filepath.Join(c.DestDir, "pool"), 0o755); err != nil {
return fmt.Errorf("create pool dir: %w", err)
}
// 2. Fetch InRelease and save as .upstream-InRelease.
inReleaseURL := c.SourceURL + "/dists/" + suite + "/InRelease"
inReleaseData, err := fetchBytes(ctx, c.HTTPClient, inReleaseURL)
if err != nil {
return fmt.Errorf("fetch InRelease: %w", err)
}
upstreamIRPath := filepath.Join(c.DestDir, "dists", suite, upstreamInRelease)
if err := atomicWrite(upstreamIRPath, inReleaseData); err != nil {
return fmt.Errorf("save upstream InRelease: %w", err)
}
// 3. Parse InRelease.
rf, err := ParseRelease(inReleaseData)
if err != nil {
return fmt.Errorf("parse InRelease: %w", err)
}
// 4. For each component/arch: find Packages.gz in SHA256 section, download, parse debs.
var allDebs []DebPackage
for _, component := range c.Config.Components {
for _, arch := range c.Config.Architectures {
if err := ctx.Err(); err != nil {
return err
}
pkgs, err := c.fetchComponentPackages(ctx, suite, component, arch, rf)
if err != nil {
return fmt.Errorf("fetch packages for %s/%s: %w", component, arch, err)
}
allDebs = append(allDebs, pkgs...)
}
}
// 5. Download all .deb files.
for _, pkg := range allDebs {
if err := ctx.Err(); err != nil {
return err
}
if err := c.downloadDeb(ctx, pkg); err != nil {
return fmt.Errorf("download %s: %w", pkg.Filename, err)
}
}
// 6. Regenerate local metadata.
if err := RegenerateMetadata(c.DestDir, c.Config); err != nil {
return fmt.Errorf("regenerate metadata: %w", err)
}
return nil
}
// fetchComponentPackages downloads Packages.gz for one component/arch pair,
// saves it as .upstream-Packages.gz, and returns the parsed package list.
func (c *Cloner) fetchComponentPackages(ctx context.Context, suite, component, arch string, rf *ReleaseFile) ([]DebPackage, error) {
wantPath := component + "/binary-" + arch + "/Packages.gz"
// Verify this path is listed in the upstream SHA256 section.
var found bool
for _, e := range rf.SHA256 {
if e.Path == wantPath {
found = true
break
}
}
if !found {
// Not an error — this component/arch may just not exist upstream.
return nil, nil
}
pkgGzURL := c.SourceURL + "/dists/" + suite + "/" + wantPath
data, err := fetchBytes(ctx, c.HTTPClient, pkgGzURL)
if err != nil {
return nil, fmt.Errorf("fetch %s: %w", pkgGzURL, err)
}
// Save as .upstream-Packages.gz so RegenerateMetadata can find it later.
destDir := filepath.Join(c.DestDir, "dists", suite, component, "binary-"+arch)
if err := os.MkdirAll(destDir, 0o755); err != nil {
return nil, err
}
upstreamPath := filepath.Join(destDir, upstreamPackagesGz)
if err := atomicWrite(upstreamPath, data); err != nil {
return nil, fmt.Errorf("save %s: %w", upstreamPath, err)
}
if c.OnProgress != nil {
c.OnProgress(wantPath, int64(len(data)))
}
// Decompress and parse.
gz, err := gzip.NewReader(strings.NewReader(string(data)))
if err != nil {
return nil, fmt.Errorf("open gzip for %s: %w", wantPath, err)
}
defer gz.Close()
raw, err := io.ReadAll(gz)
if err != nil {
return nil, fmt.Errorf("decompress %s: %w", wantPath, err)
}
pkgs, err := ParsePackages(raw)
if err != nil {
return nil, fmt.Errorf("parse Packages for %s: %w", wantPath, err)
}
return pkgs, nil
}
// downloadDeb downloads a single .deb file into the local pool tree.
func (c *Cloner) downloadDeb(ctx context.Context, pkg DebPackage) error {
url := c.SourceURL + "/" + pkg.Filename
dest := filepath.Join(c.DestDir, filepath.FromSlash(pkg.Filename))
// Skip if already on disk.
if _, err := os.Stat(dest); err == nil {
return nil
}
if err := os.MkdirAll(filepath.Dir(dest), 0o755); err != nil {
return err
}
data, err := fetchBytes(ctx, c.HTTPClient, url)
if err != nil {
return err
}
if err := atomicWrite(dest, data); err != nil {
return err
}
if c.OnProgress != nil {
c.OnProgress(pkg.Filename, int64(len(data)))
}
return nil
}

View file

@ -0,0 +1,8 @@
package apt
// Config describes which parts of an APT repository to mirror.
type Config struct {
Suite string `json:"suite"`
Components []string `json:"components"`
Architectures []string `json:"architectures"`
}

View file

@ -0,0 +1,44 @@
package apt
import (
"context"
"fmt"
"io"
"net/http"
"os"
"path/filepath"
)
// fetchBytes performs a GET request and returns the response body.
func fetchBytes(ctx context.Context, httpClient *http.Client, url string) ([]byte, error) {
req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
if err != nil {
return nil, err
}
resp, err := httpClient.Do(req)
if err != nil {
return nil, err
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return nil, fmt.Errorf("HTTP %d for %s", resp.StatusCode, url)
}
return io.ReadAll(resp.Body)
}
// atomicWrite writes data to path via a temporary file and an atomic rename.
func atomicWrite(path string, data []byte) error {
if err := os.MkdirAll(filepath.Dir(path), 0o755); err != nil {
return err
}
tmp := path + ".tmp"
if err := os.WriteFile(tmp, data, 0o644); err != nil {
return err
}
if err := os.Rename(tmp, path); err != nil {
os.Remove(tmp)
return err
}
return nil
}

View file

@ -0,0 +1,148 @@
package apt
import (
"bytes"
"compress/gzip"
"crypto/sha256"
"encoding/hex"
"fmt"
"io"
"os"
"path/filepath"
)
// InitEmptyRepo creates the directory structure for an APT mirror and writes an
// empty Release file so APT clients don't get a 404 before the first clone.
func InitEmptyRepo(localDir string, cfg Config) error {
if err := os.MkdirAll(filepath.Join(localDir, "pool"), 0o755); err != nil {
return err
}
for _, component := range cfg.Components {
for _, arch := range cfg.Architectures {
dir := filepath.Join(localDir, "dists", cfg.Suite, component, "binary-"+arch)
if err := os.MkdirAll(dir, 0o755); err != nil {
return err
}
}
}
releasePath := filepath.Join(localDir, "dists", cfg.Suite, "Release")
if _, err := os.Stat(releasePath); os.IsNotExist(err) {
releaseData := GenerateRelease(cfg.Suite, cfg.Suite, cfg.Components, cfg.Architectures, nil)
if err := atomicWrite(releasePath, releaseData); err != nil {
return fmt.Errorf("write empty Release: %w", err)
}
}
return nil
}
// RegenerateMetadata rebuilds local Packages, Packages.gz, and Release files
// for all configured component/arch pairs, keeping only packages present on disk.
func RegenerateMetadata(localDir string, cfg Config) error {
suite := cfg.Suite
var releaseEntries []ReleaseEntry
for _, component := range cfg.Components {
for _, arch := range cfg.Architectures {
entries, err := regenerateComponentArch(localDir, suite, component, arch)
if err != nil {
return fmt.Errorf("regenerate %s/%s: %w", component, arch, err)
}
releaseEntries = append(releaseEntries, entries...)
}
}
// Write dists/{suite}/Release.
releaseData := GenerateRelease(suite, suite, cfg.Components, cfg.Architectures, releaseEntries)
releasePath := filepath.Join(localDir, "dists", suite, "Release")
if err := os.MkdirAll(filepath.Dir(releasePath), 0o755); err != nil {
return fmt.Errorf("create dists dir: %w", err)
}
if err := atomicWrite(releasePath, releaseData); err != nil {
return fmt.Errorf("write Release: %w", err)
}
return nil
}
// regenerateComponentArch regenerates Packages and Packages.gz for one component/arch
// and returns the two ReleaseEntry values (plain then gz) for the Release file.
func regenerateComponentArch(localDir, suite, component, arch string) ([]ReleaseEntry, error) {
binDir := filepath.Join(localDir, "dists", suite, component, "binary-"+arch)
upstreamGzPath := filepath.Join(binDir, upstreamPackagesGz)
// 1. Read upstream Packages.gz.
gzData, err := os.ReadFile(upstreamGzPath)
if err != nil {
// Not cloned yet — skip silently.
return nil, nil
}
// 2. Decompress.
gz, err := gzip.NewReader(bytes.NewReader(gzData))
if err != nil {
return nil, fmt.Errorf("open gzip %s: %w", upstreamGzPath, err)
}
rawPackages, err := io.ReadAll(gz)
gz.Close()
if err != nil {
return nil, fmt.Errorf("decompress %s: %w", upstreamGzPath, err)
}
// 3. Parse all stanzas.
allPkgs, err := ParsePackages(rawPackages)
if err != nil {
return nil, fmt.Errorf("parse upstream Packages: %w", err)
}
// 4. Filter: keep only packages present on disk.
checker := diskPresenceChecker(localDir)
localPkgs := FilterPackages(allPkgs, checker)
// 5. Re-emit filtered stanzas.
plainBytes := emitPackages(localPkgs)
// 6. Gzip the filtered Packages.
var gzBuf bytes.Buffer
gzw := gzip.NewWriter(&gzBuf)
if _, err := gzw.Write(plainBytes); err != nil {
return nil, fmt.Errorf("gzip write: %w", err)
}
if err := gzw.Close(); err != nil {
return nil, fmt.Errorf("gzip close: %w", err)
}
gzBytes := gzBuf.Bytes()
// 7. Write Packages and Packages.gz.
if err := os.MkdirAll(binDir, 0o755); err != nil {
return nil, err
}
plainPath := filepath.Join(binDir, "Packages")
if err := atomicWrite(plainPath, plainBytes); err != nil {
return nil, fmt.Errorf("write Packages: %w", err)
}
gzPath := filepath.Join(binDir, "Packages.gz")
if err := atomicWrite(gzPath, gzBytes); err != nil {
return nil, fmt.Errorf("write Packages.gz: %w", err)
}
// 8. Compute SHA256 hashes.
plainSum := sha256.Sum256(plainBytes)
gzSum := sha256.Sum256(gzBytes)
// Release paths are relative to dists/{suite}/.
relBase := component + "/binary-" + arch + "/"
return []ReleaseEntry{
{
Hash: hex.EncodeToString(plainSum[:]),
Size: int64(len(plainBytes)),
Path: relBase + "Packages",
},
{
Hash: hex.EncodeToString(gzSum[:]),
Size: int64(len(gzBytes)),
Path: relBase + "Packages.gz",
},
}, nil
}

View file

@ -0,0 +1,128 @@
package apt
import (
"bytes"
"fmt"
"os"
"path/filepath"
"strconv"
)
// DebPackage represents a single stanza from a Debian Packages file.
type DebPackage struct {
Package string
Version string
Architecture string
Filename string // relative path, e.g. pool/main/c/curl/curl_7.88_amd64.deb
SHA256 string
Size int64
Raw []byte // original stanza bytes for faithful re-emission
}
// ParsePackages parses all stanzas from a Debian Packages control file.
// Stanzas are separated by blank lines.
func ParsePackages(data []byte) ([]DebPackage, error) {
// Normalise line endings.
data = bytes.ReplaceAll(data, []byte("\r\n"), []byte("\n"))
var pkgs []DebPackage
stanzas := bytes.Split(data, []byte("\n\n"))
for _, stanza := range stanzas {
stanza = bytes.TrimSpace(stanza)
if len(stanza) == 0 {
continue
}
pkg, err := parseStanza(stanza)
if err != nil {
return nil, fmt.Errorf("parse stanza: %w", err)
}
pkgs = append(pkgs, pkg)
}
return pkgs, nil
}
func parseStanza(stanza []byte) (DebPackage, error) {
var pkg DebPackage
pkg.Raw = stanza
lines := bytes.Split(stanza, []byte("\n"))
for _, rawLine := range lines {
line := bytes.TrimRight(rawLine, " \t")
// Continuation lines (description body etc.) start with a space — skip.
if len(line) == 0 || line[0] == ' ' || line[0] == '\t' {
continue
}
colonIdx := bytes.IndexByte(line, ':')
if colonIdx < 0 {
continue
}
key := string(bytes.TrimSpace(line[:colonIdx]))
value := string(bytes.TrimSpace(line[colonIdx+1:]))
switch key {
case "Package":
pkg.Package = value
case "Version":
pkg.Version = value
case "Architecture":
pkg.Architecture = value
case "Filename":
pkg.Filename = value
case "SHA256":
pkg.SHA256 = value
case "Size":
n, err := strconv.ParseInt(value, 10, 64)
if err != nil {
return DebPackage{}, fmt.Errorf("bad Size value %q: %w", value, err)
}
pkg.Size = n
}
}
if pkg.Package == "" {
return DebPackage{}, fmt.Errorf("stanza missing Package field")
}
return pkg, nil
}
// FilterPackages returns only the packages for which presentOnDisk returns true.
func FilterPackages(pkgs []DebPackage, presentOnDisk func(filename string) bool) []DebPackage {
var out []DebPackage
for _, p := range pkgs {
if presentOnDisk(p.Filename) {
out = append(out, p)
}
}
return out
}
// emitPackages serialises a slice of DebPackage back to Packages-file format,
// using the original Raw bytes for each stanza.
func emitPackages(pkgs []DebPackage) []byte {
var buf bytes.Buffer
for i, pkg := range pkgs {
// Strip trailing whitespace from each line within the stanza.
lines := bytes.Split(pkg.Raw, []byte("\n"))
for j, l := range lines {
lines[j] = bytes.TrimRight(l, " \t")
}
buf.Write(bytes.Join(lines, []byte("\n")))
if i < len(pkgs)-1 {
buf.WriteString("\n\n")
} else {
buf.WriteString("\n")
}
}
return buf.Bytes()
}
// diskPresenceChecker returns a presentOnDisk func rooted at localDir.
func diskPresenceChecker(localDir string) func(string) bool {
return func(filename string) bool {
p := filepath.Join(localDir, filepath.FromSlash(filename))
_, err := os.Stat(p)
return err == nil
}
}

View file

@ -0,0 +1,154 @@
package apt
import (
"bytes"
"fmt"
"strconv"
"strings"
"time"
)
// ReleaseFile represents a parsed Debian Release / InRelease file.
type ReleaseFile struct {
Origin string
Suite string
Codename string
Components string
Architectures string
SHA256 []ReleaseEntry
}
// ReleaseEntry is a single line from the SHA256 section of a Release file.
type ReleaseEntry struct {
Hash string
Size int64
Path string
}
// ParseRelease strips PGP armor (if present) and parses an InRelease or Release file.
func ParseRelease(data []byte) (*ReleaseFile, error) {
text := stripPGPArmor(string(data))
rf := &ReleaseFile{}
inSHA256 := false
for _, rawLine := range strings.Split(text, "\n") {
line := strings.TrimRight(rawLine, " \t\r")
if inSHA256 {
if line == "" || (len(line) > 0 && line[0] != ' ' && line[0] != '\t') {
// A non-indented non-empty line ends the SHA256 block.
if line != "" {
inSHA256 = false
// Fall through to parse this line as a key-value pair.
} else {
inSHA256 = false
continue
}
} else {
// Indented line: " <hash> <size> <path>"
entry, err := parseReleaseEntry(line)
if err == nil {
rf.SHA256 = append(rf.SHA256, entry)
}
continue
}
}
if strings.HasPrefix(line, "SHA256:") {
inSHA256 = true
continue
}
key, value, ok := strings.Cut(line, ":")
if !ok {
continue
}
key = strings.TrimSpace(key)
value = strings.TrimSpace(value)
switch key {
case "Origin":
rf.Origin = value
case "Suite":
rf.Suite = value
case "Codename":
rf.Codename = value
case "Components":
rf.Components = value
case "Architectures":
rf.Architectures = value
}
}
return rf, nil
}
// parseReleaseEntry parses a single indented line from the SHA256 section.
// Format: " <hash> <size> <path>"
func parseReleaseEntry(line string) (ReleaseEntry, error) {
fields := strings.Fields(line)
if len(fields) != 3 {
return ReleaseEntry{}, fmt.Errorf("malformed SHA256 entry: %q", line)
}
size, err := strconv.ParseInt(fields[1], 10, 64)
if err != nil {
return ReleaseEntry{}, fmt.Errorf("malformed size in SHA256 entry: %w", err)
}
return ReleaseEntry{Hash: fields[0], Size: size, Path: fields[2]}, nil
}
// stripPGPArmor removes the PGP signed-message wrapper if present.
// It returns only the signed body (between the header and the signature).
func stripPGPArmor(text string) string {
const beginSigned = "-----BEGIN PGP SIGNED MESSAGE-----"
const beginSig = "-----BEGIN PGP SIGNATURE-----"
if !strings.Contains(text, beginSigned) {
return text
}
// Drop the header lines (Hash: etc.) up to the first blank line.
after, found := strings.CutPrefix(text, beginSigned)
if !found {
// beginSigned not at start — find it
idx := strings.Index(text, beginSigned)
if idx < 0 {
return text
}
after = text[idx+len(beginSigned):]
}
// Skip the armor header lines (e.g. "Hash: SHA512") until the blank line.
blankIdx := strings.Index(after, "\n\n")
if blankIdx < 0 {
return text
}
body := after[blankIdx+2:]
// Cut off at the PGP signature block.
if sigIdx := strings.Index(body, beginSig); sigIdx >= 0 {
body = body[:sigIdx]
}
return strings.TrimRight(body, "\n\r ") + "\n"
}
// GenerateRelease generates a plain unsigned Release file.
func GenerateRelease(suite, codename string, components, architectures []string, entries []ReleaseEntry) []byte {
var buf bytes.Buffer
fmt.Fprintf(&buf, "Origin: ClonePack Mirror\n")
fmt.Fprintf(&buf, "Suite: %s\n", suite)
fmt.Fprintf(&buf, "Codename: %s\n", codename)
fmt.Fprintf(&buf, "Components: %s\n", strings.Join(components, " "))
fmt.Fprintf(&buf, "Architectures: %s\n", strings.Join(architectures, " "))
fmt.Fprintf(&buf, "Date: %s\n", time.Now().UTC().Format("Mon, 02 Jan 2006 15:04:05 UTC"))
fmt.Fprintf(&buf, "SHA256:\n")
for _, e := range entries {
fmt.Fprintf(&buf, " %s %d %s\n", e.Hash, e.Size, e.Path)
}
return buf.Bytes()
}

View file

@ -0,0 +1,145 @@
package apt
import (
"compress/gzip"
"context"
"fmt"
"io"
"net/http"
"os"
"path/filepath"
"strings"
"time"
)
// NewPackage is a package present upstream but absent from the local mirror.
type NewPackage struct {
Package string
Version string
Architecture string
Filename string // relative path, e.g. pool/main/c/curl/curl_7.88_amd64.deb
SHA256 string
Size int64
Component string
}
// Scanner compares an upstream APT repository against a local mirror and
// returns packages that are present upstream but missing on disk.
type Scanner struct {
SourceURL string
LocalDir string
Config Config
HTTPClient *http.Client
}
// NewScanner creates a Scanner with a 5-minute HTTP timeout.
func NewScanner(sourceURL, localDir string, cfg Config) *Scanner {
return &Scanner{
SourceURL: strings.TrimRight(sourceURL, "/"),
LocalDir: localDir,
Config: cfg,
HTTPClient: &http.Client{Timeout: 5 * time.Minute},
}
}
// Scan fetches the upstream package lists and returns packages not on disk.
func (sc *Scanner) Scan(ctx context.Context) ([]NewPackage, error) {
suite := sc.Config.Suite
// 1. Fetch InRelease.
inReleaseURL := sc.SourceURL + "/dists/" + suite + "/InRelease"
inReleaseData, err := fetchBytes(ctx, sc.HTTPClient, inReleaseURL)
if err != nil {
return nil, fmt.Errorf("fetch InRelease: %w", err)
}
rf, err := ParseRelease(inReleaseData)
if err != nil {
return nil, fmt.Errorf("parse InRelease: %w", err)
}
// 2 & 3. Fetch Packages.gz for each component/arch, save it, then filter.
var missing []NewPackage
for _, component := range sc.Config.Components {
for _, arch := range sc.Config.Architectures {
if err := ctx.Err(); err != nil {
return nil, err
}
pkgs, err := sc.fetchAndSavePackages(ctx, suite, component, arch, rf)
if err != nil {
return nil, fmt.Errorf("fetch packages for %s/%s: %w", component, arch, err)
}
// 4. Filter: only return packages whose Filename is NOT on disk.
checker := diskPresenceChecker(sc.LocalDir)
for _, pkg := range pkgs {
if !checker(pkg.Filename) {
missing = append(missing, NewPackage{
Package: pkg.Package,
Version: pkg.Version,
Architecture: pkg.Architecture,
Filename: pkg.Filename,
SHA256: pkg.SHA256,
Size: pkg.Size,
Component: component,
})
}
}
}
}
return missing, nil
}
// fetchAndSavePackages fetches Packages.gz for a component/arch, saves it as
// .upstream-Packages.gz (so RegenerateMetadata can use it), and returns parsed packages.
func (sc *Scanner) fetchAndSavePackages(ctx context.Context, suite, component, arch string, rf *ReleaseFile) ([]DebPackage, error) {
wantPath := component + "/binary-" + arch + "/Packages.gz"
// Check that this path is listed in the upstream SHA256 section.
var found bool
for _, e := range rf.SHA256 {
if e.Path == wantPath {
found = true
break
}
}
if !found {
return nil, nil
}
pkgGzURL := sc.SourceURL + "/dists/" + suite + "/" + wantPath
data, err := fetchBytes(ctx, sc.HTTPClient, pkgGzURL)
if err != nil {
return nil, fmt.Errorf("fetch %s: %w", pkgGzURL, err)
}
// Save as .upstream-Packages.gz.
destDir := filepath.Join(sc.LocalDir, "dists", suite, component, "binary-"+arch)
if err := os.MkdirAll(destDir, 0o755); err != nil {
return nil, err
}
upstreamPath := filepath.Join(destDir, upstreamPackagesGz)
if err := atomicWrite(upstreamPath, data); err != nil {
return nil, fmt.Errorf("save %s: %w", upstreamPath, err)
}
// Decompress and parse.
gz, err := gzip.NewReader(strings.NewReader(string(data)))
if err != nil {
return nil, fmt.Errorf("open gzip: %w", err)
}
defer gz.Close()
raw, err := io.ReadAll(gz)
if err != nil {
return nil, fmt.Errorf("decompress: %w", err)
}
pkgs, err := ParsePackages(raw)
if err != nil {
return nil, fmt.Errorf("parse Packages: %w", err)
}
return pkgs, nil
}