tka: stable text representation of AUMHash

This makes debugging easier, you can pass an AUMHash to a printf and get
a string that is easy to debug.

Also rearrange how directories/files work in the FS store: use the first
two characters of the string representation as the prefix directory, and
use the entire AUMHash string as the file name. This is again to aid
debugging: you can `ls` a directory and line up what prints out easily
with what you get from a printf in debug code.

Signed-off-by: David Crawshaw <crawshaw@tailscale.com>

tka/aum.go

@@ -7,6 +7,7 @@ package tka
 import (
 	"bytes"
 	"crypto/ed25519"
+	"encoding/base32"
 	"errors"
 	"fmt"
 
@@ -18,6 +19,32 @@ import (
 // AUMHash represents the BLAKE2s digest of an Authority Update Message (AUM).
 type AUMHash [blake2s.Size]byte
 
+var base32StdNoPad = base32.StdEncoding.WithPadding(base32.NoPadding)
+
+// String returns the AUMHash encoded as base32.
+// This is suitable for use as a filename, and for storing in text-preferred media.
+func (h AUMHash) String() string {
+	return base32StdNoPad.EncodeToString(h[:])
+}
+
+// UnmarshalText implements encoding.TextUnmarshaler.
+func (h *AUMHash) UnmarshalText(text []byte) error {
+	if l := base32StdNoPad.DecodedLen(len(text)); l != len(h) {
+		return fmt.Errorf("tka.AUMHash.UnmarshalText: text wrong length: %d, want %d", l, len(text))
+	}
+	if _, err := base32StdNoPad.Decode(h[:], text); err != nil {
+		return fmt.Errorf("tka.AUMHash.UnmarshalText: %w", err)
+	}
+	return nil
+}
+
+// MarshalText implements encoding.TextMarshaler.
+func (h AUMHash) MarshalText() ([]byte, error) {
+	b := make([]byte, base32StdNoPad.EncodedLen(len(h)))
+	base32StdNoPad.Encode(b, h[:])
+	return b, nil
+}
+
 // AUMKind describes valid AUM types.
 type AUMKind uint8
 

tka/tailchonk.go

@@ -6,8 +6,6 @@ package tka
 
 import (
 	"bytes"
-	"encoding/base32"
-	"encoding/hex"
 	"fmt"
 	"io/ioutil"
 	"os"
@@ -190,7 +188,8 @@ func ChonkDir(dir string) (*FS, error) {
 // fsHashInfo describes how information about an AUMHash is represented
 // on disk.
 //
-// The CBOR-serialization of this struct is stored to base/hex(hash[0])/base32(hash[1:])
+// The CBOR-serialization of this struct is stored to base/__/base32(hash)
+// where __ are the first two characters of base32(hash).
 //
 // CBOR was chosen because we are already using it and it serializes
 // much smaller than JSON for AUMs. The 'keyasint' thing isn't essential
@@ -200,12 +199,11 @@ type fsHashInfo struct {
 	AUM      *AUM      `cbor:"2,keyasint"`
 }
 
-func (c *FS) dirPrefix(h AUMHash) string {
-	return filepath.Join(c.base, hex.EncodeToString(h[:1]))
-}
-
-func (c *FS) filename(h AUMHash) string {
-	return base32.StdEncoding.WithPadding(base32.NoPadding).EncodeToString(h[1:])
+// aumDir returns the directory an AUM is stored in, and its filename
+// within the directory.
+func (c *FS) aumDir(h AUMHash) (dir, base string) {
+	s := h.String()
+	return filepath.Join(c.base, s[:2]), s
 }
 
 // AUM returns the AUM with the specified digest.
@@ -256,7 +254,8 @@ func (c *FS) ChildAUMs(prevAUMHash AUMHash) ([]AUM, error) {
 }
 
 func (c *FS) get(h AUMHash) (*fsHashInfo, error) {
-	f, err := os.Open(filepath.Join(c.dirPrefix(h), c.filename(h)))
+	dir, base := c.aumDir(h)
+	f, err := os.Open(filepath.Join(dir, base))
 	if err != nil {
 		return nil, err
 	}
@@ -266,6 +265,9 @@ func (c *FS) get(h AUMHash) (*fsHashInfo, error) {
 	if err := cbor.NewDecoder(f).Decode(&out); err != nil {
 		return nil, err
 	}
+	if out.AUM != nil && out.AUM.Hash() != h {
+		return nil, fmt.Errorf("%s: AUM does not match file name hash %s", f.Name(), out.AUM.Hash())
+	}
 	return &out, nil
 }
 
@@ -297,24 +299,15 @@ func (c *FS) scanHashes(eachHashInfo func(*fsHashInfo)) error {
 		if !prefix.IsDir() {
 			continue
 		}
-		pb, err := hex.DecodeString(prefix.Name())
-		if err != nil || len(pb) != 1 {
-			return fmt.Errorf("invalid prefix directory %q: %v", prefix.Name(), err)
-		}
-
 		files, err := os.ReadDir(filepath.Join(c.base, prefix.Name()))
 		if err != nil {
 			return fmt.Errorf("reading prefix dir: %v", err)
 		}
 		for _, file := range files {
-			remainingHash, err := base32.StdEncoding.WithPadding(base32.NoPadding).DecodeString(file.Name())
-			if err != nil {
-				return fmt.Errorf("invalid aum file %s/%s: %v", prefix.Name(), file.Name(), err)
-			}
 			var h AUMHash
-			h[0] = pb[0]
-			copy(h[1:], remainingHash)
-
+			if err := h.UnmarshalText([]byte(file.Name())); err != nil {
+				return fmt.Errorf("invalid aum file: %s: %w", file.Name(), err)
+			}
 			info, err := c.get(h)
 			if err != nil {
 				return fmt.Errorf("reading %x: %v", h, err)
@@ -422,7 +415,8 @@ func (c *FS) commit(h AUMHash, updater func(*fsHashInfo)) error {
 		return fmt.Errorf("cannot commit AUM with hash %x to %x", toCommit.AUM.Hash(), h)
 	}
 
-	if err := os.MkdirAll(c.dirPrefix(h), 0755); err != nil && !os.IsExist(err) {
+	dir, base := c.aumDir(h)
+	if err := os.MkdirAll(dir, 0755); err != nil && !os.IsExist(err) {
 		return fmt.Errorf("creating directory: %v", err)
 	}
 
@@ -430,5 +424,5 @@ func (c *FS) commit(h AUMHash, updater func(*fsHashInfo)) error {
 	if err := cbor.NewEncoder(&buff).Encode(toCommit); err != nil {
 		return fmt.Errorf("encoding: %v", err)
 	}
-	return atomicfile.WriteFile(filepath.Join(c.dirPrefix(h), c.filename(h)), buff.Bytes(), 0644)
+	return atomicfile.WriteFile(filepath.Join(dir, base), buff.Bytes(), 0644)
 }

tka/tailchonk_test.go

@@ -146,13 +146,17 @@ func TestTailchonkFS_Commit(t *testing.T) {
 		t.Fatal(err)
 	}
 
-	if got, want := chonk.filename(aum.Hash()), "HJX3LPJJQVRFSQX4QONESBU4DUO5JPORA66ZUCFS6NHZWDZTP4"; got != want {
-		t.Errorf("aum filename = %q, want %q", got, want)
+	dir, base := chonk.aumDir(aum.Hash())
+	if got, want := dir, filepath.Join(chonk.base, "VU"); got != want {
+		t.Errorf("aum dir=%s, want %s", got, want)
 	}
-	if _, err := os.Stat(filepath.Join(chonk.base, "ad", "HJX3LPJJQVRFSQX4QONESBU4DUO5JPORA66ZUCFS6NHZWDZTP4")); err != nil {
+	if want := "VU5G7NN5FGCWEWKC7SBZUSIGTQOR3VF52ED33GQIWLZU7GYPGN7Q"; base != want {
+		t.Errorf("aum base=%s, want %s", base, want)
+	}
+	if _, err := os.Stat(filepath.Join(dir, base)); err != nil {
 		t.Errorf("stat of AUM file failed: %v", err)
 	}
-	if _, err := os.Stat(filepath.Join(chonk.base, "67", "226TIYPDKQWKFD5MXUI3GRVDSDFXRBABNINTFIT5ADMCLZ464U")); err != nil {
+	if _, err := os.Stat(filepath.Join(chonk.base, "M7", "M7LL2NDB4NKCZIUPVS6RDM2GUOIMW6EEAFVBWMVCPUANQJPHT3SQ")); err != nil {
 		t.Errorf("stat of AUM parent failed: %v", err)
 	}
 }