package media import ( "encoding/hex" "strconv" ) // CIDv1 grammar (design §12.1; frontend chain/cid.ts is the twin): // multibase 'b' (RFC 4648 base32, lower case, no padding) over exactly // 0x01 · codec · 0x12 · 0x20 · digest[32], codec 0x55 (raw) or 0x70 // (dag-pb). That is 36 bytes = 58 base32 characters + 'b'. The unused two // trailing bits must be zero, so every CID has exactly one spelling. CIDv0 // ("Qm…"), upper case, other bases, codecs or hash functions are refused. const ( b32Alphabet = "abcdefghijklmnopqrstuvwxyz234567" cidBytes = 36 cidChars = 58 // ceil(36*8/5) mhSha256 = 0x12 digestLen = 32 ) // ParseCIDv1 returns the codec (CodecRaw or CodecDagPB) and the sha2-256 // digest as lower-case hex. func ParseCIDv1(s string) (codec int, digestHex string, err error) { if s == "" { return 0, "", mediaErr("cid: empty") } if len(s) >= 2 && s[0] == 'Q' && s[1] == 'm' { return 0, "", mediaErr("cid: CIDv0 is not accepted; convert it to CIDv1") } if s[0] != 'b' { return 0, "", mediaErr("cid: multibase must be 'b' (base32 lower case)") } if len(s) != cidChars+1 { return 0, "", mediaErr("cid: expected " + strconv.Itoa(cidChars+1) + " characters, got " + strconv.Itoa(len(s))) } bz, derr := base32Decode(s[1:]) if derr != nil { return 0, "", derr } if len(bz) != cidBytes { return 0, "", mediaErr("cid: wrong length") } if bz[0] != 1 { return 0, "", mediaErr("cid: version must be 1") } c := int(bz[1]) if c != CodecRaw && c != CodecDagPB { return 0, "", mediaErr("cid: codec must be raw (0x55) or dag-pb (0x70)") } if bz[2] != mhSha256 || bz[3] != digestLen { return 0, "", mediaErr("cid: multihash must be sha2-256") } return c, hex.EncodeToString(bz[4:]), nil } // CIDFromDigest is the inverse of ParseCIDv1. It panics on an unsupported // codec or a digest that is not 64 lower-case hex characters. func CIDFromDigest(codec int, digestHex string) string { if codec != CodecRaw && codec != CodecDagPB { panic("media: cid: unsupported codec") } d, ok := Raw32(digestHex) if !ok { panic("media: cid: digest must be 64 lower-case hex characters") } bz := make([]byte, 0, cidBytes) bz = append(bz, 1, byte(codec), mhSha256, digestLen) bz = append(bz, d...) return "b" + base32Encode(bz) } // IsHex32 reports whether s is exactly 64 lower-case hex characters. func IsHex32(s string) bool { _, ok := Raw32(s) return ok } // Raw32 decodes exactly 64 lower-case hex characters into 32 raw bytes // (one pass, table lookups: VM loops cost ~5k gas per character). func Raw32(s string) (string, bool) { if len(s) != 64 { return "", false } out := make([]byte, 32) for i := 0; i < 32; i++ { hi, lo := hexRev[s[2*i]], hexRev[s[2*i+1]] if hi > 15 || lo > 15 { return "", false } out[i] = hi<<4 | lo } return string(out), true } // hexRev and b32Rev map a byte to its digit value, 0xff when it is not a // lower-case hex / base32 digit. const ( hexRev = "\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\x00\x01\x02\x03\x04\x05\x06\x07\x08\x09\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\x0a\x0b\x0c\x0d\x0e\x0f\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff" b32Rev = "\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\x1a\x1b\x1c\x1d\x1e\x1f\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\x00\x01\x02\x03\x04\x05\x06\x07\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff" ) func base32Encode(bz []byte) string { out := make([]byte, 0, (len(bz)*8+4)/5) var buf uint32 bits := uint(0) for _, b := range bz { buf = buf<<8 | uint32(b) bits += 8 for bits >= 5 { bits -= 5 out = append(out, b32Alphabet[(buf>>bits)&31]) } buf &= (1 << bits) - 1 } if bits > 0 { out = append(out, b32Alphabet[(buf<<(5-bits))&31]) } return string(out) } // base32Decode is strict: lower-case alphabet only, and the unused // trailing bits must be zero. func base32Decode(s string) ([]byte, error) { out := make([]byte, 0, len(s)*5/8) var buf uint32 bits := uint(0) for i := 0; i < len(s); i++ { v := b32Rev[s[i]] if v > 31 { return nil, mediaErr("cid: invalid base32 character " + strconv.Quote(s[i:i+1])) } buf = buf<<5 | uint32(v) bits += 5 if bits >= 8 { bits -= 8 out = append(out, byte(buf>>bits)) buf &= (1 << bits) - 1 } } if bits >= 5 || buf != 0 { return nil, mediaErr("cid: non-canonical base32 padding bits") } return out, nil }