add multihash encoding

refactor to use size_t

include/mh/errors.h

@@ -7,6 +7,7 @@
 #define MH_E_TOO_SHORT -2
 #define MH_E_TOO_LONG -3
 #define MH_E_VARINT_NOT_SUPPORTED -4
+#define MH_E_DIGSET_TOO_LONG -5
 
 #define MH_E_LAST -5
 

include/mh/generic.h

@@ -1,6 +1,8 @@
-#ifndef MH_ASSERT
+#ifndef MH_GENERIC
-#define MH_ASSERT
+#define MH_GENERIC
 
 #define mh_assert_static(isTrue) void mh_assert_static(char x[1 - (!(isTrue))])
 
+#define UNUSED(x) (void)(x)
+
 #endif /* end of include guard */

include/mh/multihash.h

@@ -1,10 +1,28 @@
 #ifndef MH_HULTIHASH_H
 #define MH_HULTIHASH_H
 
+#include <stddef.h>
+
 // returns hash code or error (which is < 0)
-int mh_multihash_hash(const unsigned char multihash[], int len);
+int mh_multihash_hash(const unsigned char *multihash, size_t len);
 
 // returns length of multihash or error (which is < 0)
-int mh_multihash_length(const unsigned char multihash[], int len);
+int mh_multihash_length(const unsigned char *multihash, size_t len);
+
+// gives access to raw digset inside multihash buffer
+// returns 0 or negative error
+int mh_multihash_digset(const unsigned char *multihash, size_t len,
+	const unsigned char **digset, size_t *digset_len);
+
+// returns length in bytes of buffer needed to store multihash
+// with given hashcode and with given digset length
+// returns length or negative error code
+int mh_new_length(int code, size_t digset_len);
+
+// writes multihash into a buffer, the buffer needs to be at least
+// mh_new_length() bytes long.
+// returns negative error code or 0
+int mh_new(unsigned char *buffer, int code, const unsigned char *digset,
+	size_t digset_len);
 
 #endif /* end of include guard */

src/errors.c

@@ -13,6 +13,9 @@ const char *mh_error_string(int code) {
 		case MH_E_VARINT_NOT_SUPPORTED:
 			return "c-multihash does not yet support"
 				" varint encoding";
+		case MH_E_DIGSET_TOO_LONG:
+			return "c-multihash does not support digsets"
+				" longer than 127 bytes yet";
 		default:
 			return "unknown error code";
 	}

src/hashes.c

@@ -3,7 +3,7 @@
 #include "mh/hashes.h"
 
 #include "mh/errors.h"
-#include "mh/assert.h"
+#include "mh/generic.h"
 
 
 static const struct hash_info {

src/multihash.c

@@ -2,15 +2,30 @@
 
 #include "mh/hashes.h"
 #include "mh/errors.h"
+#include "mh/generic.h"
+
+#include <string.h>
+#include <stdlib.h>
 
 #define VARINT_MASK (1 << 7)
 
-static int check_multihash(const unsigned char mh[], int len) {
+static int check_len(size_t len) {
-	if (len < 3) {
+	if (len < 1) {
 		return MH_E_TOO_SHORT;
-	} else if (len >= 129) {
+	} else if (len >= 128) {
 		return MH_E_TOO_LONG;
-	} else if (mh[0] & VARINT_MASK) {
+	}
+
+	return MH_E_NO_ERROR;
+}
+
+static int check_multihash(const unsigned char mh[], size_t len) {
+	int error;
+
+	if (len < 3)
+		return MH_E_TOO_SHORT;
+
+	if (mh[0] & VARINT_MASK) {
 		// In near future multihash format will be
 		// extended with varints, this is how we are protecting
 		// against it.
@@ -18,12 +33,18 @@ static int check_multihash(const unsigned char mh[], int len) {
 	} else if (mh[1] & VARINT_MASK) {
 		return MH_E_VARINT_NOT_SUPPORTED;
 	}
+
+	error = check_len(mh[1]);
+	if (error)
+		return error;
+
+
 	return MH_E_NO_ERROR;
 }
 
 
 // returns hash code or error (which is < 0)
-int mh_multihash_hash(const unsigned char mh[], int len) {
+int mh_multihash_hash(const unsigned char *mh, size_t len) {
 	if (check_multihash(mh, len))
 		return check_multihash(mh, len);
 	return (int) mh[0];
@@ -31,9 +52,44 @@ int mh_multihash_hash(const unsigned char mh[], int len) {
 
 
 // returns length of multihash or error (which is < 0)
-int mh_multihash_length(const unsigned char mh[], int len) {
+int mh_multihash_length(const unsigned char *mh, size_t len) {
 	if (check_multihash(mh, len))
 		return check_multihash(mh, len);
 	return (int) mh[1];
 }
 
+// gives access to raw digset inside multihash buffer
+// returns 0 or negative error
+int mh_multihash_digset(unsigned char *multihash, size_t len, unsigned char **digset,
+		size_t *digset_len) {
+	int error = check_multihash(multihash, len);
+	if (error)
+		return error;
+
+	(*digset_len) = (size_t) mh_multihash_length(multihash, len);
+	(*digset) = multihash + 2; // Always true without varint
+
+	return MH_E_NO_ERROR;
+}
+
+int mh_new_length(int code, size_t hash_len) {
+	// right now there is no varint support
+	// so length required is 2 + hash_len
+	UNUSED(code);
+	return 2 + hash_len;
+}
+
+int mh_new(unsigned char *buffer, int code, const unsigned char *digset,
+	size_t digset_len) {
+	if (code & VARINT_MASK)
+		return MH_E_VARINT_NOT_SUPPORTED;
+	if (digset_len > 127)
+		return MH_E_DIGSET_TOO_LONG;
+
+	buffer[0] = (unsigned char) ((unsigned int) code) & 255;
+	buffer[1] = (unsigned char) digset_len;
+	memcpy(buffer + 2, digset, digset_len);
+
+	return MH_E_NO_ERROR;
+}
+

tests/c/minunit.h

@@ -4,6 +4,7 @@
 #define mu_run_test(test) do { char *message; printf(" - run: %s\n", #test); \
 	message = test(); tests_run++; \
 	if (message) return message; } while (0)
+
 int tests_run = 0;
 static char *mu_all_tests(void);
 
@@ -21,3 +22,31 @@ int main(void) {
 	return result != 0;
 }
 
+void dump(const unsigned char *data, size_t size) {
+	char ascii[17];
+	size_t i, j;
+	ascii[16] = '\0';
+	for (i = 0; i < size; ++i) {
+		printf("%02X ", ((unsigned char*)data)[i]);
+		if (((unsigned char*)data)[i] >= ' ' && ((unsigned char*)data)[i] <= '~') {
+			ascii[i % 16] = ((unsigned char*)data)[i];
+		} else {
+			ascii[i % 16] = '.';
+		}
+		if ((i+1) % 8 == 0 || i+1 == size) {
+			printf(" ");
+			if ((i+1) % 16 == 0) {
+				printf("|  %s \n", ascii);
+			} else if (i+1 == size) {
+				ascii[(i+1) % 16] = '\0';
+				if ((i+1) % 16 <= 8) {
+					printf(" ");
+				}
+				for (j = (i+1) % 16; j < 16; ++j) {
+					printf("   ");
+				}
+				printf("|  %s \n", ascii);
+			}
+		}
+	}
+}

tests/c/test_multihash_new.c

@@ -0,0 +1,54 @@
+#include <string.h>
+#include "minunit.h"
+#include "mh/multihash.h"
+
+#include "mh/hashes.h"
+
+#include "examples.h"
+
+char error_buf[256];
+
+static char *test_multihash_new_crafts_right_multihash(void) {
+	int error;
+	unsigned char buf[256]; // much bigger than needed
+	size_t digset_len = -1;
+	const unsigned char *digset = NULL;
+
+	error = mh_multihash_digset(sha1_example, sha1_example_length,
+		&digset, &digset_len);
+	mu_assert("getting digset", error == MH_E_NO_ERROR);
+
+	error = mh_new(buf, MH_H_SHA1, digset, digset_len);
+	mu_assert("creating multihash", error == MH_E_NO_ERROR);
+
+	mu_assert("crafted multihash is the same", memcmp(sha1_example, buf,
+		sha1_example_length) == 0);
+
+	return NULL;
+}
+
+static char *test_multihash_new_is_reversible(void) {
+	int error = MH_E_NO_ERROR;
+	int code = MH_H_SHA3_512;
+	const unsigned char *digset = random_512;
+	const size_t digset_len = 512 / 8;
+
+	unsigned char mh[256];
+	const size_t mh_len = mh_new_length(code, digset_len);
+
+	error = mh_new(mh, MH_H_SHA3_512, digset, digset_len);
+	mu_assert("creating multihash", error == MH_E_NO_ERROR);
+
+	mu_assert("reading code", mh_multihash_hash(mh, mh_len) == MH_H_SHA3_512);
+	mu_assert("reading length", mh_multihash_length(mh, mh_len) ==
+		(int) digset_len);
+
+	return NULL;
+}
+
+static char *mu_all_tests(void) {
+	mu_run_test(test_multihash_new_crafts_right_multihash);
+	mu_run_test(test_multihash_new_is_reversible);
+	return NULL;
+}
+