Fix big endian issues writing and reading .LYB files

The print and parse routines which handle .LYB files assume a little endian
CPU. This means libyang is failing when it is compiled for a big endian
target.

For example consider the code:
 lyb_write(out, (uint8_t *)&mod_count, 2, lybs);

mod_count is an integer and on a typical 32 bit big endian target its
format in memory is (byte0(MSB) byte1 byte2 byte3(LSB)) meaning
if mod_count has a value of 1 the first three bytes of mod_count
are zero, and as the routine only write the first two bytes, it writes
out 0. This effectively corrupts the .LYB file.

This patch modifies the routines to write and read .LYB files to use
endian agnostic methods to access data in a .LYB file.
diff --git a/src/parser_lyb.c b/src/parser_lyb.c
index e2e7baf..9b504b9 100644
--- a/src/parser_lyb.c
+++ b/src/parser_lyb.c
@@ -17,6 +17,7 @@
 #include <stdlib.h>
 #include <string.h>
 #include <inttypes.h>
+#include <endian.h>
 
 #include "libyang.h"
 #include "common.h"
@@ -32,7 +33,7 @@
 {
     int ret = 0, i, empty_chunk_i;
     size_t to_read;
-    LYB_META meta;
+    uint8_t meta_buf[LYB_META_BYTES];
 
     assert(data && lybs);
 
@@ -76,12 +77,9 @@
 
         if (empty_chunk_i > -1) {
             /* read the next chunk meta information */
-            memcpy(&meta, data + ret, LYB_META_BYTES);
-            lybs->written[empty_chunk_i] = 0;
-            lybs->inner_chunks[empty_chunk_i] = 0;
-
-            memcpy(&lybs->written[empty_chunk_i], &meta, LYB_SIZE_BYTES);
-            memcpy(&lybs->inner_chunks[empty_chunk_i], ((uint8_t *)&meta) + LYB_SIZE_BYTES, LYB_INCHUNK_BYTES);
+            memcpy(meta_buf, data + ret, LYB_META_BYTES);
+            lybs->written[empty_chunk_i] = meta_buf[0];
+            lybs->inner_chunks[empty_chunk_i] = meta_buf[1];
 
             /* remember whether there is a following chunk or not */
             lybs->position[empty_chunk_i] = (lybs->written[empty_chunk_i] == LYB_SIZE_MAX ? 1 : 0);
@@ -113,7 +111,9 @@
 static int
 lyb_read_enum(uint64_t *enum_idx, uint32_t count, const char *data, struct lyb_state *lybs)
 {
+    int ret = 0;
     size_t bytes;
+    uint64_t tmp_enum = 0;
 
     if (count < (2 << 8)) {
         bytes = 1;
@@ -125,7 +125,11 @@
         bytes = 4;
     }
 
-    return lyb_read_number(enum_idx, bytes, data, lybs);
+    /* The enum is always read into a uint64_t buffer */
+    ret = lyb_read_number(&tmp_enum, bytes, data, lybs);
+    *enum_idx = le64toh(tmp_enum);
+
+    return ret;
 }
 
 static int
@@ -133,10 +137,12 @@
 {
     int next_chunk = 0, r, ret = 0;
     size_t len = 0, cur_len;
+    uint8_t len_buf[2];
 
     if (with_length) {
-        ret += (r = lyb_read(data, (uint8_t *)&len, 2, lybs));
+        ret += (r = lyb_read(data, len_buf, 2, lybs));
         LYB_HAVE_READ_GOTO(r, data, error);
+        len = len_buf[0] | (len_buf[1] << 8);
     } else {
         /* read until the end of this subtree */
         len = lybs->written[lybs->used - 1];
@@ -190,7 +196,7 @@
 static int
 lyb_read_start_subtree(const char *data, struct lyb_state *lybs)
 {
-    LYB_META meta;
+    uint8_t meta_buf[LYB_META_BYTES];
 
     if (lybs->used == lybs->size) {
         lybs->size += LYB_STATE_STEP;
@@ -200,14 +206,11 @@
         LY_CHECK_ERR_RETURN(!lybs->written || !lybs->position || !lybs->inner_chunks, LOGMEM(NULL), -1);
     }
 
-    memcpy(&meta, data, LYB_META_BYTES);
+    memcpy(meta_buf, data, LYB_META_BYTES);
 
     ++lybs->used;
-    lybs->written[lybs->used - 1] = 0;
-    lybs->inner_chunks[lybs->used - 1] = 0;
-
-    memcpy(&lybs->written[lybs->used - 1], &meta, LYB_SIZE_BYTES);
-    memcpy(&lybs->inner_chunks[lybs->used - 1], ((uint8_t *)&meta) + LYB_SIZE_BYTES, LYB_INCHUNK_BYTES);
+    lybs->written[lybs->used - 1] = meta_buf[0];
+    lybs->inner_chunks[lybs->used - 1] = meta_buf[LYB_SIZE_BYTES];
     lybs->position[lybs->used - 1] = (lybs->written[lybs->used - 1] == LYB_SIZE_MAX ? 1 : 0);
 
     return LYB_META_BYTES;
@@ -219,14 +222,16 @@
     int r, ret = 0;
     char *mod_name = NULL, mod_rev[11];
     uint16_t rev = 0;
+    uint8_t tmp_buf[2];
 
     /* model name */
     ret += (r = lyb_read_string(data, &mod_name, 1, lybs));
     LYB_HAVE_READ_GOTO(r, data, error);
 
     /* revision */
-    ret += (r = lyb_read(data, (uint8_t *)&rev, sizeof rev, lybs));
+    ret += (r = lyb_read(data, tmp_buf, sizeof tmp_buf, lybs));
     LYB_HAVE_READ_GOTO(r, data, error);
+    rev = tmp_buf[0] | (tmp_buf[1] << 8);
 
     if (rev) {
         sprintf(mod_rev, "%04u-%02u-%02u", ((rev & 0xFE00) >> 9) + 2000, (rev & 0x01E0) >> 5, (rev & 0x001F));
@@ -430,15 +435,18 @@
     case LY_TYPE_INT16:
     case LY_TYPE_UINT16:
         ret = lyb_read_number((uint64_t *)&value->uint16, 2, data, lybs);
+        value->uint16 = le16toh(value->uint16);
         break;
     case LY_TYPE_INT32:
     case LY_TYPE_UINT32:
         ret = lyb_read_number((uint64_t *)&value->uint32, 4, data, lybs);
+        value->uint32 = le32toh(value->uint32);
         break;
     case LY_TYPE_DEC64:
     case LY_TYPE_INT64:
     case LY_TYPE_UINT64:
         ret = lyb_read_number((uint64_t *)&value->uint64, 8, data, lybs);
+        value->uint64 = le64toh(value->uint64);
         break;
     default:
         return -1;
@@ -1101,10 +1109,12 @@
 lyb_parse_data_models(struct ly_ctx *ctx, const char *data, struct lyb_state *lybs)
 {
     int i, r, ret = 0;
+    uint8_t mod_count_buf[2];
 
     /* read model count */
-    ret += (r = lyb_read(data, (uint8_t *)&lybs->mod_count, 2, lybs));
+    ret += (r = lyb_read(data, mod_count_buf, 2, lybs));
     LYB_HAVE_READ_RETURN(r, data, -1);
+    lybs->mod_count = mod_count_buf[0] | (mod_count_buf[1] << 8);
 
     lybs->models = malloc(lybs->mod_count * sizeof *lybs->models);
     LY_CHECK_ERR_RETURN(!lybs->models, LOGMEM(NULL), -1);