start writing code for *.mcX file parsing, starting with indexing logic.

docs/mc-data-spec.md

@@ -98,8 +98,8 @@ The offset of a chunk (x,z) (in chunk coordinates) in the first table can be fou
 | data  | locations (4B)     | timestamps (4B)     | chunks and unused space |
 
 ##### chunk location
-Location info for a chunk is stored as a 32 bit big-endian integer, where the first three bytes are an offset in 4KiB sectors from the start of the file.
-The last byte gives the length of the chunk in 4KiB sectors. (rounded up, of course). Where chunks are always less than 1MiB in size.
+Location info for a chunk is stored as a 32 bit big-endian integer, where the first three bytes (0xFFFFFF00) are an offset in 4KiB sectors from the start of the file.
+The last byte (0x000000FF) gives the length of the chunk in 4KiB sectors. (rounded up, of course). Where chunks are always less than 1MiB in size.
 If a chunk isn't present in the region file (e.g. because it hasn't been generated or migrated yet), both fields are zero.
 
 ##### timestamps

src/dat/mcx.c

@@ -0,0 +1,24 @@
+#include "mcx.h"
+
+#include <endian.h>
+#include <stdint.h>
+#include <stdlib.h>
+
+#include "../util/compat/endian.h"
+#include "../util/intdef.h"
+
+/* an `*.mcX` contains a `0x2000` byte long table, the first `0x1000` containing
+ * `0x400` entries of chunk data.
+ * This chunk data is big-endian, where bytes `0xFFFFFF00` represent the `0x1000` sector offset.
+ * From the start, and bytes `0x000000FF` represent the length in `0x1000` sectors. */
+void mcx_index(const u8 *restrict buf, struct mcx_chunk *restrict chunks) {
+	const u32 *ptr = (u32 *)buf;
+	for (uint i = 0; i < 0x400; i++) {
+		u32 dat = be32toh(ptr[i]);
+		chunks[i] = (struct mcx_chunk){
+			.offset = (dat >> 8) * 0x1000,
+			.length = (dat & 0xFF) * 0x1000,
+			.time = be32toh(ptr[i + 0x400]),
+		};
+	}
+}

src/dat/mcx.h

@@ -2,6 +2,21 @@
 // Licensed under the MIT Licence. See LICENSE for details
 #pragma once
 
+#include <stdlib.h>
+
+#include "../util/atrb.h"
+#include "../util/intdef.h"
+
+/* contains chunk metadata */
+struct mcx_chunk {
+	size_t offset; // byte offset for start of chunk data
+	u32 length;    // byte length of chunk (+ padding)
+	u32 time;      // modification time in epoch seconds
+};
+
+/* indexes the chunks in an `*.mcX` file, writing `0x400` of entries to `chunks` */
+void mcx_index(const u8 *restrict buf, struct mcx_chunk *restrict chunks) NONNULL((1, 2));
+
 /* the MCR (Minecraft region) and MCA (Minecraft anvil) files are similar
  * MCA is the newer variant, where it includes:
  *  - a world height of 256, rather than 128.