From 1e69cd7fb28932e39f3d70ec66b0a1c59d58a306 Mon Sep 17 00:00:00 2001
From: "Sewer." <seweratwork@gmail.com>
Date: Thu, 12 Dec 2024 20:56:13 +0000
Subject: [PATCH] patterns: Added BZip3 File Header Pattern (#329)

* Added: BZip3 Compression

* Fixed: Bound the SmallBlock data by parent's size

* Improved: Now uses the name 'Chunk' for block wrappers, such that the name 'block' matches with what the BZ3 API does

* Improved: Import rather than Include std::mem

* Added: Missing 'description' field in pragma
---
 README.md                                 |   1 +
 patterns/bzip3.hexpat                     |  64 ++++++++++++++++++++++
 tests/patterns/test_data/bzip3.hexpat.bz3 | Bin 0 -> 845 bytes
 3 files changed, 65 insertions(+)
 create mode 100644 patterns/bzip3.hexpat
 create mode 100644 tests/patterns/test_data/bzip3.hexpat.bz3

diff --git a/README.md b/README.md
index e9a378b..aa5a783 100644
--- a/README.md
+++ b/README.md
@@ -42,6 +42,7 @@ Everything will immediately show up in ImHex's Content Store and gets bundled wi
 | BSON | `application/bson` | [`patterns/bson.hexpat`](patterns/bson.hexpat) | BSON (Binary JSON) format |
 | bplist | | [`patterns/bplist.hexpat`](patterns/bplist.hexpat) | Apple's binary property list format (bplist) |
 | BSP | | [`patterns/bsp_goldsrc.hexpat`](patterns/bsp_goldsrc.hexpat) | GoldSrc engine maps format (used in Half-Life 1) |
+| BZIP3 | | [`patterns/bzip3.hexpat`](patterns/bzip3.hexpat) | GoldSrc engine maps format (used in Half-Life 1) |
 | CCHVA | | [`patterns/cchva.hexpat`](patterns/cchva.hexpat) | Command and Conquer Voxel Animation |
 | CCVXL | | [`patterns/ccvxl.hexpat`](patterns/ccvxl.hexpat) | Command and Conquer Voxel Model |
 | CCPAL | | [`patterns/ccpal.hexpat`](patterns/ccpal.hexpat) | Command and Conquer Voxel Palette |
diff --git a/patterns/bzip3.hexpat b/patterns/bzip3.hexpat
new file mode 100644
index 0000000..fe3dd71
--- /dev/null
+++ b/patterns/bzip3.hexpat
@@ -0,0 +1,64 @@
+#pragma author Sewer56
+#pragma description Parses BZip3 compression (file format) by Kamila Szewczyk
+#pragma MIME application/x-bzip3
+#pragma endian little
+#pragma magic [42 5A 33 76 31] @ 0x00
+import std.mem;
+
+// Helper function for bit counting
+fn popcount(u8 b) {
+    u32 count = 0;
+    while (b != 0) {
+        count = count + (b & 1);
+        b = b >> 1;
+    }
+    return count;
+};
+
+// Frame header structure
+struct FrameHeader {
+    char magic[5];         // "BZ3v1"
+    u32 blockSize;         // Maximum block size
+};
+
+// Small block header (for blocks < 64 bytes)
+struct SmallBlock {
+    u32 crc32;           // CRC32 checksum
+    u32 literal;         // Always 0xFFFFFFFF for small blocks
+    u8 data[parent.compressedSize - 8]; // Uncompressed data
+};
+
+// Regular block (blocks > 64 bytes)
+struct Block {
+    u32 crc32;            // CRC32 checksum of uncompressed data
+    u32 bwtIndex;         // Burrows-Wheeler transform index
+    u8  model;            // Compression model flags
+    
+    if ((model & 0x02) != 0)     
+        u32 lzpSize;      // Size after LZP compression
+    if ((model & 0x04) != 0)     
+        u32 rleSize;      // Size after RLE compression
+        
+    u8 data[parent.compressedSize - (popcount(model) * 4 + 9)];
+};
+
+// Main block structure
+struct Chunk {
+    u32 compressedSize;   // Size of compressed block
+    u32 origSize;         // Original uncompressed size
+        
+    if (compressedSize < 64) {
+        SmallBlock block;
+    } else {
+        Block block;
+    }
+};
+
+// Main parsing structure
+struct BZip3File {
+    FrameHeader header;
+    // Read blocks until end of file
+    Chunk chunks[while(!std::mem::eof())];
+};
+
+BZip3File file @ 0x0;
\ No newline at end of file
diff --git a/tests/patterns/test_data/bzip3.hexpat.bz3 b/tests/patterns/test_data/bzip3.hexpat.bz3
new file mode 100644
index 0000000000000000000000000000000000000000..0eac3e29eda3be73f2d107970089551cb5a23a63
GIT binary patch
literal 845
zcmV-T1G4-=S~GSr00002JOcm#k_G?(@9$2q6aWAK1gZuA0K~8aEx-T&|NrBOU!b>v
zn5n6HJ$|qc*_sr}V~&`4<_@ZV65{;#+G2oKSwjvp#cF3db_)EsKdM<pZ4A66kJiKE
z@|+^_7#L(Q`ld<xz12v;7xZ>t>T-~Pkzo@>+j9k*B@_ndaDY{dX&X~k`r$uEjm7Sp
zg&TCEkc22082S+VlXpDte9x#rN2&zZKJykQ4Ynwu{ML}Tk<=(v4~gceDySJk)e(pF
zsU!35mNA5sv03shN$w8HtbjlCkj@-e2sv}cefZ);>bS!vif62t3e}8NrfC0Tx`@j#
z(!hcT*Y!YVOCKUX51litc02*(euIy9<x(I?)3*>VQNe43@(5r!<v>;X8Z%A}05l5d
zQ6fQ%;f+DI5mKC$%X@04TS_nahiNq^mtJ~HvtXQQp11V?>FTg_NCmC;gU}!Q1^1<t
zDIhXPcG&MaFaTqu3)0oK_hE}_Xu3kgk+<ui;~TB<`$-6vvEDV^_0rs*)pira`gRL|
zuP-Le0Ow6E4E~`KR^9;0-Zc<@%+c3OMHMOgAfqsCFPpt|5AUdL#>^iw8>-h%V{kv%
zC&Qu8JPS&1k0x7zf>$Sd3Aw#qqQ)^->GE5qdEmOh?Y&hV@Fe=HQlxzkJ;0#hRNKoI
zy-e4P0%=A7P`Bh!8CzZ{T=?S|D`I<p*9}eqXyimoNnwzoop<hGc`lz|_^%dRfBIKk
zjVZ;!bRbT0Z2JlwKaCbi`EoDF^MNdVoOrOtS2=nWy$X4ixQ9_q<@}sIacasCmphea
zt3awj>S81soP=0TKEzMU6{3Y+hZ_F^eRwW1hI0Elw&z--m@|b1n)Z5dw+8c}9Hp!g
z9q(ZhY-7e@&C^$}n?9uC{I(P>&L0`B4{*7JM*)x7fjlD^Ep{r$8?GIb`IEh=HNSVv
zYM0k%5`)330fI*8?oJ@|E1;{$Lijt}(@~FJ%nEBXiGmyxAI<rd)2wd8-z1}LlZy=p
z+cc6bvses>9YY$n*3mYF;UYWeX$LJlb@Ha08x9>a|0Ry;D$uEZYjjY#u%KYMv)FZk
X0ln?%d=>rJ8Js+uq<Oy8CiLqq8*G<Q

literal 0
HcmV?d00001