From 9ae3b193c461b168336d1d9e272aa834705633d7 Mon Sep 17 00:00:00 2001
From: Aki <please@ignore.pl>
Date: Sun, 7 Aug 2022 21:45:05 +0200
Subject: Reimplemented part of archive format

This is getting reimplemented rather than refactor mostly in order to make
clear which parts are needed for backwards compatiblity (reading-wise) and
which are not. The current implementation has quite a number of quirks and
potential failure points despite not being large. Understanding them is not
worth it.
---
 ArchiveEx/Archive.cpp    | 82 ++++++++++++++++++++++++++++++++++++++++++++++++
 ArchiveEx/Archive.h      | 51 ++++++++++++++++++++++++++++++
 ArchiveEx/CMakeLists.txt | 13 ++++++++
 CMakeLists.txt           |  1 +
 4 files changed, 147 insertions(+)
 create mode 100644 ArchiveEx/Archive.cpp
 create mode 100644 ArchiveEx/Archive.h
 create mode 100644 ArchiveEx/CMakeLists.txt

diff --git a/ArchiveEx/Archive.cpp b/ArchiveEx/Archive.cpp
new file mode 100644
index 0000000..d714e46
--- /dev/null
+++ b/ArchiveEx/Archive.cpp
@@ -0,0 +1,82 @@
+#include "Archive.h"
+
+#include <cstdio>
+#include <memory>
+#include <vector>
+
+#include <zlib.h>
+
+
+namespace Archive
+{
+
+
+using UniqueFileHandle = std::unique_ptr<std::FILE, decltype(&std::fclose)>;
+
+
+static constexpr std::size_t DIRECTORY_MARGIN {64};
+
+
+static std::size_t BytesToBlocks(std::size_t bytes);
+
+
+Archive::Archive(const char* p) :
+	path {p},
+	header {},
+	entries {}
+{
+	UniqueFileHandle file {std::fopen(path, "rb"), &std::fclose};
+	if (!file)
+		throw "could not open archive";
+	std::size_t length = fread(&header, sizeof(Header), 1, file.get());
+	if (1 != length)
+		throw "could not read";
+	if (VERSION != header.version)
+		throw "bad version";
+	int err = std::fseek(file.get(), sizeof(Header) + header.directory.offset, SEEK_SET);
+	if (-1 == err)
+		throw "could not find directory in archive";
+	std::vector<std::uint8_t> compressed(header.directory.compressed_size);
+	length = std::fread(compressed.data(), 1, header.directory.compressed_size, file.get());
+	if (header.directory.compressed_size != length)
+		throw "could not read compressed directory";
+	const std::size_t total_entries = header.total_entries + DIRECTORY_MARGIN;
+	entries.resize(header.total_entries);
+	entries.reserve(total_entries);  // In original impl entries were switched on/off with their uncompressed size value.
+	uLongf uncompressed_size = sizeof(Entry) * total_entries;
+	err = ::uncompress(
+		reinterpret_cast<Bytef*>(entries.data()),
+		&uncompressed_size,
+		compressed.data(),
+		header.directory.compressed_size);
+	if (Z_OK != err)
+		throw "could not uncompress directory";
+}
+
+
+void
+Archive::PrintNamesOfEntries() const
+{
+	for (const auto& entry : entries)
+		std::printf("%s\n", entry.name);
+}
+
+
+std::size_t
+Archive::DirectoryBlocks() const
+{
+	const std::size_t blocks = BytesToBlocks(header.total_entries * sizeof(Entry));
+	return blocks == 0 ? 1 : blocks;
+}
+
+
+std::size_t
+BytesToBlocks(const std::size_t bytes)
+{
+	const std::size_t full = bytes / BLOCK_SIZE;
+	const std::size_t partial = (bytes % BLOCK_SIZE) > 0;
+	return full + partial;
+}
+
+
+}  // namespace Archive
diff --git a/ArchiveEx/Archive.h b/ArchiveEx/Archive.h
new file mode 100644
index 0000000..cc06dc4
--- /dev/null
+++ b/ArchiveEx/Archive.h
@@ -0,0 +1,51 @@
+#pragma once
+
+#include <cstddef>
+#include <cstdint>
+#include <vector>
+
+
+namespace Archive
+{
+
+
+static constexpr std::uint32_t VERSION {0x0010};
+static constexpr std::size_t BLOCK_SIZE {1024};
+static constexpr std::size_t NAMELEN {64};
+
+
+struct Header
+{
+	std::uint32_t version;
+	std::uint32_t total_entries;
+	struct {
+		std::uint32_t total_blocks;
+		std::uint32_t compressed_size;
+		std::uint32_t offset;
+	} directory;
+};
+
+
+struct Entry
+{
+	char name[NAMELEN];
+	std::uint32_t original_size;
+	std::uint32_t compressed_size;
+	std::uint32_t offset;
+};
+
+
+class Archive
+{
+public:
+	explicit Archive(const char* path);
+	void PrintNamesOfEntries() const;
+	std::size_t DirectoryBlocks() const;
+private:
+	const char* path;
+	Header header;
+	std::vector<Entry> entries;
+};
+
+
+}  // namespace Archive
diff --git a/ArchiveEx/CMakeLists.txt b/ArchiveEx/CMakeLists.txt
new file mode 100644
index 0000000..b2a1948
--- /dev/null
+++ b/ArchiveEx/CMakeLists.txt
@@ -0,0 +1,13 @@
+project(ArchiveEx)
+add_library(
+	ArchiveEx
+	Archive.cpp
+	)
+target_include_directories(
+	ArchiveEx
+	PUBLIC .
+	)
+target_link_libraries(
+	ArchiveEx
+	PRIVATE Zlib::zlib
+	)
diff --git a/CMakeLists.txt b/CMakeLists.txt
index fdb33d6..fea3356 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -14,6 +14,7 @@ if(MSVC)
 	set(WINDOWSSDK_LIBPATH "C:/Program Files (x86)/Windows Kits/10/Lib/10.0.19041.0" CACHE FILEPATH "Path to a versioned lib directory of selected Windows SDK")
 	set(WINDOWSSDK_PATH "C:/Program Files (x86)/Windows Kits/10/Include/10.0.19041.0" CACHE FILEPATH "Path to a versioned include directory of selected Windows SDK")
 endif()
+add_subdirectory(ArchiveEx)
 add_subdirectory(contrib)
 add_subdirectory(data)
 add_subdirectory(Datafile)
-- 
cgit v1.1