Initial decompression

1 files changed, 104 insertions, 0 deletions

diff --git a/src/main.zig b/src/main.zig
index d123e7f..34cb59c 100644
--- a/src/main.zig
+++ b/src/main.zig
@@ -4,6 +4,13 @@ const std = @import("std");
 // See spec.txt.
 
 const ZipFile = struct {
+    const CompressionMethod = enum(u16) {
+        store = 0,
+        deflate = 8,
+        lzma = 14,
+        zstd = 93,
+        xz = 95,
+    };
     // [local file header 1]
     // [encryption header 1]
     // [file data 1]
@@ -82,6 +89,51 @@ const ZipFile = struct {
     fn file_name(self: ZipFile, index: u16) []const u8 {
         return self.central_directory_headers[index].file_name;
     }
+    fn extract(self: ZipFile, index: u16, stream_or_file_in: anytype, stream_or_file_out: anytype) !void {
+        const cdh = self.central_directory_headers[index];
+        try stream_or_file_in.seekTo(cdh.relative_offset_of_local_header);
+        var lfh = try LocalFileHeader.read(self.allocator, stream_or_file_in);
+        defer lfh.deinit(self.allocator);
+        const is_encrypted = lfh.general_purpose_bit_flag.isSet(0);
+        if (is_encrypted) return error.EncryptionNotSupported;
+
+        var reader = stream_or_file_in.reader();
+        var lr = std.io.limitedReader(reader, lfh.compressed_size);
+        var limited_reader = lr.reader();
+        switch (lfh.compression_method) {
+            .store => { // no compression (store)
+                var writer = stream_or_file_out.writer();
+                try pump(limited_reader, writer, lfh.uncompressed_size, lfh.crc32);
+            },
+            .deflate => { // deflate
+                var decomp = try std.compress.deflate.decompressor(self.allocator, limited_reader, null);
+                defer decomp.deinit();
+                var decomp_reader = decomp.reader();
+                var writer = stream_or_file_out.writer();
+                try pump(decomp_reader, writer, lfh.uncompressed_size, lfh.crc32);
+            },
+            else => {
+                std.log.err("compression method {} not supported", .{lfh.compression_method});
+                return error.CompressionMethodNotSupported;
+            },
+        }
+    }
+    fn pump(reader: anytype, writer: anytype, expected_size_written: usize, expected_crc32: u32) !void {
+        var buf = [_]u8{0} ** 1024;
+        var crc32 = std.hash.Crc32.init();
+        var written: usize = 0;
+        while (true) {
+            const read = try reader.read(&buf);
+            if (read == 0) break;
+            const write = buf[0..read];
+            try writer.writeAll(write);
+
+            crc32.update(write);
+            written += read;
+        }
+        if (written != expected_size_written) return error.WrongUncompressedSize;
+        if (crc32.final() != expected_crc32) return error.WrongChecksum;
+    }
 };
 
 const CentralDirectoryHeader = struct {
@@ -143,6 +195,32 @@ const EndOfCentralDirectoryRecord = struct {
     }
 };
 
+const LocalFileHeader = struct {
+    const SIG: u32 = @as(u32, 0x04034b50);
+    version_needed_to_extract: u16,
+    general_purpose_bit_flag: std.bit_set.IntegerBitSet(16),
+    compression_method: ZipFile.CompressionMethod,
+    last_mod_file_time: u16,
+    last_mod_file_date: u16,
+    crc32: u32,
+    compressed_size: u32,
+    uncompressed_size: u32,
+    file_name_length: u16,
+    extra_field_length: u16,
+    file_name: []const u8,
+    extra_field: []const u8,
+    fn read(allocator: std.mem.Allocator, stream_or_file: anytype) !LocalFileHeader {
+        return read2(allocator, stream_or_file, LocalFileHeader, &[_]Dynamic{
+            .{ .field_name = "file_name", .length_field_name = "file_name_length" },
+            .{ .field_name = "extra_field", .length_field_name = "extra_field_length" },
+        });
+    }
+    fn deinit(self: *LocalFileHeader, allocator: std.mem.Allocator) void {
+        allocator.free(self.file_name);
+        allocator.free(self.extra_field);
+    }
+};
+
 const Dynamic = struct {
     field_name: []const u8,
     length_field_name: []const u8,
@@ -185,6 +263,18 @@ fn read2(
                 .Int => {
                     @field(t, field.name) = try reader.readIntLittle(field.type);
                 },
+                .Struct => |fsi| {
+                    if (fsi.backing_integer) |bi| {
+                        const int = try reader.readIntLittle(bi);
+                        @field(t, field.name) = @bitCast(int);
+                    } else @compileError("only packed struct with backing integer are supported, field " ++ field.name ++ " type " ++ @typeName(field.type) ++ "is not such a struct");
+                },
+                .Enum => |fei| {
+                    if (@typeInfo(fei.tag_type) == .Int) {
+                        const int = try reader.readIntLittle(fei.tag_type);
+                        @field(t, field.name) = @enumFromInt(int);
+                    } else @compileError("only enum with integer tag type are supported field " ++ field.name ++ " type " ++ @typeName(field.type) ++ "is not such a struct");
+                },
                 else => @compileError("don't know  how to handle field " ++ field.name ++ " of type " ++ @tagName(fti)),
             }
         }
@@ -210,3 +300,17 @@ test "foo2" {
     try std.testing.expectEqualStrings(zf.file_name(0), "hello.txt");
     try std.testing.expectEqualStrings(zf.file_name(1), "foo.txt");
 }
+
+test "extract" {
+    const test_zip = @embedFile("hello.zip");
+    var fbs = std.io.fixedBufferStream(test_zip);
+    var zf = try ZipFile.from(std.testing.allocator, &fbs);
+    defer zf.deinit();
+    var out = [_]u8{0} ** 1024;
+    var fbs_out = std.io.fixedBufferStream(&out);
+    try zf.extract(0, &fbs, &fbs_out);
+    try std.testing.expectEqualStrings("Hello, Zip!", fbs_out.getWritten());
+    fbs_out.reset();
+    try zf.extract(1, &fbs, &fbs_out);
+    try std.testing.expectEqualStrings("hi there\n", fbs_out.getWritten());
+}