Working put test, passing, but not complete

README.md

@@ -5,3 +5,7 @@ Do it simply and do it well.
 ## The Bitcask paper
 
 https://riak.com/assets/bitcask-intro.pdf
+
+## Implementation in Ruby
+
+https://dinesh.wiki/posts/build-your-own-persistent-kv-store/

src/Bitcask/bitcask.zig

@@ -1,17 +1,64 @@
 const std = @import("std");
 const expect = std.testing.expect;
+const Dir = std.fs.Dir;
+const mem = std.mem;
 
-pub const BitcaskFileError = error{
+const BitcaskFileError = error{
+    CannotInitialize,
     AccessDenied,
     OutOfMemory,
     FileNotFound,
+    Unknown
 };
 
-pub const BitCask = struct {
+const CaskEntry = struct {
+    keySize: u32,
+    valSize: u32,
+    valType: EntryType,
+    key:     []const u8,
+    val:     *const anyopaque,
+};
+
+const KV = struct {
+    key: []const u8,
+    val: []const u8
+};
+
+const EntryType = enum(u8) {
+    str = 0,
+    int = 1,
+    flt = 2,
+};
+
+
+const BitCask = struct {
+
+    const KeyDirEntry = struct {
+        file_name: []const u8,
+        offset: u4,
+        size: u4,
+    };
+
+
     const FILE_THRESHOLD_SIZE = 1000;
 
+    var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
+    const allocator = arena.allocator();
     
-    // std.StringArrayHashMap
+
+    // TODO fix to a file descriptor or reference for fixed-size, may only need u4? depends on max file size I suppose
+    // u4 = key size of 4.29GB
+
+    var keyDir: std.StringArrayHashMap(KeyDirEntry) = undefined;
+    var current_file: std.fs.File = undefined;
+    var current_file_name: [*]const u8 = "";
+
+    // bitcask:close(BitCaskHandle) Close a Bitcask data store and flush all pending writes (if any) to disk
+    pub fn close() void {
+        std.debug.print("Closing bitcask...\n", .{});
+        arena.deinit();
+        current_file.close();
+    }
 
     // From the Bitcask paper, the API should look something like this
 
@@ -29,66 +76,187 @@ pub const BitCask = struct {
     // writer and not just a reader) and sync on put (if this writer would
     // prefer to sync the write file after every write operation).
     // The directory must be readable and writable by this process, and
-    // only one process may open a Bitcask with read write at a time.
+    // only one process may open a Bitcask with read or write at a time.
     // bitcask:open(DirectoryName) Open a new or existing Bitcask datastore for read-only access.
     // → BitCaskHandle | {error, any()} The directory and all files in it must be readable by this process.
-    fn open(directory_name: []const u8) BitcaskFileError!void {
+    pub fn open(directory_name: []const u8) (std.fs.File.OpenError || BitcaskFileError)!void {
-        std.debug.print("Opening bitcask in dir {s}\n", .{directory_name});
+        // init keyDir
-        return error.FileNotFound;
+        keyDir = std.StringArrayHashMap(KeyDirEntry).init(
+            allocator,
+        );
+        const file_name = "cask1.db";
+        current_file_name = file_name;
+        // Open file
+        std.debug.print("\nOpening bitcask in dir {s}/\n", .{directory_name});
+        var bitcask_dir = try std.fs.Dir.openDir(std.fs.cwd(), directory_name, .{.access_sub_paths=false});
+        // TODO get all files in cask and have some order to open the latest one
+        var err_or_file = bitcask_dir.openFile(file_name, .{.mode=.read_write});
+        
+        // If the file does not exist, create it
+        if (err_or_file == std.fs.File.OpenError.FileNotFound) {
+            std.debug.print("No file found, creating new one\n", .{});
+            _ = try bitcask_dir.createFile(file_name, .{});  // catch this error, then I can remove error union
+            current_file = try bitcask_dir.openFile(file_name, .{.mode=.read_write});
+        } else if (@TypeOf(err_or_file) != std.fs.File.OpenError) {
+            std.debug.print("File {s} found\n", .{file_name});
+            current_file = try err_or_file;
+        } else {
+            std.debug.print("Encountered unknown error while opening bitcask: {any}\n", .{err_or_file});
+            return error.CannotInitialize;
+        }
+
+//         std.debug.print("Current File open: {}\n", .{current_file});
+        return;
     }
 
     // bitcask:get(BitCaskHandle, Key) Retrieve a value by key from a Bitcask datastore.
     // → not found | {ok, Value}
-    fn get(key: []const u8) error{NotImplemented}!void {
+    pub fn get(key: []const u8) error{OutOfMemory, NoCaskFound, Unknown}!CaskEntry {
         std.debug.print("Getting value with key {s}\n", .{key});
-        return error.NotImplemented;
+        const optional_cask = keyDir.get(key);
+        if (optional_cask == null) {
+            return error.NoCaskFound;
+        }
+        const cask = optional_cask.?; // shorthand for 'optional_cask orelse unreachable'
+        std.debug.print("Got cask from file {s}\n", .{cask.file});
+        return cask;
+        // TODO get entry from db rather than return the keyDir entry
     }
 
     // bitcask:put(BitCaskHandle, Key, Value) Store a key and value in a Bitcask datastore.
     // → ok | {error, any()}
-    fn put() error{NotImplemented}!void {
+    pub fn put(data: KV) BitcaskFileError!void {
-        return error.NotImplemented;
+
+//         const key_type: EntryType = EntryType.str;
+//         const key_size = getEntrySize(data.key, key_type);
+//         const val_type: EntryType = EntryType.str;
+//         const val_size = getEntrySize(data.val, val_type);
+//
+
+
+        const entry_bytes = serializeCaskEntry(&allocator, data) catch |err| {
+            std.debug.print("Encountered unknown error while serializing bitcask: {any}\n", .{err});
+            return BitcaskFileError.Unknown;
+        };
+        std.debug.print("Writing data to cask: {}\n", .{data});
+        const i = current_file.write(entry_bytes) catch |err| {
+            std.debug.print("Encountered unknown error while writing to bitcask: {any}\n", .{err});
+            return BitcaskFileError.Unknown;
+        };
+        std.debug.print("After write: {any}\n", .{i});
+
+
+        //try keyDir.put("key", data.val);
+
+        return;
     }
 
-    // bitcask:delete(BitCaskHandle, Key) Delete a key from a Bitcask datastore.
+    // bitcask:delete(BitCaskHandle, Key) Delete a key from a Bitcask datastore.  Note that deletion is simply a write of a special tombstone value, which will be removed on the next merge
     // → ok | {error, any()}
-    fn delete() error{NotImplemented}!void {
+    pub fn delete() error{NotImplemented}!void {
         return error.NotImplemented;
     }
 
     // bitcask:list keys(BitCaskHandle) List all keys in a Bitcask datastore.
     // → [Key] | {error, any()}
-    fn list() error{NotImplemented}!void {
+    pub fn list() error{NotImplemented}!void {
         return error.NotImplemented;
     }
 
     // bitcask:fold(BitCaskHandle,Fun,Acc0) Fold over all K/V pairs in a Bitcask datastore.
     // → Acc Fun is expected to be of the form: F(K,V,Acc0) → Acc.
-    fn fold() error{NotImplemented}!void {
+    pub fn fold() error{NotImplemented}!void {
         return error.NotImplemented;
     }
 
     // bitcask:merge(DirectoryName) Merge several data files within a Bitcask datastore into a more
     // → ok | {error, any()} compact form. Also, produce hintfiles for faster startup.
-    fn merge() error{NotImplemented}!void {
+    pub fn merge() error{NotImplemented}!void {
         return error.NotImplemented;
     }
 
     // bitcask:sync(BitCaskHandle) Force any writes to sync to disk.
     // → ok
-    fn sync() error{NotImplemented}!void {
+    pub fn sync() error{NotImplemented}!void {
         return error.NotImplemented;
     }
 
-    // bitcask:close(BitCaskHandle) Close a Bitcask data store and flush all pending writes (if any) to disk
+    fn getEntrySize(data: *anyopaque, t: EntryType) u32 {
-    fn close() error{NotImplemented}!void {
+        switch (t) {
-        return error.NotImplemented;
+            .str => {
+                const ptr = @ptrCast([*]const u8, data);
+                return ptr.len;
+            },
+            .int => {
+                return @sizeOf(i32);
+            },
+            .flt => {
+                return @sizeOf(f64);
             }
+        }
+    }
+
+    fn serializeCaskEntry(aloc: *const mem.Allocator, entry: KV) error{OutOfMemory}![]u8 {
+
+        var serialized_data = try std.ArrayList(u8).initCapacity(aloc.*, entry.key.len + entry.val.len + 8);
+
+        // Serialize u32 into 4 u8 bytes
+        const keyBytes = std.mem.toBytes(entry.key.len);
+        try serialized_data.appendSlice(&keyBytes);        // Serialize u32 into 4 u8 bytes
+        // Serialize u32 into 4 u8 bytes
+        const valBytes = std.mem.toBytes(entry.val.len);
+        try serialized_data.appendSlice(&valBytes);
+        //try serialized_data.append(@enumToInt(entry.valType));
+
+        for (entry.key) |byte| {
+            try serialized_data.append(byte);
+        }
+
+        for (entry.val) |byte| {
+            try serialized_data.append(byte);
+        }
+
+        // Convert the val any type to the type specified
+//         switch (entry.valType) {
+//             .str => {
+//                 const str_val = @ptrCast([*]const u8, entry.val);
+//                 for (str_val) |byte| {
+//                     try serialized_data.append(byte);
+//                 }
+//             },
+//             .int => {
+//                 const int_val = @ptrCast(*const i32, entry.val);
+//                 try serialized_data.appendSlice(mem.asBytes(int_val));
+//             },
+//             .flt => {
+//                 const flt_val = @ptrCast(*const f64, entry.val);
+//                 try serialized_data.appendSlice(mem.asBytes(flt_val));
+//             }
+//         }
+
+        return try serialized_data.toOwnedSlice();
+    }
+
+//     fn getValType(caskVal: any) u8 {
+//         switch (@TypeOf(caskVal)) {
+//             u8 =>
+//         }
+//     }
+
 };
 
+// pub fn main() !void {
+//     const bc = BitCask;
+//     try bc.open("data");
+//     defer bc.close();
+//
+//     try bc.put(.{.key = "id", .val = "abc123"});
+// }
+
 test "Bitcask spec implementation: open" {
     const bc = BitCask;
-    try bc.open("File");
+    try bc.open("data");
+    defer bc.close();
     // bc.open("File") catch |err| {
     //     try expect(err == error.FileNotFound);
     // };
@@ -97,12 +265,33 @@ test "Bitcask spec implementation: open" {
 
 test "Bitcask spec implementation: get" {
     const bc = BitCask;
-    try bc.get("key");
+    //bc.init();
+    try bc.open("data");
+    defer bc.close();
+    _ = bc.get("key") catch |err| {
+        try expect(err == error.NoCaskFound);
+    };
+    //try bc.put();
+    const cask = try bc.get("key");
+    try expect(std.mem.eql(u8, cask.file, "file.txt"));
+    try expect(cask.offset == 0);
+    try expect(cask.size == 0);
 }
 
 test "Bitcask spec implementation: put" {
     const bc = BitCask;
-    try bc.put();
+    try bc.open("data");
+    defer bc.close();
+
+    const key = "id";
+    const val = "1";
+
+    try bc.put(.{
+        .key = key,
+        .val = val
+    });
+
+    try bc.put(.{.key = "user1", .val = "likes apples so much"});
 }
 
 test "Bitcask spec implementation: delete" {