More spec compliant Blob.prototype.type (#2340)

* Make `Blob.prototype. type` more spec compliant * Add a few more checks for isNumber() * Fix `make headers` * Safer JSValue.isString() * More tests for blob.slice * Make `Blob.prototype.type` more spec compliant * Add isASCII check * Fix types * Fix failing type test * Update blob.zig * Update blob.zig * Fix .eql check on empty values --------- Co-authored-by: Jarred Sumner <[email protected]>
oven-sh · Mar 8, 2023 · f97c8ac · f97c8ac
1 parent c04f582
commit f97c8ac
Show file tree

Hide file tree

Showing 16 changed files with 261 additions and 63 deletions.
diff --git a/packages/bun-types/bun.d.ts b/packages/bun-types/bun.d.ts
@@ -620,8 +620,26 @@ declare module "bun" {
      *
      * @param begin - start offset in bytes
      * @param end - absolute offset in bytes (relative to 0)
+     * @param contentType - MIME type for the new FileBlob
      */
-    slice(begin?: number, end?: number): FileBlob;
+    slice(begin?: number, end?: number, contentType?: string): FileBlob;
+
+    /**
+     * Offset any operation on the file starting at `begin`
+     *
+     * Similar to [`TypedArray.subarray`](https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/TypedArray/subarray). Does not copy the file, open the file, or modify the file.
+     *
+     * If `begin` > 0, {@link Bun.write()} will be slower on macOS
+     *
+     * @param begin - start offset in bytes
+     * @param contentType - MIME type for the new FileBlob
+     */
+    slice(begin?: number, contentType?: string): FileBlob;
+
+    /**
+     * @param contentType - MIME type for the new FileBlob
+     */
+    slice(contentType?: string): FileBlob;
 
     /**
      * Incremental writer for files and pipes.

diff --git a/packages/bun-types/globals.d.ts b/packages/bun-types/globals.d.ts
@@ -565,7 +565,29 @@ declare class Blob implements BlobInterface {
    * @param end The index that sets the end of the view.
    *
    */
-  slice(begin?: number, end?: number): Blob;
+  slice(begin?: number, end?: number, contentType?: string): Blob;
+
+  /**
+   * Create a new view **without 🚫 copying** the underlying data.
+   *
+   * Similar to [`BufferSource.subarray`](https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/BufferSource/subarray)
+   *
+   * @param begin The index that sets the beginning of the view.
+   * @param end The index that sets the end of the view.
+   *
+   */
+  slice(begin?: number, contentType?: string): Blob;
+
+  /**
+   * Create a new view **without 🚫 copying** the underlying data.
+   *
+   * Similar to [`BufferSource.subarray`](https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/BufferSource/subarray)
+   *
+   * @param begin The index that sets the beginning of the view.
+   * @param end The index that sets the end of the view.
+   *
+   */
+  slice(contentType?: string): Blob;
 
   /**
    * Read the data from the blob as a string. It will be decoded from UTF-8.

diff --git a/packages/bun-types/tests/array.test-d.ts b/packages/bun-types/tests/array.test-d.ts
@@ -13,6 +13,4 @@ async function* listReleases() {
   }
 }
 
-const releases = await Array.fromAsync(listReleases());
-
-export {};
+export const releases = await Array.fromAsync(listReleases());
diff --git a/src/bun.js/api/server.zig b/src/bun.js/api/server.zig
@@ -376,7 +376,9 @@ pub const ServerConfig = struct {
             }
 
             if (arg.getTruthy(global, "maxRequestBodySize")) |max_request_body_size| {
-                args.max_request_body_size = @intCast(u64, @max(0, max_request_body_size.toInt64()));
+                if (max_request_body_size.isNumber()) {
+                    args.max_request_body_size = @intCast(u64, @max(0, max_request_body_size.toInt64()));
+                }
             }
 
             if (arg.getTruthy(global, "error")) |onError| {

diff --git a/src/bun.js/base.zig b/src/bun.js/base.zig
@@ -1773,6 +1773,8 @@ pub const ArrayBuffer = extern struct {
     }
 
     pub fn createEmpty(globalThis: *JSC.JSGlobalObject, comptime kind: JSC.JSValue.JSType) JSValue {
+        JSC.markBinding(@src());
+
         return switch (comptime kind) {
             .Uint8Array => Bun__createUint8ArrayForCopy(globalThis, null, 0, false),
             .ArrayBuffer => Bun__createArrayBufferForCopy(globalThis, null, 0),

diff --git a/src/bun.js/bindings/bindings.cpp b/src/bun.js/bindings/bindings.cpp
@@ -2455,10 +2455,6 @@ bool JSC__JSValue__isPrimitive(JSC__JSValue JSValue0)
 {
     return JSC::JSValue::decode(JSValue0).isPrimitive();
 }
-bool JSC__JSValue__isString(JSC__JSValue JSValue0)
-{
-    return JSC::JSValue::decode(JSValue0).isString();
-}
 bool JSC__JSValue__isSymbol(JSC__JSValue JSValue0)
 {
     return JSC::JSValue::decode(JSValue0).isSymbol();

diff --git a/src/bun.js/bindings/bindings.zig b/src/bun.js/bindings/bindings.zig
@@ -159,6 +159,9 @@ pub const ZigString = extern struct {
     }
 
     pub fn eql(this: ZigString, other: ZigString) bool {
+        if (this.len == 0 or other.len == 0)
+            return this.len == other.len;
+
         const left_utf16 = this.is16Bit();
         const right_utf16 = other.is16Bit();
 
@@ -2276,6 +2279,10 @@ pub const JSGlobalObject = extern struct {
         return this.bunVM().allocator;
     }
 
+    pub fn throwOutOfMemory(this: *JSGlobalObject) void {
+        this.throwValue(this.createErrorInstance("Out of memory", .{}));
+    }
+
     pub fn throwInvalidArguments(
         this: *JSGlobalObject,
         comptime fmt: string,
@@ -3454,8 +3461,11 @@ pub const JSValue = enum(JSValueReprInt) {
         return res;
     }
 
-    pub fn isString(this: JSValue) bool {
-        return cppFn("isString", .{this});
+    pub inline fn isString(this: JSValue) bool {
+        if (!this.isCell())
+            return false;
+
+        return jsType(this).isStringLike();
     }
     pub fn isBigInt(this: JSValue) bool {
         return cppFn("isBigInt", .{this});
@@ -3968,7 +3978,6 @@ pub const JSValue = enum(JSValueReprInt) {
         "isObject",
         "isPrimitive",
         "isSameValue",
-        "isString",
         "isSymbol",
         "isTerminationException",
         "isUInt32AsAnyInt",

diff --git a/src/bun.js/bindings/headers-cpp.h b/src/bun.js/bindings/headers-cpp.h
@@ -1,4 +1,4 @@
-//-- AUTOGENERATED FILE -- 1677776166
+//-- AUTOGENERATED FILE -- 1678254453
 // clang-format off
 #pragma once
 

diff --git a/src/bun.js/bindings/headers.h b/src/bun.js/bindings/headers.h
@@ -1,5 +1,5 @@
 // clang-format off
-//-- AUTOGENERATED FILE -- 1677776166
+//-- AUTOGENERATED FILE -- 1678254453
 #pragma once
 
 #include <stddef.h>
@@ -331,7 +331,6 @@ CPP_DECL bool JSC__JSValue__isNumber(JSC__JSValue JSValue0);
 CPP_DECL bool JSC__JSValue__isObject(JSC__JSValue JSValue0);
 CPP_DECL bool JSC__JSValue__isPrimitive(JSC__JSValue JSValue0);
 CPP_DECL bool JSC__JSValue__isSameValue(JSC__JSValue JSValue0, JSC__JSValue JSValue1, JSC__JSGlobalObject* arg2);
-CPP_DECL bool JSC__JSValue__isString(JSC__JSValue JSValue0);
 CPP_DECL bool JSC__JSValue__isSymbol(JSC__JSValue JSValue0);
 CPP_DECL bool JSC__JSValue__isTerminationException(JSC__JSValue JSValue0, JSC__VM* arg1);
 CPP_DECL bool JSC__JSValue__isUInt32AsAnyInt(JSC__JSValue JSValue0);

diff --git a/src/bun.js/bindings/headers.zig b/src/bun.js/bindings/headers.zig
@@ -244,7 +244,6 @@ pub extern fn JSC__JSValue__isNumber(JSValue0: JSC__JSValue) bool;
 pub extern fn JSC__JSValue__isObject(JSValue0: JSC__JSValue) bool;
 pub extern fn JSC__JSValue__isPrimitive(JSValue0: JSC__JSValue) bool;
 pub extern fn JSC__JSValue__isSameValue(JSValue0: JSC__JSValue, JSValue1: JSC__JSValue, arg2: *bindings.JSGlobalObject) bool;
-pub extern fn JSC__JSValue__isString(JSValue0: JSC__JSValue) bool;
 pub extern fn JSC__JSValue__isSymbol(JSValue0: JSC__JSValue) bool;
 pub extern fn JSC__JSValue__isTerminationException(JSValue0: JSC__JSValue, arg1: *bindings.VM) bool;
 pub extern fn JSC__JSValue__isUInt32AsAnyInt(JSValue0: JSC__JSValue) bool;

diff --git a/src/bun.js/javascript.zig b/src/bun.js/javascript.zig
@@ -450,6 +450,10 @@ pub const VirtualMachine = struct {
         return VMHolder.vm.?;
     }
 
+    pub fn mimeType(this: *VirtualMachine, str: []const u8) ?bun.HTTP.MimeType {
+        return this.rareData().mimeTypeFromString(this.allocator, str);
+    }
+
     pub const GCLevel = enum(u3) {
         none = 0,
         mild = 1,

diff --git a/src/bun.js/rare_data.zig b/src/bun.js/rare_data.zig
@@ -27,6 +27,18 @@ file_polls_: ?*JSC.FilePoll.HiveArray = null,
 
 global_dns_data: ?*JSC.DNS.GlobalData = null,
 
+mime_types: ?bun.HTTP.MimeType.Map = null,
+
+pub fn mimeTypeFromString(this: *RareData, allocator: std.mem.Allocator, str: []const u8) ?bun.HTTP.MimeType {
+    if (this.mime_types == null) {
+        this.mime_types = bun.HTTP.MimeType.createHashTable(
+            allocator,
+        ) catch @panic("Out of memory");
+    }
+
+    return this.mime_types.?.get(str);
+}
+
 pub fn filePolls(this: *RareData, vm: *JSC.VirtualMachine) *JSC.FilePoll.HiveArray {
     return this.file_polls_ orelse {
         this.file_polls_ = vm.allocator.create(JSC.FilePoll.HiveArray) catch unreachable;

diff --git a/src/bun.js/webcore/blob.zig b/src/bun.js/webcore/blob.zig
@@ -2391,7 +2391,7 @@ pub const Blob = struct {
         callframe: *JSC.CallFrame,
     ) callconv(.C) JSC.JSValue {
         var allocator = globalThis.allocator();
-        var arguments_ = callframe.arguments(2);
+        var arguments_ = callframe.arguments(3);
         var args = arguments_.ptr[0..arguments_.len];
 
         if (this.size == 0) {
@@ -2410,51 +2410,84 @@ pub const Blob = struct {
         // If the optional end parameter is not used as a parameter when making this call, let relativeEnd be size.
         var relativeEnd: i64 = @intCast(i64, this.size);
 
+        if (args.ptr[0].isString()) {
+            args.ptr[2] = args.ptr[0];
+            args.ptr[1] = .zero;
+            args.ptr[0] = .zero;
+            args.len = 3;
+        } else if (args.ptr[1].isString()) {
+            args.ptr[2] = args.ptr[1];
+            args.ptr[1] = .zero;
+            args.len = 3;
+        }
+
         var args_iter = JSC.Node.ArgumentsSlice.init(globalThis.bunVM(), args);
         if (args_iter.nextEat()) |start_| {
-            const start = start_.toInt64();
-            if (start < 0) {
-                // If the optional start parameter is negative, let relativeStart be start + size.
-                relativeStart = @intCast(i64, @max(start + @intCast(i64, this.size), 0));
-            } else {
-                // Otherwise, let relativeStart be start.
-                relativeStart = @min(@intCast(i64, start), @intCast(i64, this.size));
+            if (start_.isNumber()) {
+                const start = start_.toInt64();
+                if (start < 0) {
+                    // If the optional start parameter is negative, let relativeStart be start + size.
+                    relativeStart = @intCast(i64, @max(start +% @intCast(i64, this.size), 0));
+                } else {
+                    // Otherwise, let relativeStart be start.
+                    relativeStart = @min(@intCast(i64, start), @intCast(i64, this.size));
+                }
             }
         }
 
         if (args_iter.nextEat()) |end_| {
-            const end = end_.toInt64();
-            // If end is negative, let relativeEnd be max((size + end), 0).
-            if (end < 0) {
-                // If the optional start parameter is negative, let relativeStart be start + size.
-                relativeEnd = @intCast(i64, @max(end + @intCast(i64, this.size), 0));
-            } else {
-                // Otherwise, let relativeStart be start.
-                relativeEnd = @min(@intCast(i64, end), @intCast(i64, this.size));
+            if (end_.isNumber()) {
+                const end = end_.toInt64();
+                // If end is negative, let relativeEnd be max((size + end), 0).
+                if (end < 0) {
+                    // If the optional start parameter is negative, let relativeStart be start + size.
+                    relativeEnd = @intCast(i64, @max(end +% @intCast(i64, this.size), 0));
+                } else {
+                    // Otherwise, let relativeStart be start.
+                    relativeEnd = @min(@intCast(i64, end), @intCast(i64, this.size));
+                }
             }
         }
 
         var content_type: string = "";
+        var content_type_was_allocated = false;
         if (args_iter.nextEat()) |content_type_| {
-            if (content_type_.isString()) {
-                var zig_str = content_type_.getZigString(globalThis);
-                var slicer = zig_str.toSlice(bun.default_allocator);
-                defer slicer.deinit();
-                var slice = slicer.slice();
-                var content_type_buf = allocator.alloc(u8, slice.len) catch unreachable;
-                content_type = strings.copyLowercase(slice, content_type_buf);
+            inner: {
+                if (content_type_.isString()) {
+                    var zig_str = content_type_.getZigString(globalThis);
+                    var slicer = zig_str.toSlice(bun.default_allocator);
+                    defer slicer.deinit();
+                    var slice = slicer.slice();
+                    if (!strings.isAllASCII(slice)) {
+                        break :inner;
+                    }
+
+                    if (globalThis.bunVM().mimeType(slice)) |mime| {
+                        content_type = mime.value;
+                        break :inner;
+                    }
+
+                    content_type_was_allocated = slice.len > 0;
+                    var content_type_buf = allocator.alloc(u8, slice.len) catch unreachable;
+                    content_type = strings.copyLowercase(slice, content_type_buf);
+                }
             }
         }
 
-        const len = @intCast(SizeType, @max(relativeEnd - relativeStart, 0));
+        const len = @intCast(SizeType, @max(relativeEnd -| relativeStart, 0));
 
         // This copies over the is_all_ascii flag
         // which is okay because this will only be a <= slice
         var blob = this.dupe();
         blob.offset = @intCast(SizeType, relativeStart);
         blob.size = len;
+
+        // infer the content type if it was not specified
+        if (content_type.len == 0 and this.content_type.len > 0 and !this.content_type_allocated)
+            content_type = this.content_type;
+
         blob.content_type = content_type;
-        blob.content_type_allocated = content_type.len > 0;
+        blob.content_type_allocated = content_type_was_allocated;
 
         var blob_ = allocator.create(Blob) catch unreachable;
         blob_.* = blob;
@@ -2474,19 +2507,35 @@ pub const Blob = struct {
         globalThis: *JSC.JSGlobalObject,
         value: JSC.JSValue,
     ) callconv(.C) bool {
-        var zig_str = value.getZigString(globalThis);
-        if (zig_str.is16Bit())
-            return false;
+        var zig_str = if (value.isString())
+            value.getZigString(globalThis)
+        else
+            ZigString.Empty;
+
+        if (!zig_str.isAllASCII()) {
+            zig_str = ZigString.Empty;
+        }
 
-        var slice = zig_str.trimmedSlice();
-        if (strings.eql(slice, this.content_type))
+        if (zig_str.eql(ZigString.init(this.content_type))) {
             return true;
+        }
 
         const prev_content_type = this.content_type;
         {
-            defer if (this.content_type_allocated) bun.default_allocator.free(prev_content_type);
-            var content_type_buf = globalThis.allocator().alloc(u8, slice.len) catch unreachable;
-            this.content_type = strings.copyLowercase(slice, content_type_buf);
+            var slicer = zig_str.toSlice(bun.default_allocator);
+            defer slicer.deinit();
+            const allocated = this.content_type_allocated;
+            defer if (allocated) bun.default_allocator.free(prev_content_type);
+            if (globalThis.bunVM().mimeType(slicer.slice())) |mime| {
+                this.content_type = mime.value;
+                this.content_type_allocated = false;
+                return true;
+            }
+            var content_type_buf = globalThis.allocator().alloc(u8, slicer.len) catch {
+                globalThis.throwOutOfMemory();
+                return false;
+            };
+            this.content_type = strings.copyLowercase(slicer.slice(), content_type_buf);
         }
 
         this.content_type_allocated = true;
@@ -2602,13 +2651,22 @@ pub const Blob = struct {
                         // Normative conditions for this member are provided
                         // in the § 3.1 Constructors.
                         if (options.get(globalThis, "type")) |content_type| {
-                            if (content_type.isString()) {
-                                var content_type_str = content_type.toSlice(globalThis, bun.default_allocator);
-                                defer content_type_str.deinit();
-                                var slice = content_type_str.slice();
-                                var content_type_buf = allocator.alloc(u8, slice.len) catch unreachable;
-                                blob.content_type = strings.copyLowercase(slice, content_type_buf);
-                                blob.content_type_allocated = true;
+                            inner: {
+                                if (content_type.isString()) {
+                                    var content_type_str = content_type.toSlice(globalThis, bun.default_allocator);
+                                    defer content_type_str.deinit();
+                                    var slice = content_type_str.slice();
+                                    if (!strings.isAllASCII(slice)) {
+                                        break :inner;
+                                    }
+                                    if (globalThis.bunVM().mimeType(slice)) |mime| {
+                                        blob.content_type = mime.value;
+                                        break :inner;
+                                    }
+                                    var content_type_buf = allocator.alloc(u8, slice.len) catch unreachable;
+                                    blob.content_type = strings.copyLowercase(slice, content_type_buf);
+                                    blob.content_type_allocated = true;
+                                }
                             }
                         }
                     }