mirror of
https://github.com/denoland/deno.git
synced 2024-11-01 09:24:20 -04:00
408 lines
11 KiB
Rust
408 lines
11 KiB
Rust
// Copyright 2018-2022 the Deno authors. All rights reserved. MIT license.
|
|
|
|
mod blob;
|
|
mod compression;
|
|
mod message_port;
|
|
mod timers;
|
|
|
|
use deno_core::error::range_error;
|
|
use deno_core::error::type_error;
|
|
use deno_core::error::AnyError;
|
|
use deno_core::include_js_files;
|
|
use deno_core::op;
|
|
use deno_core::url::Url;
|
|
use deno_core::ByteString;
|
|
use deno_core::Extension;
|
|
use deno_core::OpState;
|
|
use deno_core::Resource;
|
|
use deno_core::ResourceId;
|
|
use deno_core::U16String;
|
|
use deno_core::ZeroCopyBuf;
|
|
use encoding_rs::CoderResult;
|
|
use encoding_rs::Decoder;
|
|
use encoding_rs::DecoderResult;
|
|
use encoding_rs::Encoding;
|
|
use serde::Deserialize;
|
|
use serde::Serialize;
|
|
use std::borrow::Cow;
|
|
use std::cell::RefCell;
|
|
use std::fmt;
|
|
use std::path::PathBuf;
|
|
use std::usize;
|
|
|
|
use crate::blob::op_blob_create_object_url;
|
|
use crate::blob::op_blob_create_part;
|
|
use crate::blob::op_blob_from_object_url;
|
|
use crate::blob::op_blob_read_part;
|
|
use crate::blob::op_blob_remove_part;
|
|
use crate::blob::op_blob_revoke_object_url;
|
|
use crate::blob::op_blob_slice_part;
|
|
pub use crate::blob::Blob;
|
|
pub use crate::blob::BlobPart;
|
|
pub use crate::blob::BlobStore;
|
|
pub use crate::blob::InMemoryBlobPart;
|
|
|
|
pub use crate::message_port::create_entangled_message_port;
|
|
use crate::message_port::op_message_port_create_entangled;
|
|
use crate::message_port::op_message_port_post_message;
|
|
use crate::message_port::op_message_port_recv_message;
|
|
pub use crate::message_port::JsMessageData;
|
|
pub use crate::message_port::MessagePort;
|
|
|
|
use crate::timers::op_now;
|
|
use crate::timers::op_sleep;
|
|
use crate::timers::op_sleep_sync;
|
|
use crate::timers::op_timer_handle;
|
|
use crate::timers::StartTime;
|
|
pub use crate::timers::TimersPermission;
|
|
|
|
/// Load and execute the javascript code.
|
|
pub fn init<P: TimersPermission + 'static>(
|
|
blob_store: BlobStore,
|
|
maybe_location: Option<Url>,
|
|
) -> Extension {
|
|
Extension::builder()
|
|
.js(include_js_files!(
|
|
prefix "deno:ext/web",
|
|
"00_infra.js",
|
|
"01_dom_exception.js",
|
|
"01_mimesniff.js",
|
|
"02_event.js",
|
|
"02_structured_clone.js",
|
|
"02_timers.js",
|
|
"03_abort_signal.js",
|
|
"04_global_interfaces.js",
|
|
"05_base64.js",
|
|
"06_streams.js",
|
|
"08_text_encoding.js",
|
|
"09_file.js",
|
|
"10_filereader.js",
|
|
"11_blob_url.js",
|
|
"12_location.js",
|
|
"13_message_port.js",
|
|
"14_compression.js",
|
|
"15_performance.js",
|
|
))
|
|
.ops(vec![
|
|
op_base64_decode::decl(),
|
|
op_base64_encode::decl(),
|
|
op_base64_atob::decl(),
|
|
op_base64_btoa::decl(),
|
|
op_encoding_normalize_label::decl(),
|
|
op_encoding_new_decoder::decl(),
|
|
op_encoding_decode::decl(),
|
|
op_encoding_encode_into::decl(),
|
|
op_blob_create_part::decl(),
|
|
op_blob_slice_part::decl(),
|
|
op_blob_read_part::decl(),
|
|
op_blob_remove_part::decl(),
|
|
op_blob_create_object_url::decl(),
|
|
op_blob_revoke_object_url::decl(),
|
|
op_blob_from_object_url::decl(),
|
|
op_message_port_create_entangled::decl(),
|
|
op_message_port_post_message::decl(),
|
|
op_message_port_recv_message::decl(),
|
|
compression::op_compression_new::decl(),
|
|
compression::op_compression_write::decl(),
|
|
compression::op_compression_finish::decl(),
|
|
op_now::decl::<P>(),
|
|
op_timer_handle::decl(),
|
|
op_sleep::decl(),
|
|
op_sleep_sync::decl::<P>(),
|
|
])
|
|
.state(move |state| {
|
|
state.put(blob_store.clone());
|
|
if let Some(location) = maybe_location.clone() {
|
|
state.put(Location(location));
|
|
}
|
|
state.put(StartTime::now());
|
|
Ok(())
|
|
})
|
|
.build()
|
|
}
|
|
|
|
#[op]
|
|
fn op_base64_decode(input: String) -> Result<ZeroCopyBuf, AnyError> {
|
|
let mut input = input.into_bytes();
|
|
input.retain(|c| !c.is_ascii_whitespace());
|
|
Ok(b64_decode(&input)?.into())
|
|
}
|
|
|
|
#[op]
|
|
fn op_base64_atob(mut s: ByteString) -> Result<ByteString, AnyError> {
|
|
s.retain(|c| !c.is_ascii_whitespace());
|
|
|
|
// If padding is expected, fail if not 4-byte aligned
|
|
if s.len() % 4 != 0 && (s.ends_with(b"==") || s.ends_with(b"=")) {
|
|
return Err(
|
|
DomExceptionInvalidCharacterError::new("Failed to decode base64.").into(),
|
|
);
|
|
}
|
|
|
|
Ok(b64_decode(&s)?.into())
|
|
}
|
|
|
|
fn b64_decode(input: &[u8]) -> Result<Vec<u8>, AnyError> {
|
|
// "If the length of input divides by 4 leaving no remainder, then:
|
|
// if input ends with one or two U+003D EQUALS SIGN (=) characters,
|
|
// remove them from input."
|
|
let input = match input.len() % 4 == 0 {
|
|
true if input.ends_with(b"==") => &input[..input.len() - 2],
|
|
true if input.ends_with(b"=") => &input[..input.len() - 1],
|
|
_ => input,
|
|
};
|
|
|
|
// "If the length of input divides by 4 leaving a remainder of 1,
|
|
// throw an InvalidCharacterError exception and abort these steps."
|
|
if input.len() % 4 == 1 {
|
|
return Err(
|
|
DomExceptionInvalidCharacterError::new("Failed to decode base64.").into(),
|
|
);
|
|
}
|
|
|
|
let cfg = base64::Config::new(base64::CharacterSet::Standard, true)
|
|
.decode_allow_trailing_bits(true);
|
|
let out = base64::decode_config(input, cfg).map_err(|err| match err {
|
|
base64::DecodeError::InvalidByte(_, _) => {
|
|
DomExceptionInvalidCharacterError::new(
|
|
"Failed to decode base64: invalid character",
|
|
)
|
|
}
|
|
_ => DomExceptionInvalidCharacterError::new(&format!(
|
|
"Failed to decode base64: {:?}",
|
|
err
|
|
)),
|
|
})?;
|
|
|
|
Ok(out)
|
|
}
|
|
|
|
#[op]
|
|
fn op_base64_encode(s: ZeroCopyBuf) -> Result<String, AnyError> {
|
|
Ok(b64_encode(&s))
|
|
}
|
|
|
|
#[op]
|
|
fn op_base64_btoa(s: ByteString) -> Result<String, AnyError> {
|
|
Ok(b64_encode(s))
|
|
}
|
|
|
|
fn b64_encode(s: impl AsRef<[u8]>) -> String {
|
|
let cfg = base64::Config::new(base64::CharacterSet::Standard, true)
|
|
.decode_allow_trailing_bits(true);
|
|
base64::encode_config(s.as_ref(), cfg)
|
|
}
|
|
|
|
#[derive(Deserialize)]
|
|
#[serde(rename_all = "camelCase")]
|
|
struct DecoderOptions {
|
|
label: String,
|
|
ignore_bom: bool,
|
|
fatal: bool,
|
|
}
|
|
|
|
#[op]
|
|
fn op_encoding_normalize_label(label: String) -> Result<String, AnyError> {
|
|
let encoding = Encoding::for_label_no_replacement(label.as_bytes())
|
|
.ok_or_else(|| {
|
|
range_error(format!(
|
|
"The encoding label provided ('{}') is invalid.",
|
|
label
|
|
))
|
|
})?;
|
|
Ok(encoding.name().to_lowercase())
|
|
}
|
|
|
|
#[op]
|
|
fn op_encoding_new_decoder(
|
|
state: &mut OpState,
|
|
options: DecoderOptions,
|
|
) -> Result<ResourceId, AnyError> {
|
|
let DecoderOptions {
|
|
label,
|
|
fatal,
|
|
ignore_bom,
|
|
} = options;
|
|
|
|
let encoding = Encoding::for_label(label.as_bytes()).ok_or_else(|| {
|
|
range_error(format!(
|
|
"The encoding label provided ('{}') is invalid.",
|
|
label
|
|
))
|
|
})?;
|
|
|
|
let decoder = if ignore_bom {
|
|
encoding.new_decoder_without_bom_handling()
|
|
} else {
|
|
encoding.new_decoder_with_bom_removal()
|
|
};
|
|
|
|
let rid = state.resource_table.add(TextDecoderResource {
|
|
decoder: RefCell::new(decoder),
|
|
fatal,
|
|
});
|
|
|
|
Ok(rid)
|
|
}
|
|
|
|
#[derive(Deserialize)]
|
|
#[serde(rename_all = "camelCase")]
|
|
struct DecodeOptions {
|
|
rid: ResourceId,
|
|
stream: bool,
|
|
}
|
|
|
|
#[op]
|
|
fn op_encoding_decode(
|
|
state: &mut OpState,
|
|
data: ZeroCopyBuf,
|
|
options: DecodeOptions,
|
|
) -> Result<U16String, AnyError> {
|
|
let DecodeOptions { rid, stream } = options;
|
|
|
|
let resource = state.resource_table.get::<TextDecoderResource>(rid)?;
|
|
|
|
let mut decoder = resource.decoder.borrow_mut();
|
|
let fatal = resource.fatal;
|
|
|
|
let max_buffer_length = decoder
|
|
.max_utf16_buffer_length(data.len())
|
|
.ok_or_else(|| range_error("Value too large to decode."))?;
|
|
|
|
let mut output = vec![0; max_buffer_length];
|
|
|
|
if fatal {
|
|
let (result, _, written) =
|
|
decoder.decode_to_utf16_without_replacement(&data, &mut output, !stream);
|
|
match result {
|
|
DecoderResult::InputEmpty => {
|
|
output.truncate(written);
|
|
Ok(output.into())
|
|
}
|
|
DecoderResult::OutputFull => {
|
|
Err(range_error("Provided buffer too small."))
|
|
}
|
|
DecoderResult::Malformed(_, _) => {
|
|
Err(type_error("The encoded data is not valid."))
|
|
}
|
|
}
|
|
} else {
|
|
let (result, _, written, _) =
|
|
decoder.decode_to_utf16(&data, &mut output, !stream);
|
|
match result {
|
|
CoderResult::InputEmpty => {
|
|
output.truncate(written);
|
|
Ok(output.into())
|
|
}
|
|
CoderResult::OutputFull => Err(range_error("Provided buffer too small.")),
|
|
}
|
|
}
|
|
}
|
|
|
|
struct TextDecoderResource {
|
|
decoder: RefCell<Decoder>,
|
|
fatal: bool,
|
|
}
|
|
|
|
impl Resource for TextDecoderResource {
|
|
fn name(&self) -> Cow<str> {
|
|
"textDecoder".into()
|
|
}
|
|
}
|
|
|
|
#[derive(Serialize)]
|
|
#[serde(rename_all = "camelCase")]
|
|
struct EncodeIntoResult {
|
|
read: usize,
|
|
written: usize,
|
|
}
|
|
|
|
#[op]
|
|
fn op_encoding_encode_into(
|
|
input: String,
|
|
mut buffer: ZeroCopyBuf,
|
|
) -> Result<EncodeIntoResult, AnyError> {
|
|
// Since `input` is already UTF-8, we can simply find the last UTF-8 code
|
|
// point boundary from input that fits in `buffer`, and copy the bytes up to
|
|
// that point.
|
|
let boundary = if buffer.len() >= input.len() {
|
|
input.len()
|
|
} else {
|
|
let mut boundary = buffer.len();
|
|
|
|
// The maximum length of a UTF-8 code point is 4 bytes.
|
|
for _ in 0..4 {
|
|
if input.is_char_boundary(boundary) {
|
|
break;
|
|
}
|
|
debug_assert!(boundary > 0);
|
|
boundary -= 1;
|
|
}
|
|
|
|
debug_assert!(input.is_char_boundary(boundary));
|
|
boundary
|
|
};
|
|
|
|
buffer[..boundary].copy_from_slice(input[..boundary].as_bytes());
|
|
|
|
Ok(EncodeIntoResult {
|
|
// The `read` output parameter is measured in UTF-16 code units.
|
|
read: input[..boundary].encode_utf16().count(),
|
|
written: boundary,
|
|
})
|
|
}
|
|
|
|
pub fn get_declaration() -> PathBuf {
|
|
PathBuf::from(env!("CARGO_MANIFEST_DIR")).join("lib.deno_web.d.ts")
|
|
}
|
|
|
|
#[derive(Debug)]
|
|
pub struct DomExceptionQuotaExceededError {
|
|
pub msg: String,
|
|
}
|
|
|
|
impl DomExceptionQuotaExceededError {
|
|
pub fn new(msg: &str) -> Self {
|
|
DomExceptionQuotaExceededError {
|
|
msg: msg.to_string(),
|
|
}
|
|
}
|
|
}
|
|
|
|
#[derive(Debug)]
|
|
pub struct DomExceptionInvalidCharacterError {
|
|
pub msg: String,
|
|
}
|
|
|
|
impl DomExceptionInvalidCharacterError {
|
|
pub fn new(msg: &str) -> Self {
|
|
DomExceptionInvalidCharacterError {
|
|
msg: msg.to_string(),
|
|
}
|
|
}
|
|
}
|
|
|
|
impl fmt::Display for DomExceptionQuotaExceededError {
|
|
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
|
|
f.pad(&self.msg)
|
|
}
|
|
}
|
|
impl fmt::Display for DomExceptionInvalidCharacterError {
|
|
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
|
|
f.pad(&self.msg)
|
|
}
|
|
}
|
|
|
|
impl std::error::Error for DomExceptionQuotaExceededError {}
|
|
|
|
impl std::error::Error for DomExceptionInvalidCharacterError {}
|
|
|
|
pub fn get_error_class_name(e: &AnyError) -> Option<&'static str> {
|
|
e.downcast_ref::<DomExceptionQuotaExceededError>()
|
|
.map(|_| "DOMExceptionQuotaExceededError")
|
|
.or_else(|| {
|
|
e.downcast_ref::<DomExceptionInvalidCharacterError>()
|
|
.map(|_| "DOMExceptionInvalidCharacterError")
|
|
})
|
|
}
|
|
pub struct Location(pub Url);
|