This commit fixes hang in web workers occuring when sending
"undefined" as message value. It is a temporary band-aid
until proper structured close is implemented.
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
This commit adds new binary target called "denort".
It is a "lite" version of "deno" binary that can only execute
code embedded inside the binary itself.
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
When we were doing single process in-memory reports getting the source
from the runtime was practical, but now that we're writing to disk this
conflicts with the format tools taking raw v8 coverage dumps expect.
This commit adds new option to "Worker" Web API that allows to
configure permissions.
New "Worker.deno.permissions" option can be used to define limited
permissions to the worker thread by either:
- inherit set of parent thread permissions
- use limited subset of parent thread permissions
- revoke all permissions (full sandbox)
In order to achieve this functionality "CliModuleLoader"
was modified to accept "initial permissions", which are used
for top module loading (ie. uses parent thread permission set
to load top level module of a worker).
This commit removes "js" module from "cli".
It contained stuff related to TypeScript compiler (snapshot,
declaration files) and thus it was moved to "tsc" module.
This commit rewrites initialisation of the "shared queue" and
in effect prevents from double execution of "core/core.js" and
"core/error.js".
Previously both of these files were executed every time a "JsRuntime"
was created. That lead to a situation where one copy of each script
was included in the snapshot and then another copy would be
executed after loading the snapshot.
Effectively "JsRuntime::shared_init" was removed; instead execution
of those scripts and actual initialisation of shared queue
was split into two helper functions: "JsRuntime::js_init" and
"JsRuntime::share_queue_init".
Additionally stale TODO comments were removed.
Merging multiple runs isn't quite right because we
rely on a 0 count to signal that a block hasn't been called.
Other tools like c8 expect this to be true as-well so we
need to do our best to merge coverage files rather
than duplicating them.
The child process kept running and printing "hello" to stdout.
This commit also removes the dependency on reqwest and instead
switches to the re-export from the fetch crate.
Brings back commit 1a2e7741c3.
This commit adds back "/json/list" endpoint to
inspector server which was erroneously removed
during server rewrite.
Co-authored-by: Ben Noordhuis <info@bnoordhuis.nl>
Fix bug in regular expression and make the regular expression more
strict.
In a string passed to new RegExp(), '[\t\s]' is identical to '[ts]' and
not `/[\t\s]/`. For that, the backslash needs to be escaped in the
string. Futhermore, `\t` is the tab character and is included in the
special regexp value `\s` so is unnecessary.
That would reduce the RegExp to new RegExp(`^${value}\\s*;?`) but
there's no point in matching 0 or more space characters followed by 0 or
one semi-colons as that will match no matter what follows `value`.
To make it more strict, require one of space, semicolon, or
end-of-string after value.
Allowlist checking already uses hosts but for some reason
requests, revokes and the runtime permissions API use URLs.
- BREAKING(lib.deno.unstable.d.ts): Change
NetPermissionDescriptor::url to NetPermissionDescriptor::host
- fix(runtime/permissions): Don't add whole URLs to the
allowlist on request
- fix(runtime/permissions): Harden strength semantics:
({ name: "net", host: "127.0.0.1" } is stronger than
{ name: "net", host: "127.0.0.1:8000" }) for blocklisting
- refactor(runtime/permissions): Use tuples for hosts, make
the host optional in Permissions::{query_net, request_net, revoke_net}()
This PR refactors "cli/flags.rs" and "runtime/permissions.rs" so
that "allow_read", "allow_write" and "allow_net" themselves
have allowlists, instead of storing them in additional fields.
This commit rewrites "chown_test.ts" to use the GNU "id" command
instead of python. This won't work on Windows, but these tests aren't
currently run on Windows anyway.
Implementors of `deno_core::JsRuntime` might want to do additional actions
during each turn of event loop, eg. `deno_runtime::Worker` polls inspector,
`deno_runtime::WebWorker` receives/dispatches messages from/to worker host.
Previously `JsRuntime::mod_evaluate` was implemented in such fashion that it
only polled `JsRuntime`'s event loop. This behavior turned out to be wrong
in the example of `WebWorker` which couldn't receive/dispatch messages because
its implementation of event loop was never called.
This commit rewrites "mod_evaluate" to return a handle to receiver that resolves
when module's promise resolves. It is now implementors responsibility to poll
event loop after calling `mod_evaluate`.
This commit migrates all ops to use new resource table
and "AsyncRefCell".
Old implementation of resource table was completely
removed and all code referencing it was updated to use
new system.
This commit moves Deno JS runtime, ops, permissions and
inspector implementation to new "deno_runtime" crate located
in "runtime/" directory.
Details in "runtime/README.md".
Co-authored-by: Ryan Dahl <ry@tinyclouds.org>
This commit does major refactor of "Worker" and "WebWorker",
in order to decouple them from "ProgramState" and "Flags".
The main points of interest are "create_main_worker()" and
"create_web_worker_callback()" functions which are responsible
for creating "Worker" and "WebWorker" in CLI context.
As a result it is now possible to factor out common "runtime"
functionality into a separate crate.
This commit adds "Deno.core.createPrepareStackTrace". This function
was moved from "cli/rt/40_error_stack.js" to unify handling of stack frames in core
(before this PR there was implicit dependency on logic in "core/error.rs::JsError").
Unfortunately formatting logic must still be duplicated in "cli/error.js::PrettyJsError"
to provide coloring, but currently there's no solution to this problem.
"createPrepareStackTrace" can accept a single argument; a function that takes
a location and provides source mapped location back.
This commit disables source mapping of errors
for standalone binaries. Since applying source
maps relies on using file fetcher infrastructure
it's not feasible to use it for standalone binaries
that are not supposed to use that infrastructure.
This commit makes the file watcher continue to work even if module
resolution fails at the initial attempt, allowing us to execute `run`
or `bundle` subcommand when a script has invalid syntax. In such
cases, the watcher observes a single file that is specified as an
command line argument.
Fixes panic occurring in worker when "self.close()" is called
at the top level, ie. worker shuts down while
module evaluation promise hasn't yet resolved.
This commit fixes order of events for "onload" event.
Previously handler attached using "window.onload" was
always fired before handlers added using "addEventListener".
Factored out "init_v8_flags", "init_logger" and
"get_subcommand" from "main" function.
Also "Worker" was removed in favor of moving
logic to "MainWorker" and "WebWorker" respectively.
This commit removes unnecessary conversion into BTreeMap. The value
that gets converted into BTreeMap is _originally_ BTreeMap, so this
conversion is just superfluous.
Additionally, a test is added so that we can make sure the keys in the
emitted lockfile are sorted alphabetically.
This commit adds support for "--watch" flag for "bundle"
and "fmt" subcommands.
In addition to this, it refactors "run --watch" command so that
module resolution will occur every time the file watcher detects
file addition/deletion, which allows the watcher to observe a file
that is newly added to the dependency as well.
This commit does major overhaul of "upgrade" subcommand,
reducing complexity & giving more sensible console output.
Removes gz support for archives.
Uses last part of url instead of scraping to get latest version.
This commit simplifies code responsible for serialization
of web worker events. Instead of using "PrettyJsError"
it's enough to downcast to "deno_core::JsError", making
the code less specific to cli/ implementation.
This commit renames "fmt_errors::JsError" to "PrettyJsError"
to avoid confusion with "deno_core::JsError".
Consequently "CoreJsError" aliases to "deno_core::JsError"
were removed.
Additionally source mapping step has been removed from
"PrettyJsError::create" to better separate domains.
This commit moves following tools into a single "tools"
module located at "cli/tools/mod.rs":
- formatter
- linter
- test runner
- coverage collector
- installer
- binary upgrader
- repl
This commit renames "fs" module in "cli/" to "fs_util". This is purely
cosmetic change; there were a few places which aliased "crate::fs"
to "deno_fs" which was very confusing with "fs" module in ops.
This commit merges implementations of "collect_files" and "files_in_subtree",
leaving only the former. Additionally it was ensured that directories are not yielded
from this function.
This commit reworks "collect_files" utility to accept "ignore" parameter
which allows to filter out files in a single iteration instead of walking
file tree second time to excude "ignored" files.
This commit rewrites scripts in "tools/" directory
to use Deno instead of Python. In return it allows
to remove huge number of Python packages in "third_party/".
This commit removes ProgramState::permissions field.
Having permissions parsed from CLI flags stored on globally
accessible state object made it easy to mistakenly use these
permissions in situations which required "runtime" permissions.
This commit rewrites file watcher used with --watch flag.
Instead of creating new watcher after each restart, only a single
watcher is created for whole process. Additionally debouncing
mechanism has been added to prevent infinite restart loops
if multiple files were changed in quick succession.
Co-authored-by: bartossh <lenart.consulting@gmail.com>
Previously, entering a single ']' would cause repl to forever accepting
new lines, due to that `ValidationResult::Invalid` would actually be
consumed by the editor itself while continue building the lines. Instead
we should mark it as `Valid` and send the bad input for evaluation to
get the proper error from V8.
Before:
```
> ]
(you can keep entering new line here, and it will never consume input
until you Ctrl-C)
```
After:
```
> ]
Uncaught SyntaxError: Unexpected token ']'
>
```
This commit renames two assertion functions to better align with JS API:
- assertStringContains -> assertStringIncludes
- assertArrayContains -> assertArrayIncludes
This changes the behavior of keyboard interrupts (ctrl+c) to continue,
clearing the current line instead of exiting.
Exit can still be done with ctrl+d or by calling close().
This writes all evaluaton results to stdout regardless if the result is
an error or not.
This matches the behavior of other read-eval-print-loops like Node.
This commit fixes test runner by awaitning "Deno.runTests()" call,
which ensures proper error is returned when there's an unresolved
promise that's being awaited.
This extracts prelude injection into a helper function and moves the
prelude string literal into a top level static string to help trim some
of the fat out of the run function.
This commit changes implementation of top-level-await in "deno_core".
Previously promise returned from module evaluation was not awaited,
leading to out-of-order execution of modules that have TLA. It's been
fixed by changing "JsRuntime::mod_evaluate" to be an async function
that resolves when the promise returned from module evaluation also
resolves. When waiting for promise resolution event loop is polled
repeatedly, until there are no more dynamic imports or pending
ops.
This commit fixes the inspection of functions. The current
implementation gets the name of the type of the function
from "f.__proto__.constructor.name", and it throws when
the prototype is set to null.
This commit checks the prototype before accessing its
constructor name and uses the generic name 'Function'
if the prototype is not available.
* Revert "refactor: Worker is not a Future (#7895)"
This reverts commit f4357f0ff9.
* Revert "refactor(core): JsRuntime is not a Future (#7855)"
This reverts commit d8879feb8c.
* Revert "fix(core): module execution with top level await (#7672)"
This reverts commit c7c7677825.
This commit rewrites deno::Worker to not implement Future
trait.
Instead there are two separate methods:
- Worker::poll_event_loop() - does single tick of event loop
- Worker::run_event_loop() - runs event loop to completion
Additionally some cleanup to Worker's field visibility was done.
This commit rewrites deno_core::JsRuntime to not implement Future
trait.
Instead there are two separate methods:
- JsRuntime::poll_event_loop() - does single tick of event loop
- JsRuntime::run_event_loop() - runs event loop to completion
This commit renames occurrences of "isolate" variable name
to "js_runtime". This was outstanding debt after renaming
deno_core::CoreIsolate to JsRuntime.
This changes the signature of InspectorSession.post_message to take a
&str rather than a String avoiding the need call str.to_string at each
call site.
This commit fixes implementation of top level await in "deno_core".
Previously promise returned from module execution was ignored causing to execute
modules out-of-order.
With this commit promise returned from module execution is stored on "JsRuntime"
and event loop is polled until the promise resolves.
This makes use of a default referrer when its empty in repl mode so that
dynamic imports work in the global evaluation context.
Co-authored-by: Bartek Iwanczuk <biwanczuk@gmail.com>
This ports the REPL over to Rust and makes use of an inspector session to run a REPL on top of any isolate which lets make full use of rustylines various things like validators and completors without having to introduce a bunch of hard to test internal ops and glue code.
An accidental but good side effect of this is that the multiple line input we previously had is now an editable multi-line input prompt that is correctly stored in the history as a single entry.
This extracts the inspector session specific bits from CoverageCollector
into a standalone struct dubbed `InspectorSesssion` which can be used as
a general purpose session to communicate with the inspector on the same
thread as an isolate/inspector pair lives on.
Previously when we used the websocket to talk to the inspector we used
the runtime domain to send a "runIfWaitingForDebugger" message.
However this is not required since we now talk to the inspector directly
and no longer send said message so this removes the enabling of the
domain entirely.
This commits removes "CompilerWorker" in favor of
using "JsRuntime".
"cli/ops/compiler.rs" has been removed in favor of inline
registration of ops in "cli/tsc.rs"
This commit removes cli/tsc/40_error_stack.js as it is not
needed in TSC host. All errors originating in TSC are terminal
and don't require source mapping hence we can rely on default
stack traces provided by deno_core.
Additionally tsc/06_util.js was removed and its code moved
to tsc/99_main_compiler.js
This makes inspector registration with the server with optional and
explicit to allow for inspectors to exist without spinning up the
server.
As a side effect of the server being explicitly passed around it also
makes it possible to bind multiple servers.
This commit adds support for stack traces in "deno_core".
Implementation of "Display" trait for "JsError" has been updated
and in consequence "deno_core::js_check" became obsolete and
removed.
This removes the dependency on global state and instead relies on the
runtime's internal state to get the script sources it saw when it
collected code coverage for them.
- remove "CliState.workers" and "CliState.next_worker_id", instead
store them on "OpState" using type aliases.
- remove "CliState.global_timer" and "CliState.start_time", instead
store them on "OpState" using type aliases.
- remove "CliState.is_internal", instead pass it to Worker::new
- move "CliState::permissions" to "OpState"
- move "CliState::main_module" to "OpState"
- move "CliState::global_state" to "OpState"
- move "CliState::check_unstable()" to "GlobalState"
- change "cli_state()" to "global_state()"
- change "deno_core::ModuleLoader" trait to pass "OpState" to callbacks
- rename "CliState" to "CliModuleLoader"
Moves op_close and op_resources to deno_core::ops and exports them.
Adds serde dependency to deno_core and reexports it.
Moves JS implementation of those ops to Deno.core and reexports them in Deno.
This reverts the changes introduced by PR #5276, which made
`deno install «script»` automatically bundle the script's dependencies.
It broke the `deno install` command for a large number of scripts.
This reverts commit 34e98fa59c.
Closes: #7492
This commit adds basic support for collecting coverage
data using "deno test".
Currently the report is only a text added to the end
of output from "deno test".
Removes:
- "deno_core::StartupData"
- "deno_core::Script"
- "deno_core::OwnedScript"
Changes to "JsRuntime":
- remove "new_with_loader()"
- remove "with_heap_limits()"
- rename "IsolateOptions" to "RuntimeOptions" and make public
- "JsRuntime::new()" takes "RuntimeOptions" as a single param
Provides a concrete state type that can be dynamically added. This is necessary for op crates.
* renames BasicState to OpState
* async ops take `Rc<RefCell<OpState>>`
* sync ops take `&mut OpState`
* removes `OpRegistry`, `OpRouter` traits
* `get_error_class_fn` moved to OpState
* ResourceTable moved to OpState
This commit fixes panic caused by "unimplemented!()"
calls for some variants of "nix::errno::Errno".
Catch-all variant now returns "Error" class name
instead of panicking.
Co-authored-by: Bert Belder <bertbelder@gmail.com>
This commit changes "deno info" subcommand logic.
- Modules are no longer loaded into V8 isolate - analysis
is done using ModuleGraph.
- Removed deno_core::Deps structure.
- Modules are no longer type-checked and transpiled -
"compiled" file is shown only if it is already available.
- Added number of unique dependencies for root module.
- Changed tree output:
- file size is shown next to the dependency
- repeated dependencies are marked with "*"
- used less spaces in prefix to save terminal width
deno_core/
- rename core_isolate.rs to runtime.rs
- rename CoreIsolate to JsRuntime
- rename JSError to JsError
- rename JSStackFrame to JsStackFrame
cli/
- update references from deno_core::CoreIsolate to deno_core::JsRuntime
- rename deno_core::JSError to deno_core::JsError
- rename fmt_errors::JSError to fmt_errors::JsError
This commit adds "EmitTranspileOptions" to "transpile()" function,
that allows to configure transpilation process based on the
currently loaded "tsconfig.json".
All benchmarks are done in Rust and can be invoked with
`cargo bench`.
Currently this has it's own "harness" that behaves like
`./tools/benchmark.py` did.
Because of this tests inside `cli/bench` are currently not run.
This should be switched to the language provided harness
once the `#[bench]` attribute has been stabilized.
This ensures that all spaces are set to be "+" in the string rather than
just the first and brings deno into line with how browsers handle spaces
in URLSearchParams, see #7001.
The following crates were _not_ upgraded to avoid having multiple
versions of the same crate in the dependency tree:
* tokio-tungstenite v0.10.1 -> v0.11.0
* swc_common v0. 8.0 -> v0. 9.1
* swc_ecmascript v0. 1.0 -> v0. 3.0
* webpki-roots v0.19.0 -> v0.20.0
* nix v0.17.0 -> v0.18.0
- Fix protocol regex.
- Truncate repeated leading slashes in file paths.
- Make drive letter support platform-independent.
- Drop the hostname if a drive letter is parsed.
- Fix drive letter normalization and basing.
- Allow basing over the host.
- Fix same-protocol basing.
- Remove Windows UNC path support.
- Reverts #6418. This is non-standard. Wouldn't be too much of a problem but it
makes other parts of the spec hard to realize.
This PR is intentionally ugly. It duplicates all of the code in cli/js2/ into
cli/tsc/ ... because it's very important that we all understand that this code
is unnecessarily duplicated in our binary. I hope this ugliness provides the
motivation to clean it up.
The typescript git submodule is removed, because it's a very large repo and
contains all sorts of stuff we don't need. Instead the necessary files are
copied directly into the deno repo. Hence +200k lines.
COMPILER_SNAPSHOT.bin size
```
master 3448139
this branch 3320972
```
Fixes #6812
This commit adds a fallback mechanism for absent compiled source file.
Because imported type declaration files are not emitted by TS compiler
and their imports are not elided users often hit "No such file or directory"
error. With this commit in such situation an empty source file will be
provided to V8 with a warning to the user suggesting using "import type"/
"export type" syntax instead.
This commits adds support for "--config" flag in "deno install"
subcommand. Specified configuration file is copied alongside
source code to installation directory.
- Add more support for generics
- Add the --private flag - displays documentation for
not exported and private nodes
- Display more attributes like abstract, static and readonly
- Display type aliases
- Refactor module to use the Display trait
- Use a bit more color
This PR removes direct dependency on dprint-plugin-typescript
and swc_ecma_visit. Both of these dependencies have been
pushed out to deno_lint crate. This should make it a bit easier
to do the upgrades and prevent having conflicting crate requirements.
This commit adds a "--no-check" option to following subcommands:
- "deno cache"
- "deno info"
- "deno run"
- "deno test"
The "--no-check" options allows to skip type checking step and instead
directly transpiles TS sources to JS sources.
This solution uses `ts.transpileModule()` API and is just an interim
solution before implementing it fully in Rust.
Currently, the documentation makes it sound like the test subcommand's filter
flag could accept some kind of pattern matching value like a glob or a regex,
although the function "createFilterFn" accepts a regex as an argument, there's
no way to pass an actual regex value from the CLI.
This commit makes it possible to pass a string that could be cast as regex
when string matches "^/.*/$".
With this change, a user can use the filter flag as follow:
deno test --filter "/test-.+/"
Also tested that `\` get escaped properly, on MacOS at least, and this is
also a valid flag:
deno test --filter "/test-\d+/"
This commit adds incremental compilation capabilities to internal TS compiler.
Instead of using "ts.createProgram()" API for compilation step (during deno
startup), "ts.createIncrementalProgram()" API is used instead.
Thanks to TS' ".tsbuildinfo" file that already stores all necessary metadata
for compilation I was able to remove our own invention that is ".graph" file.
".tsbuildinfo" file is stored alongside compiled source and is used to
cache-bust outdated dependencies, facilitated by the "version" field.
The value for "version" field is computed in Rust during loading of module
graph and is basically a hash of the file contents.
Please keep in mind that incremental compilation is only used for initial
compilation (or dynamic imports compilation) - bundling and runtime compiler
APIs haven't been changed at all.
Due to problems with source map I changed compilation settings to inline
source map (inlineSourceMap instead of sourceMap).
The following used to fail in Deno despite working in the browser:
```javascript
new Request('http://localhost/', {method: 'POST', body: new URLSearchParams({hello: 'world'})}).text().then(console.log)
```
This commit adds alternate dispatch method to core JS API.
"Deno.core.dispatchByName()" works like "Deno.core.dispatch()",
but takes op name instead of op id as a first argument.
* refactor "compile" and "runtimeCompile" in "compiler.ts" and factor out
separate methods for "compile" and "bundle" operations
* remove noisy debug output from "compiler.ts"
* provide "Serialize" implementations for enums in "msg.rs"
* rename "analyze_dependencies_and_references" to "pre_process_file" and
move it to "tsc.rs"
* refactor ModuleGraph to use more concrete types and properly annotate
locations where errors occur
* remove dead code from "file_fetcher.rs" - "SourceFile.types_url" is no
longer needed, as type reference parsing is done in "ModuleGraph"
* remove unneeded field "source_path" from ".meta" files stored for
compiled source file (towards #6080)
This commit provides a "system_loader_es5.js" bundle loader which will be added
to the bundle when the target is < ES2017, which is the minimum target syntax
required for "system_loader.js".
Supports #5913 (via Deno.bundle()) with a couple caveats:
* Allowing "deno bundle" to take a different target is not supported, as we
specifically ignore "target" when passed in a TypeScript config file. This is
because deno bundle is really intended to generate bundles that work in Deno.
It is an unintentional side effect that some bundles are loadable in browsers.
* While a target of "es3" will be accepted, the module loader will still only be
compatible with ES5 or later. Realistically no one should be expecting bundles
generated by Deno to be used on IE8 and prior, and there is just too much
"baggage" to support that at this point.
This is a minor variation of 75bb9d, which exposed some sort of internal V8 bug.
Ref #6358
This is 100% authored by Kitson Kelly. Github might change the author when landing
so I'm leaving this in:
Co-authored-by: Kitson Kelly <me@kitsonkelly.com>
This commit provides a "system_loader_es5.js" bundle loader which will be added
to the bundle when the target is < ES2017, which is the minimum target syntax
required for "system_loader.js".
Supports #5913 (via Deno.bundle()) with a couple caveats:
* Allowing "deno bundle" to take a different target is not supported, as we
specifically ignore "target" when passed in a TypeScript config file. This is
because deno bundle is really intended to generate bundles that work in Deno.
It is an unintentional side effect that some bundles are loadable in browsers.
* While a target of "es3" will be accepted, the module loader will still only be
compatible with ES5 or later. Realistically no one should be expecting bundles
generated by Deno to be used on IE8 and prior, and there is just too much
"baggage" to support that at this point.
The tests for testing that `Deno.truncateSync` and `Deno.truncate`
require write permissions seem to not call the functions they are
testing *at all* and are calling `Deno.mkdir` and `Deno.mkdirSync`
instead.
This commit replaces those calls with calls to `Deno.truncateSync`
and `Deno.truncate` respectively.
Currently WebAssembly runtime errors don't propagate up to the user as
they use urls to denote where the error occurred which get caught by the source-map
pipeline which doesn't support the wasm scheme.
This commit:
* added default file globs so "deno lint" can be run
without arguments (just like "deno fmt")
* added test for globs in "deno lint"
* upgrade "deno_lint" crate to v0.1.9
This commit fixes several regressions in TS compiler:
* double compilation of same module during same process run
* compilation of JavaScript entry point with non-JS imports
* unexpected skip of emit during compilation
Additional checks were added to ensure "allowJs" setting is
used in TS compiler if JavaScript has non-JS dependencies.
Currently sync operations on stdin are failing because tokio::Stdin
cannot be converted to a std::File.
This commit replaces tokio::stdin with a raw file descriptor
wrapped in a std::fs::File which can be converted to a
tokio::File and back again making the synchronous version
of op_read actually work.
$HOME is meaningless on Windows. It may be set by users or by third
party software, but it is non-standard and should not be relied upon.
Likewise, $USERPROFILE is meaningless on other platforms.
This reverts commit c4c6a8dae4
There is some controversy about this change because vscode doesn't interpret the fragments correctly. Needs more discussion before landing.
This commit fixes regression that caused TS dependencies
not being compiled.
Check was added that ensures TS compiler is run if
any of dependencies in module graph is TS/TSX/JSX.
This PR addresses many problems with module graph loading
introduced in #5029, as well as many long standing issues.
"ModuleGraphLoader" has been wired to "ModuleLoader" implemented
on "State" - that means that dependency analysis and fetching is done
before spinning up TS compiler worker.
Basic dependency tracking for TS compilation has been implemented.
Errors caused by import statements are now annotated with import
location.
Co-authored-by: Ryan Dahl <ry@tinyclouds.org>
Since everything that Deno loads is treated as an ES Module,
it means that all code is treated as "use strict" except for
when using the REPL. This PR changes that so code in the
REPL is also always evaluated with "use strict". There are
also a couple other places where we load code as scripts
which should also use "use strict" just in case.
This commit changes how error occurring in SWC are handled.
Changed lexer settings to properly handle TS decorators.
Changed output of SWC error to annotate with position in file.
This commit fixes a bug introduced in #5029 that caused bad
handling of redirects during module analysis.
Also ensured that duplicate modules are not downloaded.
This commit disabled flaky test in cli/module_graph.rs.
The test uses 019_media_types.ts which was known to be flaky
for some time. Test should be reenabled once test HTTP server
is rewritten to Rust.
This commit completely overhauls how module analysis is
performed in TS compiler by moving the logic to Rust.
In the current setup module analysis is performed using
"ts.preProcessFile" API in a special TS compiler worker
running on a separate thread.
"ts.preProcessFile" allowed us to build a lot of functionality
in CLI including X-TypeScript-Types header support
and @deno-types directive support. Unfortunately at the
same time complexity of the ops required to perform
supporting tasks exploded and caused some hidden
permission escapes.
This PR introduces "ModuleGraphLoader" which can parse
source and load recursively all dependent source files; as
well as declaration files. All dependencies used in TS
compiler and now fetched and collected upfront in Rust
before spinning up TS compiler.
To achieve feature parity with existing APIs this commit
includes a lot of changes:
* add "ModuleGraphLoader"
- can fetch local and remote sources
- parses source code using SWC and extracts imports, exports, file references, special
headers
- this struct inherited all of the hidden complexity and cruft from TS version and requires
several follow up PRs
* rewrite cli/tsc.rs to perform module analysis upfront and send all required source code to
TS worker in one message
* remove op_resolve_modules and op_fetch_source_files from cli/ops/compiler.rs
* run TS worker on the same thread
This commit removes "check_stderr" setting from itest! macro used
to generate integration tests. Without this setting on tests discarded
output of stderr making it very hard to debug the problem in test.
Numerous tests were changed by adding "--quiet" flag to not display
"Compile"/"Download" prompts.
This commit fixes panic occurring if $DENO_DIR is set to a relative
path, eg. "DENO_DIR=denodir deno run main.ts".
Before creating DenoDir instance given path is checked and if necessary
resolved against current working directory.
Additional sanity checks were put in place to ensure all caches
receive absolute path for the location.
This commit fixes problems with source maps in Chrome Devtools
by substituting source map URL generated by TS compiler with
actual file URL pointing to DENO_DIR.
Dummy value of "source_map_url" has been removed from
"ScriptOrigin".
Also fixes lock file which used compiled source code to generate
lock hash; it now uses source code of the file that is
being compiled.
This PR hot-fixes permission escapes in dynamic imports, workers
and runtime compiler APIs.
"permissions" parameter was added to public APIs of SourceFileFetcher
and appropriate permission checks are performed during loading of
local and remote files.
* Prepend underscores to private modules
* Remove collectUint8Arrays() It would be a misuse of Deno.iter()'s result.
* Move std/_util/async.ts to std/async
* Move std/util/sha*.ts to std/hash
Importing .wasm files is non-standardized therefore deciding to
support current functionality past 1.0 release is risky.
Besides that .wasm import posed many challenges in our codebase
due to complex interactions with TS compiler which spawned
thread for each encountered .wasm import.
This commit removes:
- cli/compilers/wasm.rs
- cli/compilers/wasm_wrap.js
- two integration tests related to .wasm imports
This commit updates "deno_typescript" crate to properly map
bundle entrypoint file to internal specifier.
All import specifiers were remapped from "file:///a/b/c.ts" to
"$deno$/a/b/c.ts", but that was not the case for entrypoint file
"main.ts" and "compiler.ts".
Because of that internal stack traces were inconsistent; showing
"file:///some/random/path/on/ci/machine.ts" URL in frames that
originate from "main.ts" or "compiler.ts" and "$deno$/file.ts"
for all other imports.
This PR removes op_cache and refactors how Deno interacts with TS compiler.
Ultimate goal is to completely sandbox TS compiler worker; it should operate on
simple request -> response basis. With this commit TS compiler no longer
caches compiled sources as they are generated but rather collects all sources
and sends them back to Rust when compilation is done.
Additionally "Diagnostic" and its children got refactored to use "Deserialize" trait
instead of manually implementing JSON deserialization.
The issue is solved by proxying websocket messages over a pair of
`futures::mpsc::unbounded` channels. As these are are implemented in
the 'futures' crate, they can't participate in Tokio's cooperative
task yielding.
refactor: Parse URLs more sequentially. This makes it easier to change matching behaviour depending on the protocol.
fix: Fail when a host isn't given for certain protocols.
fix: Convert back-slashes info forward-slashes.
This PR removes the hack in CLI that allows to run scripts with shorthand: deno script.ts.
Removing this functionality because it hacks around short-comings of clap our CLI parser. We agree that this shorthand syntax is desirable, but it needs to be rethinked and reimplemented. For 1.0 we should go with conservative approach that is correct.
This PR adds prepare_load hook method to ModuleLoader trait. It allows implementors to perform preparation work before starting actual module loading into isolate. It's meant to be used in CLI; where "transpilation" step will be explicitly performed during prepare_load instead of doing it adhoc for each module if needed.
Keep in mind Buffer.toString() still exists, but returns [object Object].
Reason for removal of Buffer.toString() was that it implicitly used
TextDecoder with fixed "utf-8" encoding and no way to customize
the encoding.
This change is to prevent needed a separate stat syscall for each file
when using readdir.
For consistency, this PR also modifies std's `WalkEntry` interface to
extend `DirEntry` with an additional `path` field.
This commit removes "combined" interfaces from cli/js/io.ts; in the
like of "ReadCloser", "WriteCloser" in favor of using intersections
of concrete interfaces.
Changed `URL.port` implementation to match [WHATWG
specifications](https://url.spec.whatwg.org/#port-state).
This PR matches the behaviour of other browsers:
1. a `TypeError` must be thrown when passing an URL with an invalid
port to the constructor.
2. When setting an invalid port, using property setter, I haven't found
what should happen in this case, so I mimic **Firefox** & **Node**
behaviour. If an invalid port is set, it will use the previous value.
**Chrome** sets the value to `'0'` if an invalid port is set. I prefer
to keep the previous valid value. (I can use Chrome's behaviour if you
think it's better, it's a simple value change)
```
url.port = '3000'; // valid
url.port = 'deno'; // invalid
assertEquals(url.port, '3000');
```
3. If the port value equals the current protocol default port value,
`port` will be an empty string.
When creating a console instance, one must pass "printFunc" arg
which is used internally by Console to output messages.
Due to numerous refactors there was a single method ("console.clear()")
that used "Deno.stdout" instead of "printFunc".
This commit unifies how "Console" outpus message, by using
"printFunc" in all methods; consequently "Deno.stdout" is no longer
imported in "cli/js/console.ts" making it a standalone module that doesn't
depend on any CLI-specific APIs.
Deno.runTests() interface is not yet good enough to be exposed
publicly with stability guarantees.
This commit removes public API related to testing: Deno.runTests()
and Deno.TestMessage, but keeps them exposed on Deno.internal object
so they can be used with "deno test" subcommand.
* Allows union and intersection types to have same multi-line behaviour as arguments (more like prettier).
* Stops line breaks before `extends` in conditional type, which would have been a parser error.
* Check for `// dprint-ignore-file` comment before parsing. This allow files that panic in swc to be ignored.
* Fixes a bug in "multi-line" detection.
Refactors Event and EventTarget so that they better encapsulate their
non-public data as well as are more forward compatible with things like
DOM Nodes.
Moves `dom_types.ts` -> `dom_types.d.ts` which was always the intention,
it was a legacy of when we used to build the types from the code and the
limitations of the compiler. There was a lot of cruft in `dom_types`
which shouldn't have been there, and mis-alignment to the DOM standards.
This generally has been eliminated, though we still have some minor
differences from the DOM (like the removal of some deprecated
methods/properties).
Adds `DOMException`. Strictly it shouldn't inherit from `Error`, but
most browsers provide a stack trace when one is thrown, so the behaviour
in Deno actually better matches the browser.
`Event` still doesn't log to console like it does in the browser. I
wanted to get this raised and that could be an enhancement later on (it
currently doesn't either).
- Removes the __fetch namespace from `deno types`
- Response.redirect should be a static.
- Response.body should not be AsyncIterable.
- Disables the deno_proxy benchmark
- Makes std/examples/curl.ts buffer the body before printing to stdout
Fixes #4000 and fixes #4476. Now always tries to fetch reference types
for JS files. Does not throw if it fails, since Typescript compiler will
complain if the file is not there(it will try to fetch it again first)
and people who just use JS should not be bothered by this error.
Not sure about my test, it passes and catches the bug but maybe there is
a better way to express it.
* Reduce "testing" interfaces
* Use a callback instead of a generator for Deno.runTests()
* Default RunTestsOptions::reportToConsole to true
* Compose TestMessage into a single interface
* Properly track isFile, isSymlink, isDirectory
These don't exhaust all the possibilities, so none of them should be
defined as "neither of the others".
* empty
1. Array elements are now grouped the same as in Node.js
2. Limit to 100 (Node.js default) elements to display in iterable
3. Print each element in new line if excessing max line length (same as in Node.js)
4. Print length of the TypedArray
5. Print information about empty items in Array
This is a first pass implementation which is still missing several important
features:
- support for --inspect-brk (#4503)
- support for source maps (#4501)
- support for piping console.log to devtools console (#4502)
Co-authored-by: Bert Belder <bertbelder@gmail.com>
Co-authored-by: Matt Harrison <mt.harrison86@gmail.com>
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
* Remove DENO_BUILD_MODE and DENO_BUILD_PATH
Also remove outdated docs related to ninja/gn.
* fix
* remove parameter to build_mode()
* remove arg parsing from benchmark.py
This a complex boring PR that shifts around code (primarily) in cli/fs.rs and
cli/ops/fs.rs. The gain of this refactoring is to ease the way for #4188 and
#4017, and also to avoid some future development pain.
Mostly there is no change in functionality. Except:
* squashed bugs where op_utime and op_chown weren't using `resolve_from_cwd`
* eliminated the use of the external `remove_dir_all` crate.
* op_chmod now only queries metadata to verify file/dir exists on Windows (it
will already fail on Unix if it doesn't)
* op_chown now verifies the file/dir's existence on Windows like chmod does.
* add tests for "Deno.core.encode" and "Deno.core.decode" for empty inputs
* use "Deno.core.encode" in "TextEncoder"
* use "Deno.core.decode" in "TextDecoder"
* remove "core_decode" and "core_encode" benchmarks
After splitting "failFast" and "exitOnFail" arguments, there was a situation where failing tests did not exit with code 1.
* fixed argument value passed to Deno.runTests() in deno test
* fixed argument value passed to Deno.runTests() in std/testing/runner.ts
* added integration tests for deno test to ensure failFast and exitOnFail work as expected
* don't write test file to file system, but keep it in memory
This PR brings assertOps and assertResources sanitizers to Deno.test() API.
assertOps checks that test doesn't leak async ops, ie. there are no unresolved
promises originating from Deno APIs. Enabled by default, can be disabled using
Deno.TestDefinition.disableOpSanitizer.
assertResources checks that test doesn't leak resources, ie. all resources used
in test are closed. For example; if a file is opened during a test case it must be
explicitly closed before test case finishes. It's most useful for asynchronous
generators. Enabled by default, can be disabled using
Deno.TestDefinition.disableResourceSanitizer.
We've used those sanitizers in internal runtime tests and it proved very useful in
surfacing incorrect tests which resulted in interference between the tests.
All tests have been sanitized.
Closes #4208
Fixes #4101
Previously, we would just provide the raw JSON to the TypeScript
compiler worker, but TypeScript does not transform JSON. This caused
a problem when emitting a bundle, that the JSON would just be "inlined"
into the output, instead of being transformed into a module.
This fixes this problem by providing the compiled JSON to the TypeScript
compiler, so TypeScript just sees JSON as a "normal" TypeScript module.
This PR attempts to fix intermittent errors occurring on Windows for "cli/tests/unit_test_runner.ts." Runner has been reworked to create only single TCP listener instead of one listener per worker.
Additionally worker doesn't close TCP socket - it waits for parent process to close the socket and only then exits.
This commit changes output of default test reporter to resemble output from Rust test runner;
first the name of running test is printed with "...", then after test has run result is printed on the same line.
* Split "Deno.TestEvent.Result" into "TestStart" and "TestEnd";
* changes TestReporter interface to support both events;
Co-authored-by: Ryan Dahl <ry@tinyclouds.org>
This commits add two new methods to "Deno.core" namespace: "encode" and "decode".
Those methods are bound in Rust to provide a) fast b) generally available of encoding and decoding UTF-8 strings.
Both methods are now used in "cli/js/dispatch_json.ts".
* drop server guard before unit test result check
To prevent cascading test failures when js_unit_test http server
guard is dropped before asserting that tests were successful.
This is really a band-aid and doesn't solve underlying issue with
http server.
* Update CLI for unit_test_runner.ts
* Change cli/js/tests/unit_test_runner.ts command line interface to work in 3
modes:
- "one-off" - run tests that match permissions of currently running
process
- "master" - run tests for all possible permission combinations, by
spawning subprocesses running in "worker" mode and communicating via
TCP socket; requires elevated permissions
- "worker" - run tests for set of permissions provided by CLI arg;
requires elevated permissions to setup TCP connection to "master";
after initial setup process drops permissions to given set
* Support filtering of tests by string passed after "--" CLI arg
* Update cli/js/tests/README.md
Rewrites "cli/js/unit_test_runner.ts" to communicate with spawned subprocesses
using TCP socket.
* Rewrite "Deno.runTests()" by factoring out testing logic to private "TestApi"
class. "TestApi" implements "AsyncIterator" that yields "TestEvent"s,
which is an interface for different types of event occuring during running
tests.
* Add "reporter" argument to "Deno.runTests()" to allow users to provide custom
reporting mechanism for tests. It's represented by "TestReporter" interface,
that implements hook functions for each type of "TestEvent". If "reporter"
is not provided then default console reporting is used (via
"ConsoleReporter").
* Change how "unit_test_runner" communicates with spawned suprocesses. Instead
of parsing text data from child's stdout, a TCP socket is created and used
for communication. "unit_test_runner" can run in either "master" or "worker"
mode. Former is responsible for test discovery and establishing needed
permission combinations; while latter (that is spawned by "master") executes
tests that match given permission set.
* Use "SocketReporter" that implements "TestReporter" interface to send output
of tests to "master" process. Data is sent as stringified JSON and then
parsed by "master" as structured data. "master" applies it's own reporting
logic to output tests to console (by reusing default "ConsoleReporter").
- moves signal definition from "cli/js/process.ts" to "cli/js/signals.ts"
- removes "cli/js/dispatch.ts"
- removes "cli/js/types.ts"
- moves web specific utilities to "cli/js/web/util.ts"
- moves compiler implementation to "cli/js/compiler/" directory
- moves more APIs to "cli/js/web":
* "console.ts"
* "console_table.ts"
* "performance.ts"
* "timers.ts"
* "workers.ts"
- removes some dead code from "cli/js/"
Listener and UDPConn are AsyncIterables instead of AsyncIterators.
The [Symbol.asyncIterator]()s are defined as generators and the
next() methods are gone.
"Listener/Socket has been closed" errors are now BadResource.
All Deno runtime test files were moved to cli/js/tests/ directory.
It makes a clear distinction that cli/js/tests/ contains code
that is run under Deno runtime as opposed to code in cli/js/ which
is used to create bundle and snapshot with "deno_typescript".
Following JS ops were moved to separate files in cli/js/ops directory:
- io
- process
- worker_host
- web_worker
- plugins
- timers
- signal
- permissions
Following JS ops were moved to separate files in cli/js/ops directory:
- compiler
- dispatch_json
- dispatch_minimal
- errors
- fetch
- fs_events
- os
- random
- repl
- resources
- runtime_compiler
- runtime
- tty
There's a lot of variation in doc comments and internal code about
whether chmod/0o777-style permissions are called `mode` or `perm`. (For
example, mkdir and writeFile choose differently.)
Had proposed earlier to go consistently with `perm`, but on balance devs
prefer to go with `mode`.
There's a lot of variation in doc comments and internal code about
whether the first parameter to file system calls is `path` or `name` or
`filename`. For consistency, have made it always be `path`.
* refactor: preliminary cleanup of Deno.runTests()
* Change time measurement to use new Date() instead of
performance.now(). Because there is no guarantee that tests are
run with "--allow-hr" using new Date() guarantees higher
precision of 1ms instead of 2ms.
* Support String type filter in "skip" and "only".
* Split "exitOnFail" into "exitOnFail" and "failFast".
Former tells if "runTests()" should exit with code 1 on test
failure, while latter tells if "runTests()" should stop
running tests on first failure.
* Use "defer" to wait for unhandled promise rejection - this bit
is funky and doesn't seem right, but for now it's just a rewrite
from using "setTimeout". Intended to be fixed in later commits.
* Remove global "__DENO_TEST_REGISTRY", don't expose list of
registered tests (to be addressed in follow up commits)
* Remove arbitrary slow test threshold; use uniform coloring
instead
- Cleanup "tsCompilerOnMessage" by factoring out separate methods for each
request type:
* "compile"
* "runtimeCompile"
* "runtimeTranspile"
- Simplify control flow of compiler workers by a) no longer calling "close()" in worker runtime after a
single message; b) explicitly shutting down worker from host after a single message
Co-authored-by: Ryan Dahl <ry@tinyclouds.org>
Rewrite "testPerm" helper function used for testing of internal
runtime code. It's been renamed to "unitTest" and provides API that
is extensible in the future by accepting optional "UnitTestOptions"
argument. "test" helper was also removed and replaced by
overloaded version of "unitTest" that takes only function argument.
"UnitTestOptions" currently supports "perms" and "skip"
options, where former works exactly as first argument to "testPerm"
did, while the latter allows to conditionally skip tests.
* add "assertOps" test assertion which makes sure test case
is not "leaking" ops - ie. after test finishes there are no
pending async ops
* apply "assertOps" to all tests in "cli/js/"
* fix numerous tests leaking ops
* document problem with edge case in "clearInterval"
and "clearTimeout" implementation where they
may leak async ops
* move "cli/js/worker_test.ts" to "cli/tests/worker_test.ts" and
run as integration test; workers leak ops because of missing
"terminate" implementation
Fixes #4602
We turned off `allowJs` by default, to keep the compiler from grabbing
a bunch of files that it wouldn't actually do anything useful with. On
the other hand, this caused problems with bundles, where the compiler
needs to gather all the dependencies, including JavaScript ones. This
fixes this so that when we are bundling, we analyse JavaScript imports
in the compiler.
* Rename MkdirOption interface to MkdirOptions
It was hard to remember which Options interfaces were spelled in the
singular and which in the plural (and anyway this one contained two
options).
Also added MkdirOptions to cli/js/deno.ts. All the analogous interfaces
were exported there.
* Rename RemoveOption interface to RemoveOptions
This was the last remaining Option interface spelled in the singular.
Easier to remember if they're all Option**s**; plus we may want to add extra
options here later.
Rewrite "normalize_path()" to remove all intermediate components from the path, ie. "./" and "../". It's very similar in functionality to fs::canonicalize(), however "normalize_path() doesn't resolve symlinks.
Previously, bundles always utilised top level await, even if the bundled
modules didn't require top level await. Now, analysis of the bundle is
done and if none of the bundled modules are asynchronously executed,
then the bundle as a whole will be synchronously executed.
Fixes #4055
Fixes #4123
- Added `ServerRequest.finalize()`: consuming all unread body stream and trailers.
- This is cleanup method for reading next request from same keep-alive connection.
- Needed when handler didn't consume all body and trailers even after responding.
- refactor: `ServerRequest._bodyStream()`, `ServerRequestBody` are removed.
- Now using `bodyReader()` and `chunkedBodyReader()` instead.
- fix: Trailers should only be read `transfer-encoding` is `chunked` and `trailer` header is set and its value is valid.
- fix: use `Headers.append()` on reading trailers.
- fix: delete `trailer` field from headers after reading trailers.
- reorg: Several functions related to IO are moved into `http/io.ts`
This PR introduces the mutex guard for the test cases which depends on the
permission prompt mocking utility. permission_request test cases depend on
the mocked (dummy) value of the permission_prompt result. The value is stored
at static STUB_PROMPT_VALUE: AtomicBool and these test cases share this
value. So we should lock at the start of these test cases.
Before this change cargo test permission failed 6 times out of 20. After this
change, 0 times out of 20 (on my mac).
To better reflect changes in error types in JS from #3662 this PR changes
default error type used in ops from "ErrBox" to "OpError".
"OpError" is a type that can be sent over to JSON; it has all
information needed to construct error in JavaScript. That
made "GetErrorKind" trait useless and so it was removed altogether.
To provide compatibility with previous use of "ErrBox" an implementation of
"From<ErrBox> for OpError" was added, however, it is an escape hatch and
ops implementors should strive to use "OpError" directly.
This PR fixes an issue where we recursively analysed imports on plain JS files
in the compiler irrespective of "checkJs" being true. This caused problems
where when analysing the imports of those files, we would mistake some
import like structures (AMD/CommonJS) as dependencies and try to resolve
the "modules" even though the compiler would not actually look at those files.
Fixes #4031
When a bundle contains a single module, we were incorrectly determining
the module name, resulting in a non-functional bundle. This PR corrects
that determination.
Example:
$ python2 -c 'open("\x80\x7F", "w")'
$ deno eval 'Deno.readDirSync(".")'
thread 'main' panicked at 'called `Option::unwrap()` on a `None` value', cli/ops/fs.rs:373:16
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
fatal runtime error: failed to initiate panic, error 5
Aborted (core dumped)
Before this commit they made deno panic, now they are silently skipped.
Not ideal but arguably better than panicking.
No test because what characters are and aren't allowed in filenames is
highly file system-dependent.
Closes #3950
Moves to using a minimal System loader for bundles generated by Deno.
TypeScript in 3.8 will be able to output TLA for modules, and the loader
is written to take advantage of that as soon as we update Deno to TS
3.8.
System also allows us to support `import.meta` and provide more ESM
aligned assignment of exports, as well as there is better handling of
circular imports.
The loader is also very terse versus to try to save overhead.
Also, fixed an issue where abstract classes were not being re-exported.
Fixes #2553
Fixes #3559
Fixes #3751
Fixes #3825
Refs #3301
* rewrite test runner in Rust
* migrate "test" and "runTests" functions from std to "Deno" namespace
* use "Deno.test()" to run internal JS unit tests
* remove std downloads for Deno subcommands
* Use PathBuf for DenoSubcommand::Bundle's out_file
* Use PathBuf for DenoSubcommand::Format's files
* Use PathBuf for DenoSubcommand::Install's dir
* Use PathBuf for read/write whitelists
* establish basic event loop for workers
* make "self.close()" inside worker
* remove "runWorkerMessageLoop() - instead manually call global function
in Rust when message arrives. This is done in preparation for structured clone
* refactor "WorkerChannel" and use distinct structs for internal
and external channels; "WorkerChannelsInternal" and "WorkerHandle"
* move "State.worker_channels_internal" to "Worker.internal_channels"
* add "WorkerEvent" enum for child->host communication;
currently "Message(Buf)" and "Error(ErrBox)" variants are supported
* add tests for nested workers
* add tests for worker throwing error on startup
In denoland/deno#2335 a conditional was added to make sure
toAsyncIterator didn't skip chunks because the reader returned data and
EOF in a single call, fixing #2330.
Later, in denoland/deno#2591, the `Reader` interface changed to
`Promise<number | EOF>`. Since the reader no longer returns data and EOF
in a single call, this conditional is not necessary. We can just return
`{ done: true }` when we get `EOF`.
Co-authored-by: Arun Srinivasan <rulfzid@gmail.com>
Co-authored-by: Arun Srinivasan <rulfzid@gmail.com>
- Exports diagnostic items from `diagnostics.ts` which are missing at
runtime.
- Returns an array of diagnostics, instead of an object with a property
of `items`. This is because of the way Rust deals with certain
structures, and shouldn't be exposed in the APIs.
This change simplifies how we execute V8. Previously V8 Isolates jumped
around threads every time they were woken up. This was overly complex and
potentially hurting performance in a myriad ways. Now isolates run on
their own dedicated thread and never move.
- blocking_json spawns a thread and does not use a thread pool
- op_host_poll_worker and op_host_resume_worker are non-operational
- removes Worker::get_message and Worker::post_message
- ThreadSafeState::workers table contains WorkerChannel entries instead
of actual Worker instances.
- MainWorker and CompilerWorker are no longer Futures.
- The multi-threaded version of deno_core_http_bench was removed.
- AyncOps no longer need to be Send + Sync
This PR is very large and several tests were disabled to speed
integration:
- installer_test_local_module_run
- installer_test_remote_module_run
- _015_duplicate_parallel_import
- _026_workers
For some reason, the unit tests for Deno.remove() were not being imported to
unit_tests.ts and, consequently, not being executed. Thus, I imported them,
refactored some existent ones and wrote new ones for the symlink removal case.
Since the creation of a symlink is not implemented for Windows yet, assertions
that consider this state were added when the tests are executed in this OS.
This flag was added to evaluate performance relative to tokio's threaded
runtime. Although it's faster in the HTTP benchmark, it's clear the runtime
is not the only perf problem.
Removing this flag will simplify further refactors, in particular
adopting the #[tokio::main] macro. This will be done in a follow up.
Ultimately we expect to move to the current thread runtime with Isolates
pinned to specific threads, but that will be a much larger refactor. The
--current-thread just complicates that effort.
Before:
```
▶ target/debug/deno https://deno.land/std/examples/echo_server.ts
error: Uncaught PermissionDenied: run again with the --allow-net flag
► $deno$/dispatch_json.ts:40:11
at DenoError ($deno$/errors.ts:20:5)
at unwrapResponse ($deno$/dispatch_json.ts:40:11)
at sendSync ($deno$/dispatch_json.ts:67:10)
at listen ($deno$/net.ts:170:15)
at https://deno.land/std/examples/echo_server.ts:4:23
```
```
▶ target/debug/deno --allow-read=/usr https://deno.land/std/examples/cat.ts /etc/passwd
error: Uncaught PermissionDenied: run again with the --allow-read flag
► $deno$/dispatch_json.ts:40:11
at DenoError ($deno$/errors.ts:20:5)
at unwrapResponse ($deno$/dispatch_json.ts:40:11)
at sendAsync ($deno$/dispatch_json.ts:91:10)
```
After:
```
▶ target/debug/deno https://deno.land/std/examples/echo_server.ts
error: Uncaught PermissionDenied: network access to "0.0.0.0:8080", run again with the --allow-net flag
► $deno$/dispatch_json.ts:40:11
at DenoError ($deno$/errors.ts:20:5)
at unwrapResponse ($deno$/dispatch_json.ts:40:11)
at sendSync ($deno$/dispatch_json.ts:67:10)
at listen ($deno$/net.ts:170:15)
at https://deno.land/std/examples/echo_server.ts:4:23
```
```
▶ target/debug/deno --allow-read=/usr https://deno.land/std/examples/cat.ts /etc/passwd
error: Uncaught PermissionDenied: read access to "/etc/passwd", run again with the --allow-read flag
► $deno$/dispatch_json.ts:40:11
at DenoError ($deno$/errors.ts:20:5)
at unwrapResponse ($deno$/dispatch_json.ts:40:11)
at sendAsync ($deno$/dispatch_json.ts:91:10)
```
* move is_dyn_import argument from Loader::resolve to Loader::load - it was always kind of strange that resolve() checks permissions.
* change argument type from &str to &ModuleSpecifier where applicable
Ref #3712. This change allowed the deno_typescript crate to reference
cli/js/lib.deno_runtime.d.ts which breaks "cargo package". We intend to
reintroduce a revised version of this patch later once "cargo
package" is working and tested.
This reverts commit 737ab94ea1.
* split ops/worker.rs into ops/worker_host.rs and ops/web_worker.rs
* refactor js/workers.ts and factor out js/worker_main.ts - entry point for WebWorker runtime
* BREAKING CHANGE: remove support for blob: URL in Worker
* BREAKING CHANGE: remove Deno namespace support and noDenoNamespace option in Worker constructor
* introduce WebWorker struct which is a stripped down version of cli::Worker
* refactored RecursiveLoad - it was renamed to RecursiveModuleLoad, it does not take ownership of isolate anymore - a struct implementing Stream that yields SourceCodeInfo
* untangled module loading logic between RecursiveLoad and isolate - that logic is encapsulated in EsIsolate and RecursiveModuleLoad, where isolate just consumes modules as they become available - does not require to pass Arc<Mutex<Isolate>> around anymore
* removed EsIsolate.mods_ in favor of Modules and moved them inside EsIsolate
* EsIsolate now requires "loader" argument during construction - struct that implements Loader trait
* rewrite first methods on isolate as async
