Implementation of generics for `#[op2]`, along with some refactoring to
improve the ergonomics of ops with generics parameters:
- The ops have generics on the struct rather than the associated
methods, which allows us to trait-ify ops (impossible when they are on
the methods)
- The decl() method can become a trait-associated const field which
unlocks future optimizations
Callers of ops need to switch from:
`op_net_connect_tcp::call::<TestPermission>(conn_state, ip_addr)` to
`op_net_connect_tcp::<TestPermission>::call(conn_state, ip_addr)`.
This is a reproduction and fix for a very obscure bug where the Deno
runtime locks up we end up polling an empty JoinSet and attempt to
resolve ops after-the-fact. There's a small footgun in the JoinSet API
where polling it while empty returns Ready(None), which means that it
never holds on to the waker. This means that if we aren't testing for
this particular return value and don't stash the waker ourselves for a
future async op to eventually queue, we can end up losing the waker
entirely and the op wakes up, notifies tokio, which notifies the
JoinSet, which then has nobody to notify 😢.
Co-authored-by: Luca Casonato <hello@lcas.dev>
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
…nclusion" (#19519)"
This reverts commit 28a4f3d0f5.
This change causes failures when used outside Deno repo:
```
============================================================
Deno has panicked. This is a bug in Deno. Please report this
at https://github.com/denoland/deno/issues/new.
If you can reliably reproduce this panic, include the
reproduction steps and re-run with the RUST_BACKTRACE=1 env
var set and include the backtrace in your report.
Platform: linux x86_64
Version: 1.34.3+b37b286
Args: ["/opt/hostedtoolcache/deno/0.0.0-b37b286f7fa68d5656f7c180f6127bdc38cf2cf5/x64/deno", "test", "--doc", "--unstable", "--allow-all", "--coverage=./cov"]
thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: Failed to read "/home/runner/work/deno/deno/core/00_primordials.js"
Caused by:
No such file or directory (os error 2)', core/runtime/jsruntime.rs:699:8
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
```
Implements `Result` in fast-calls. Note that the approach here is
slightly different. Rather than store the last result in the `OpState`,
we put it into the `OpCtx` which saves us a lookup and lock in the error
case. We do not have to lock this field as it's guaranteed only one
runtime and thread can ever access it.
The fastcall path for many ops can avoid doing a great deal of work,
even for `Result` return values. In the previous iteration of `ops`, all
`Result`-returning functions would fetch and lock the `OpState`,
regardless of whether it was used or not.
Relands #19463. This time the `ExtensionFileSourceCode` enum is
preserved, so this effectively just splits feature
`include_js_for_snapshotting` into `exclude_js_sources` and
`runtime_js_sources`, adds a `force_include_js_sources` option on
`extension!()`, and unifies `ext::Init_ops_and_esm()` and
`ext::init_ops()` into `ext::init()`.
This is a new op system that will eventually replace `#[op]`.
Features
- More maintainable, generally less-coupled code
- More modern Rust proc-macro libraries
- Enforces correct `fast` labelling for fast ops, allowing for visual
scanning of fast ops
- Explicit marking of `#[string]`, `#[serde]` and `#[smi]` parameters.
This first version of op2 supports integer and Option<integer>
parameters only, and allows us to start working on converting ops and
adding features.
Fixes #19568
Values are not coerced to the desired type during deserialisation. This
makes serde_v8 stricter.
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
`ZeroCopyBuf` was convenient to use, but sometimes it did hide details
that some copies were necessary in certain cases. Also it made it way to easy
for the caller to pass around and convert into different values. This commit
splits `ZeroCopyBuf` into `JsBuffer` (an array buffer coming from V8) and
`ToJsBuffer` (a Rust buffer that will be converted into a V8 array buffer).
As a result some magical conversions were removed (they were never used)
limiting the API surface and preparing for changes in #19534.
A few easy migrations of module code from the runtime to the module map.
The module map already has a few places where it needs a handle scope,
so we're not coupling it any further with the v8 runtime.
`init_runtime_module_map` is replaced with an option to reduce API
surface of JsRuntime.
`module_resolve_callback` now lives in the `ModuleMap` and we use a
annex data to avoid having to go through the `Rc<RefCell<...>>` stored
in the `JsRuntime`'s isolate.
This is a quick first refactoring to split the tests out of runtime and
move runtime-related code to a top-level runtime module.
There will be a followup to refactor imports a bit, but this is the
major change that will most likely conflict with other work and I want
to merge it early.
… (#19463)"
This reverts commit ceb03cfb03.
This is being reverted because it causes 3.5Mb increase in the binary
size,
due to runtime JS code being included in the binary, even though it's
already snapshotted.
CC @nayeemrmn
Remove `ExtensionFileSourceCode::LoadedFromFsDuringSnapshot` and feature
`include_js_for_snapshotting` since they leak paths that are only
applicable in this repo to embedders. Replace with feature
`exclude_js_sources`. Additionally the feature
`force_include_js_sources` allows negating it, if both features are set.
We need both of these because features are additive and there must be a
way of force including sources for snapshot creation while still having
the `exclude_js_sources` feature. `force_include_js_sources` is only set
for build deps, so sources are still excluded from the final binary.
You can also specify `force_include_js_sources` on any extension to
override the above features for that extension. Towards #19398.
But there was still the snapshot-from-snapshot situation where code
could be executed twice, I addressed that by making `mod_evaluate()` and
scripts like `core/01_core.js` behave idempotently. This allowed
unifying `ext::init_ops()` and `ext::init_ops_and_esm()` into
`ext::init()`.
This commit migrates "deno_core" from using "FuturesUnordered" to
"tokio::task::JoinSet". This makes every op to be a separate Tokio task
and should unlock better utilization of kqueue/epoll.
There were two quirks added to this PR:
- because of the fact that "JoinSet" immediately polls spawn tasks,
op sanitizers can give false positives in some cases, this was
alleviated by polling event loop once before running a test with
"deno test", which gives canceled ops an opportunity to settle
- "JsRuntimeState::waker" was moved to "OpState::waker" so that FFI
API can still use threadsafe functions - without this change the
registered wakers were wrong as they would not wake up the
whole "JsRuntime" but the task associated with an op
---------
Co-authored-by: Matt Mastracci <matthew@mastracci.com>
It's not used anymore. Subsequently allows removing
`ModuleMap::op_state`, allowing `ModuleMap` to have a sane default so
`JsRuntime::module_map` no longer needs to be optional.
This cleans up `JsRuntime` a bit more:
* We no longer print cargo's rerun-if-changed messages in `JsRuntime` --
those are printed elsewhere
* We no longer special case the OwnedIsolate for snapshots. Instead we
make use of an inner object that has the `Drop` impl and allows us to
`std::mem::forget` it if we need to extract the isolate for a snapshot
* The `snapshot` method is only available on `JsRuntimeForSnapshot`, not
`JsRuntime`.
* `OpState` construction is slightly cleaner, though I'd still like to
extract more
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
Rather than disallowing `ext:` resolution, clear the module map after
initializing extensions so extension modules are anonymized. This
operation is explicitly called in `deno_runtime`. Re-inject `node:`
specifiers into the module map after doing this.
Fixes #17717.
Dispatches op per-realm, and allows JsRealm to be garbage collected.
Slight improvement to benchmarks, but opens opportunity to clean up event loop.
---------
Co-authored-by: Matt Mastracci <matthew@mastracci.com>
This commit ensures that JavaScript functions generated for registered
ops have proper names set up - the function name matches the name
of the op.
A test was added in `core/runtime.rs` that verifies this.
Closes https://github.com/denoland/deno/issues/19206
**THIS PR HAS GIT CONFLICTS THAT MUST BE RESOLVED**
This is the release commit being forwarded back to main for 1.33.4
Please ensure:
- [x] Everything looks ok in the PR
- [ ] The release has been published
To make edits to this PR:
```shell
git fetch upstream forward_v1.33.4 && git checkout -b forward_v1.33.4 upstream/forward_v1.33.4
```
Don't need this PR? Close it.
cc @levex
Co-authored-by: levex <levex@users.noreply.github.com>
Co-authored-by: Levente Kurusa <lkurusa@kernelstuff.org>
Partially supersedes #19016.
This migrates `spawn` and `spawn_blocking` to `deno_core`, and removes
the requirement for `spawn` tasks to be `Send` given our single-threaded
executor.
While we don't need to technically do anything w/`spawn_blocking`, this
allows us to have a single `JoinHandle` type that works for both cases,
and allows us to more easily experiment with alternative
`spawn_blocking` implementations that do not require tokio (ie: rayon).
Async ops (+~35%):
Before:
```
time 1310 ms rate 763358
time 1267 ms rate 789265
time 1259 ms rate 794281
time 1266 ms rate 789889
```
After:
```
time 956 ms rate 1046025
time 954 ms rate 1048218
time 924 ms rate 1082251
time 920 ms rate 1086956
```
HTTP serve (+~4.4%):
Before:
```
Running 10s test @ http://localhost:4500
2 threads and 10 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 68.78us 19.77us 1.43ms 86.84%
Req/Sec 68.78k 5.00k 73.84k 91.58%
1381833 requests in 10.10s, 167.36MB read
Requests/sec: 136823.29
Transfer/sec: 16.57MB
```
After:
```
Running 10s test @ http://localhost:4500
2 threads and 10 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 63.12us 17.43us 1.11ms 85.13%
Req/Sec 71.82k 3.71k 77.02k 79.21%
1443195 requests in 10.10s, 174.79MB read
Requests/sec: 142921.99
Transfer/sec: 17.31MB
```
Suggested-By: alice@ryhl.io
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
**THIS PR HAS GIT CONFLICTS THAT MUST BE RESOLVED**
This is the release commit being forwarded back to main for 1.33.3
Please ensure:
- [x] Everything looks ok in the PR
- [x] The release has been published
To make edits to this PR:
```shell
git fetch upstream forward_v1.33.3 && git checkout -b forward_v1.33.3 upstream/forward_v1.33.3
```
Don't need this PR? Close it.
cc @levex
Co-authored-by: Levente Kurusa <lkurusa@kernelstuff.org>
This now allows circular imports across extensions.
Instead of load + eval of all ESM files in declaration order, all files
are only loaded. Eval is done recursively by V8, only evaluating
files that are listed in `Extension::esm_entry_point` fields.
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
The root cause of denoland/deno_std#3320, I believe, is that
`pending_promise_rejections` is a `HashMap` whose entries are in
arbitrary order, and as a result either of the two errors (`AddrInUse`
and `TypeError`) may be selected when determining which one to report. I
changed the field to a `VecDeque` so that the first error (`AddrInUse`
in this case) is always selected.
1. Boxed `File` and `FileSystem` to allow more easily passing this
through the CLI code (as shown within this pr).
2. `StdFileResource` is now `FileResource`. `FileResource` now contains
an `Rc<dyn File>`.
**THIS PR HAS GIT CONFLICTS THAT MUST BE RESOLVED**
This is the release commit being forwarded back to main for 1.33.2
Please ensure:
- [x] Everything looks ok in the PR
- [x] The release has been published
To make edits to this PR:
```shell
git fetch upstream forward_v1.33.2 && git checkout -b forward_v1.33.2 upstream/forward_v1.33.2
```
Don't need this PR? Close it.
cc @levex
Co-authored-by: levex <levex@users.noreply.github.com>
Co-authored-by: Levente Kurusa <lkurusa@kernelstuff.org>
Fixes #18979.
This changes the predicate for allowing `ext:` specifier resolution from
`snapshot_loaded_and_not_snapshotting` to `ext_resolution_allowed` which
is only set to true during the extension module loading phase. Module
loaders as used in core
are now declared as `ExtModuleLoader` rather than `dyn ModuleLoader`.
This commits changes "deno_core" to use jemalloc allocator as an
allocator
for V8 array buffers. This greatly improves our GC characteristics as we
are using
a lot of short lived array buffers. They no longer go through the
expensive
malloc/free cycle using the default Rust allocator, but instead use
jemallocator's
memory pool.
As a result the flamegraphs for WS/HTTP server flamegraphs no longer
show
stacks for malloc/free around ops that use ZeroCopyBuf and &[u8].
---------
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
