Some tests were silently failing after #3358 and #3434 because pool.spawn_ok
was used which doesn't panic on errors. For reference, the failure looked like this:
thread '<unnamed>' panicked at 'assertion failed: match isolate.poll_unpin(cx) { Poll::Ready(Ok(_)) => true, _ => false, }', core/isolate.rs:1408:7
After landing #3358 the benchmarks exploded indicating problems with workers and deno_core_http_bench.
This PR dramatically fixes thread/syscall count that showed up on benchmarks. Thread count is not back to previous levels but difference went from hundreds/thousands to about ~50.
- removes global `RESOURCE_TABLE` - resource tables are now created per `Worker`
in `State`
- renames `CliResource` to `StreamResource` and moves all logic related
to it to `cli/ops/io.rs`
- removes `cli/resources.rs`
- adds `state` argument to `op_read` and `op_write` and consequently adds
`stateful_minimal_op` to `State`
- IMPORTANT NOTE: workers don't have access to process stdio - this is
caused by fact that dropping worker would close stdout for process
(because it's constructed from raw handle, which closes underlying file
descriptor on drop)
This patch does not work with the recent bundler changes (#3325).
Unfortunately I didn't merge master before landing this patch. It has
something to do with console.log not working inside the compiler worker.
This reverts commit fd62379eaf.
- removes global `RESOURCE_TABLE` - resource tables are now created per `Worker`
in `State`
- renames `CliResource` to `StreamResource` and moves all logic related
to it to `cli/ops/io.rs`
- removes `cli/resources.rs`
- adds `state` argument to `op_read` and `op_write` and consequently adds
`stateful_minimal_op` to `State`
- IMPORTANT NOTE: workers don't have access to process stdio - this is
caused by fact that dropping worker would close stdout for process
(because it's constructed from raw handle, which closes underlying file
descriptor on drop)
Towards simplifying (or better removing entirely) the CoreResource
trait. Resources should be any bit of privileged heap allocated memory
that needs to be referenced from JS, not very specific trait
implementations. Therefore CoreResource should be pushed towards being
as general as possible.
- Fixes cargo publish on deno_typescript, deno_cli_snapshots, and
deno_cli.
- Combines cli_snapshots and js into one directory.
- Extracts TS version at compile time rather than runtime
- Bumps version awkwardly - it was necessary to test end-to-end
publishing. Sorry.
- Adds git submodule deno_typescript/typescript
Instead of using core/snapshot_creator.rs, instead two crates are
introduced which allow building the snapshot during build.rs.
Rollup is removed and replaced with our own bundler. This removes
the Node build dependency. Modules in //js now use Deno-style imports
with file extensions, rather than Node style extensionless imports.
This improves incremental build time when changes are made to //js files
by about 40 seconds.
Deno.core.dispatch() used to push the "control" buf onto the shared
array buffer before calling into V8, with the idea that it was one less
argument to parse. Turns out there is no more overhead passing the
control ArrayBuffer directly over. Furthermore this optimization was
making the refactors outlined in #2730 more complex. Therefore it is
being removed.
The rules are now as follows:
* In `import` statements, as mandated by the WHATWG specification,
the import specifier is always treated as a URL.
If it is a relative URL, it must start with either / or ./ or ../
* A script name passed to deno as a command line argument may be either
an absolute URL or a local path.
- If the name starts with a valid URI scheme followed by a colon, e.g.
'http:', 'https:', 'file:', 'foo+bar:', it always interpreted as a
URL (even if Deno doesn't support the indicated protocol).
- Otherwise, the script name is interpreted as a local path. The local
path may be relative, and operating system semantics determine how
it is resolved. Prefixing a relative path with ./ is not required.
1. Separate Snapshot and Script StartupData functions based on cfg "no-snapshot-init"
2. Replace deprecated Once::ONCE_INIT with Once::new (https://github.com/rust-lang/rust/pull/61757)
3. Elide lifetime
4. Fix typos
It's unnecessary indirection and is preventing the ability to easily
pass isolate references into the dispatch and dyn_import closures.
Note: this changes how StartupData::Script is executed. It's no longer done
during Isolate::new() but rather lazily on first poll or execution.
This patch makes it so that RecursiveLoad doesn't own the Isolate, so
Worker::execute_mod_async does not consume itself.
Previously Worker implemented Loader, but now ThreadSafeState does.
This is necessary preparation work for dynamic import (#1789) and import
maps (#1921)
* Compiler no longer has its own Tokio runtime. Compiler handles one
message and then exits.
* Uses the simpler ts.CompilerHost interface instead of
ts.LanguageServiceHost.
* avoids recompiling the same module by introducing a hacky but simple
`hashset<string>` that stores the module names that have been already
compiled.
* Removes the CompilerConfig op.
* Removes a lot of the mocking stuff in compiler.ts like `this._ts`. It
is not useful as we don't even have tests.
* Turns off checkJs because it causes fmt_test to die with OOM.
This patch provides a work-around for an apparent V8 bug where
initializing multiple isolates concurrently leads to a crash on
Windows.
At the time of writing the cause of this crash is not exactly
understood, but it seems to be related to the V8 internal
function win64_unwindinfo::RegisterNonABICompliantCodeRange(),
which didn't exist in older versions of V8.
* In order to prevent ArrayBuffers from getting garbage collected by V8,
we used to store a v8::Persistent<ArrayBuffer> in a map. This patch
introduces a custom ArrayBuffer allocator which doesn't use Persistent
handles, but instead stores a pointer to the actual ArrayBuffer data
alongside with a reference count. Since creating Persistent handles
has quite a bit of overhead, this change significantly increases
performance. Various HTTP server benchmarks report about 5-10% more
requests per second than before.
* Previously the Persistent handle that prevented garbage collection had
to be released manually, and this wasn't always done, which was
causing memory leaks. This has been resolved by introducing a new
`PinnedBuf` type in both Rust and C++ that automatically re-enables
garbage collection when it goes out of scope.
* Zero-copy buffers are now correctly wrapped in an Option if there is a
possibility that they're not present. This clears up a correctness
issue where we were creating zero-length slices from a null pointer,
which is against the rules.
Op dispatch is now dynamically dispatched, so slightly less efficient.
The immeasurable perf hit is a reasonable trade for the API simplicity
that is gained here.
Additionally, instead of polling ops in a loop until none of them are
ready, the isolate will now yield to the task system after delivering
the first batch of completed ops to the javascript side.
Although this makes performance a bit worse (about 15% fewer
requests/second on the 'deno_core_http_bench' benchmark), we feel that
the advantages are worth it:
* It resolves the extremely high worst-case latency that we were seeing
on deno_core_http_bench, in particular when using the multi-threaded
Tokio runtime, which would sometimes exceed a full second.
* Before this patch, the implementation of Isolate::poll() had to loop
through all sub-futures and poll each one of them, which doesn't scale
well as the number of futures managed by the isolate goes up. This
could lead to poor performance when e.g. a server is servicing
thousands of connected clients.
This change is made in preparation for using FuturesUnordered to track
futures that are spawned by the isolate. FuturesUnordered sets up
notififications for every future that it finds to be not ready when
polled, which causes a crash if attempted outside of a task context.
