This changes the error handling model of ServerHttp2Stream,
ServerHttp2Request and ServerHttp2Response.
An 'error' emitted on ServerHttp2Stream will not go to
'uncaughtException' anymore, but to the server 'streamError'.
On the stream 'error', ServerHttp2Request will emit 'abort', while
ServerHttp2Response would do nothing.
It also updates respondWith* to the new error handling.
Fixes: https://github.com/nodejs/node/issues/14963
PR-URL: https://github.com/nodejs/node/pull/14991
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com>
The benchmark script for dns contained functions with args declared
but never used. This fix removes those arguments from the function
signatures.
No test existed for the dns benchmark so one was added to the
parallel suite.
To improve performance the tests are limited to 1 invocation to a
single endpoint.
PR-URL: https://github.com/nodejs/node/pull/14936
Reviewed-By: Rich Trott <rtrott@gmail.com>
Reviewed-By: Ruben Bridgewater <ruben@bridgewater.de>
Reviewed-By: Refael Ackermann <refack@gmail.com>
This changes the error handling model of ServerHttp2Stream,
ServerHttp2Request and ServerHttp2Response.
An 'error' emitted on ServerHttp2Stream will not go to
'uncaughtException' anymore, but to the server 'streamError'.
On the stream 'error', ServerHttp2Request will emit 'abort', while
ServerHttp2Response would do nothing
See: https://github.com/nodejs/node/issues/14963
PR-URL: https://github.com/nodejs/node/pull/14991
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com>
The `double` fields in `performance_state` could previously have
been aligned at 4-byte instead of 8-byte boundaries, which would
have made creating an Float64Array them as a array buffer view
for an ArrayBuffer extending over the entire struct an invalid
operation.
Ref: 67269fd7f3
Comments out related flaky failure
PR-URL: https://github.com/nodejs/node/pull/14996
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Refael Ackermann <refack@gmail.com>
Node.js exposes `console.group()` and `console.groupEnd()` via the
inspector. These functions have no apparent effect when called from
Node.js without the inspector. We cannot easily hide them when Node.js
is started without the inspector because we support opening the
inspector during runtime via `inspector.port()`.
Implement a minimal `console.group()`/`console.groupEnd()`. More
sophisticated implementations are possible, but they can be done in
userland and/or features can be added to this at a later time.
`console.groupCollapsed()` is implemented as an alias for
`console.group()`.
PR-URL: https://github.com/nodejs/node/pull/14910
Fixes: https://github.com/nodejs/node/issues/1716
Ref: https://github.com/nodejs/node/issues/12675
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Timothy Gu <timothygu99@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
An initial implementation of the Performance Timing API for Node.js.
This is the same Performance Timing API implemented by modern browsers
with a number of Node.js specific properties. The User Timing mark()
and measure() APIs are implemented, garbage collection timing, and
node startup milestone timing.
```js
const { performance } = require('perf_hooks');
performance.mark('A');
setTimeout(() => {
performance.mark('B');
performance.measure('A to B', 'A', 'B');
const entry = performance.getEntriesByName('A to B', 'measure')[0];
console.log(entry.duration);
}, 10000);
```
The implementation is at the native layer and makes use of uv_hrtime().
This should enable *eventual* integration with things like Tracing
and Inspection.
The implementation is extensible and should allow us to add new
performance entry types as we go (e.g. for measuring i/o perf,
etc).
Documentation and a test are provided.
PR-URL: https://github.com/nodejs/node/pull/14680
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
`test-fs-chmod` modifies the permissions on files in `test/fixtures`.
This change has the test use the temp directory instead. One of the
fixture files is not used by any other test, so it has been deleted.
I took this opportunity to remove `console.log()` statements from the
test.
PR-URL: https://github.com/nodejs/node/pull/14926
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
* ensure that UV_... props are constants
* improve error message ... the error message when passing
in `err >= 0` to `util._errnoException()` was less than
useful.
* refine uses of process.binding('uv')
PR-URL: https://github.com/nodejs/node/pull/14933
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
The global `process` object had its prototype replaced with a
fresh object that had `EventEmitter.prototype` as its prototype.
With this change, the original `process.constructor.prototype` is
modified to have `EventEmitter.prototype` as its prototype, reflecting
that `process` objects are also `EventEmitter`s.
Fixes: https://github.com/nodejs/node/issues/14699
PR-URL: https://github.com/nodejs/node/pull/14715
Reviewed-By: Anna Henningsen <anna@addaleax.net>
* id values of -1 are allowed. They indicate that the id was never
correctly assigned to the async resource. These will appear in any
call graph, and will only be apparent to those using the async_hooks
module, then reported in an issue.
* ids < -1 are still not allowed and will cause the application to
exit the process; because there is no scenario where this should ever
happen.
* Add asyncId range checks to emitAfterScript().
* Fix emitBeforeScript() range checks which should have been || not &&.
* Replace errors with entries in internal/errors.
* Fix async_hooks tests that check for exceptions to match new
internal/errors entries.
NOTE: emit{Before,After,Destroy}() must continue to exit the process
because in the case of an exception during hook execution the state of
the application is unknowable. For example, an exception could cause a
memory leak:
const id_map = new Map();
before(id) {
id_map.set(id, /* data object or similar */);
},
after(id) {
throw new Error('id never dies!');
id_map.delete(id);
}
Allowing a recoverable exception may also cause an abort because of a
stack check in Environment::AsyncHooks::pop_ids() that verifies the
async id and pop'd ids match. This case would be more difficult to debug
than if fatalError() (lib/async_hooks.js) was called immediately.
try {
async_hooks.emitBefore(null, NaN);
} catch (e) { }
// do something
async_hooks.emitAfter(5);
It also allows an edge case where emitBefore() could be called twice and
not have the pop_ids() CHECK fail:
try {
async_hooks.emitBefore(5, NaN);
} catch (e) { }
async_hooks.emitBefore(5);
// do something
async_hooks.emitAfter(5);
There is the option of allowing mismatches in the stack and ignoring the
check if no async hooks are enabled, but I don't believe going this far
is necessary.
PR-URL: https://github.com/nodejs/node/pull/14722
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Refael Ackermann <refack@gmail.com>
v8::MaybeLocal::ToLocalChecked() will abort on an empty handle.
AsyncWrap::MakeCallback returns an empty handle if the domain is
disposed, but this should not cause the process to abort. So instead
return v8::Undefined() and allow the error to be handled normally.
PR-URL: https://github.com/nodejs/node/pull/14722
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Refael Ackermann <refack@gmail.com>
On macOS, the parent process might not receive a message if it
is sent to soon, and then subsequent messages are also sometimes not
received.
(Is this a bug or expected operating system behavior like the
way a file watcher is returned before it's actually watching the file
system on/ macOS?)
Send a second message after a delay on macOS.
While at it, minor refactoring to the test:
* Blank line after loading `common` module per test-writing guide
* Wrap arrow function in braces where implicit return is not needed
* Remove unnecessary unref in subprocess
PR-URL: https://github.com/nodejs/node/pull/14780
Fixes: https://github.com/nodejs/node/issues/14747
Reviewed-By: Anna Henningsen <anna@addaleax.net>
