Skip to content

Conversation

@github-actions
Copy link
Contributor

What's Changed

New Contributors

Full Changelog: v6.21.1...v6.21.2

mcollina and others added 30 commits July 22, 2024 10:10
Signed-off-by: Matteo Collina <hello@matteocollina.com>
Signed-off-by: Matteo Collina <hello@matteocollina.com>
Signed-off-by: Matteo Collina <hello@matteocollina.com>
Signed-off-by: Matteo Collina <hello@matteocollina.com>
* fix: restore externalized Node.js dep compatibility

Restore the ability to build Undici compatible with Node.js'
`configure --shared-builtin-undici/undici-path ...` build option.

Scopes the `hasApk` conditional to only cover the part that requires
`apk`.

Makes the WASM optimizer (binaryen) optional to allow building on
Linux distributions that do not package `binaryen` and must be able
to rebuild everything (including tooling) from source.

* ci: add workflow for externalized Node.js dep

Add a workflow to test building Undici in a way compatible with Node.js
built with `configure --shared-builtin-undici/undici-path ...`.

This configuration is used by downstream Node.js packagers (e.g. Fedora)
who require the ability to be able to build everything from source.
Signed-off-by: Matteo Collina <hello@matteocollina.com>
* fix: memory leak

Holding a reference to the stream in the finalization registry causes a memory leak, even when consuming the body.

* docs: add comment explaining the strong reference

* typo
Signed-off-by: Matteo Collina <hello@matteocollina.com>
…t a TTY

Signed-off-by: Matteo Collina <hello@matteocollina.com>
Signed-off-by: Matteo Collina <hello@matteocollina.com>
* test: add test for memory leak

* lint
The wasm-opt binary may be available in different place than the local
directory (`./wasm-opt`) – for example, in /usr/bin/wasm-opt.
Similarly to the other parametrized WASM build options,
use WASM_OPT environment variable to identify the wanted binary,
with fallback to the previous value.

Even with the environment variable available,
the `hasOptimizer` is kept to make the optimization optional.

Signed-off-by: Jan Staněk <jstanek@redhat.com>
Signed-off-by: Matteo Collina <hello@matteocollina.com>
Signed-off-by: Matteo Collina <hello@matteocollina.com>
Co-authored-by: Jean-Baptiste Richardet <JbIPS@users.noreply.github.com>
Co-authored-by: KaKa <23028015+climba03003@users.noreply.github.com>
Co-authored-by: Carlos Fuentes <me@metcoder.dev>
Co-authored-by: Mert Can Altin <mertgold60@gmail.com>
Co-authored-by: Carlos Fuentes <me@metcoder.dev>
(cherry picked from commit b725457)

Co-authored-by: Aras Abbasi <aras.abbasi@googlemail.com>
…blocking (#3495) (#3673)

* fix: refactor fast timers, fix UND_ERR_CONNECT_TIMEOUT on event loop blocking

* also test native timers

* improve the commentary of fasttimers

* revert changes in client-h1.js

* make fasttimers independent from performance.now

* less flaky?

(cherry picked from commit 8a09d77)

Co-authored-by: Aras Abbasi <aras.abbasi@googlemail.com>
(cherry picked from commit 3faf140)

Co-authored-by: Aras Abbasi <aras.abbasi@googlemail.com>
* fix: use fasttimers for all connection timeouts

* Apply suggestions from code review

* activate some tests

* also use fastTimers in connect.js

* fix tests

* fix tests

* fix tests

* fix: use native timeouts for TIMEOUT_IDLE, rename TIMEOUT_IDLE to TIMEOUT_KEEP_ALIVE (#3554)

* fix: use native timeouts for TIMEOUT_IDLE, rename TIMEOUT_IDLE to TIMEOUT_KEEP_ALIVE

* ensure that fast timers and native timers are set properly

* .

* .

* rename import

* more informative connection error

* ignore request-timeout binary file, rename clearAll to reset

* fix

* add test

* use queueMicrotask earlier in the socket callbacks

(cherry picked from commit dca0aa0)

Co-authored-by: Aras Abbasi <aras.abbasi@googlemail.com>
(cherry picked from commit 990df2c)

Co-authored-by: Aras Abbasi <aras.abbasi@googlemail.com>
…ass them faster (#3656) (#3679)

* test: less flaky timers acceptance test

* use fake timers

* tick 10

* one tick

---------

Co-authored-by: Carlos Fuentes <me@metcoder.dev>
(cherry picked from commit 5308d29)

Co-authored-by: Aras Abbasi <aras.abbasi@googlemail.com>
(cherry picked from commit 773ba01)

Co-authored-by: Khafra <maitken033380023@gmail.com>
…onses (#3662) (#3689)

(cherry picked from commit 2d524cf)

Co-authored-by: Aras Abbasi <aras.abbasi@googlemail.com>
Co-authored-by: Aras Abbasi <aras.abbasi@googlemail.com>
(cherry picked from commit 862c035)

Co-authored-by: Khafra <maitken033380023@gmail.com>
mcollina and others added 23 commits October 9, 2024 09:48
Signed-off-by: Matteo Collina <hello@matteocollina.com>
(cherry picked from commit 3d0ce67)

Co-authored-by: Aras Abbasi <aras.abbasi@googlemail.com>
(cherry picked from commit db8e642)

Co-authored-by: tsctx <91457664+tsctx@users.noreply.github.com>
(cherry picked from commit d6c44f3)

Co-authored-by: Carlos Fuentes <me@metcoder.dev>
…mise (#3716) (#3723)

(cherry picked from commit cda5f94)

Co-authored-by: Aras Abbasi <aras.abbasi@googlemail.com>
* web: mark as uncloneable when possible

This tells node that the marked instances from undici are not cloneable,
so that attempts to cloning those throw `DataCloneError`.

* test: add test cases for platform objects uncloneable

* fix: move to webidl

* fix: move it under webidl

* fixup: rename it to markAsUncloneable

* fix: mark more web instances uncloneable

* fixup

---------

Co-authored-by: Khafra <maitken033380023@gmail.com>
(cherry picked from commit 1ab2382)
(cherry picked from commit e5c242d)

Co-authored-by: tsctx <91457664+tsctx@users.noreply.github.com>
* fix: handle undefined deref of weakref socket

* exit early

---------

Co-authored-by: Aras Abbasi <aras.abbasi@googlemail.com>
(cherry picked from commit f98fbef)

Co-authored-by: hochoy <hochoy@users.noreply.github.com>
* fix: range end is zero-indexed

* tests: update tests for retry handler

(cherry picked from commit 2b81fbc)

Co-authored-by: D Trombett <d@trombett.org>
Signed-off-by: Matteo Collina <hello@matteocollina.com>
* fix(#3736): back-port 183f8e9 to v6.x

* Backport #3769: fix http2 test
* fix(#3817): send servername for SNI on TLS (#3821)

* fix(#3817): send servername for SNI on TLS

* fix: set host header to servername

* refactor: attach regardless

(cherry picked from commit b93a834)

* feat: missing interceptor

* fix: lint
* fix: sending formdata bodies with http2 (#3863)

(cherry picked from commit e49b575)

* fix: bad merge

---------

Co-authored-by: Khafra <maitken033380023@gmail.com>
… when http2 goaway (#3877)

Co-authored-by: 沈鸿飞 <shen.hongfei@outlook.com>
* Update return type

* add type tests

(cherry picked from commit a1fb2cc)

Co-authored-by: Qayyuum Harun <69181532+mqayyuum@users.noreply.github.com>
* Use crypto.randomInt()

Signed-off-by: Matteo Collina <hello@matteocollina.com>

* fixup

Signed-off-by: Matteo Collina <hello@matteocollina.com>

---------

Signed-off-by: Matteo Collina <hello@matteocollina.com>
Signed-off-by: Matteo Collina <hello@matteocollina.com>
@mcollina
Copy link
Member

Nope, this didn't work. I'll release manually.

@mcollina mcollina closed this Mar 12, 2025
@Uzlopak Uzlopak deleted the release/v6.21.2 branch August 21, 2025 11:32
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.