* test: re-enable nan test: typedarrays-test.js
Fixes#28414.
I've confirmed this fix wfm on Linux. Pushing into a PR to get CI to run
it out on Win and Mac platforms too.
* chore: clarify comment
* test: fix NAN test string alignment
* test: (wip) add ldflags, archive file for libc++
* test: (wip) add libc++ to CircleCI
* test: (wip) add llvm flags
* test: (wip) change ldflag syntax
* test: (wip) build libc++abi as static
* fix: correct ldflags
* test: add ld env
* fix: do not commit this
* test: add lld from src to circleci
* test: add lld link to ld
* chore: preserve third_party
* seems legit
* sam swears this works kinda sort of sometimes'
:
* build: add gn visibility patch
* chore: update patches
* build: check for flatten_relative_to = false
* build: upload zip files, add to release.js validation
* debug: what the hell gn
* build: add libcxx gni to lint ignore
Linting the file adjusted the licenses array, which only contains
one value, and causes the gn check to fail later
* build: also use nan-spec-runner flags on Windows
* build: add linked flags for win32 only
* build: build libc++ as source on win
* build: clean up patch, add -fPIC for IA32
* build: delete libcxx .a files from root
* build: rename libc++.zip, clean up upload per platform
* build: fix gni lint
* ci: add libcxx gen to circleci config
* build: correct libcxx-object syntax
Co-authored-by: Samuel Attard <sam@electronjs.org>
Co-authored-by: Charles Kerr <charles@charleskerr.com>
Co-authored-by: clavin <clavin@electronjs.org>
Co-authored-by: Samuel Attard <sattard@slack-corp.com>
Co-authored-by: PatchUp <73610968+patchup[bot]@users.noreply.github.com>
Co-authored-by: Samuel Attard <sam@electronjs.org>
The `vendor` directory is a remnant of times long since gone when we had
brightray and libchromiumcontent (never forget). It is currently the
hope to a single header file that afaics we do not need anymore and
randomly pyyaml gets cloned there but that is easily movable.
This commit removes all references to the vendor directory from scripts,
docs, helpers, etc. and removes the directory itself.
* chore: graceful handling of notes with sub-lists
Handle multine release notes that contain their own bullet points.
Also, if a release note begins with a bullet point, remove it because it
will confuse the markdown parser to have two bullet points.
* refactor: use js instead of boto for s3put
* replace merge-electron-checksums.py with js, remove boto
* fix auth
* actually pass the env to execute()
Trop annotations are in the form of "(Also in 7.3, 8, 9)" with links to
the sibling branches.
Previously seen in b43e601b83 but is now
free of optional chaining and nullish coalescing, to run on Node < 14 :)
* build: add arm64 macOS publish jobs
* update ci-release-build to run new arm64 macOS publish jobs
* fixup circleci config validate issues
Co-authored-by: John Kleinschmidt <jkleinsc@github.com>
* chore: add patches to prevent installation of non-arm pip packages
* chore: add patches for apple-silicon
* build: add apple silicon build
* ci: add testing of new arm binary
* chore: remove / update for upstreamed patches
* Skip content tracing on macos on arm
* build: ensure that spec native modules are rebuilt for arm64 on apple-silicon
* chore: fix patches
* chore: fix broken patch
* chore: fix arm64 DCHECK
* build: add MAS arm64 build
* build: disable arm2 tests
* chore: update patches
* build: actually build MAS version of apple silicon app
Co-authored-by: John Kleinschmidt <jkleinsc@github.com>
* chore: ensure release notes always come from Clerk
Now with tests!
* chore: move sinon devDependency into `spec-main`
* refactor: tweak note-spec variable for readability
PR numbers are almost always listed parenthetically in the commit message;
but when something is committed manually, it could be missing. This change
uses octokit's listPullRequestsAssociatedWithCommit() as a second approach
to finding a commit's PR.
Last night's Releases WG meeting noted that "re-enable pdf viewer" was
missing from the notes. This PR fixes that omission.
When zip contains files larger than 4GB 4bytes length headers are
not sufficient anymore. Zip64 defines an extra header 0x0001 which
may contain 8byte lengthed (16 exabytes) lengths [uncompressed and compressed].
Read this value when performing extra data cleaning and override the
bogus value if the header is available.
Read https://blog.yaakov.online/zip64-go-big-or-go-home/ for more
information on Zip64 extra header.
This is the first known implementation of zip64 stripping.
* build: speed up release builds by using workflows
* Don't get depot-tools twice
* Do not use source from cache
* Checkout boto and requests for release builds
* Make sure vendor dirs are persisted.
* Use git cache for releases.
* show sccache stats right after build
* Make messages easier for sudowoodo to consume.
Since electron zip are including build dates the checksum of each of
these zip files is time dependant. In order to fix this issue strip all
the dates contained in each of the zip entries.
This commit uploads the output of the windows toolchain profiler.
The windows toolchain profiler is aimed at allowing the compilation
of electron release artifacts using the same windows toolchain on user
builder. This combined with the ability of electron of outputting
deterministic build should allow in future to check electron build
artifacts against CI artifacts.
See: 97959b5e5c include windows toolchain profiler (contd #20949)
* chore: add code required to use chromes spellchecker
* chore: fix linting
* chore: manifests needs buildflags now
* chore: add dictionarySuggestions to the context menu event when the spellchecker is active
* chore: enable by default for windows builds
* chore: add patch to remove incognito usage in the spellchecker
* chore: add dependencies on spellcheck common and flags
* chore: conditionally include spell check panel impl
* chore: fix deps for spellcheck feature flags
* chore: add patch for electron resources
* chore: add dependency on //components/language/core/browser
* chore: patches to make hunspell work on windows
* build: collect hunspell dictionaries into a zip file and publish
* chore: clean up patches
* chore: add docs and set spell checker url method
* chore: fix error handling
* chore: fix hash logic
* build: update hunspell filename generator
* fix: default spellchecker list to the current system locale if we can
* docs: document the language getter
* chore: patch IDS_ resources for linux builds
* feat: add spellcheck webpref flag to disable the builtin spellchecker
* chore: fix docs typo
* chore: clean up spellchecker impl as per feedback
* remove unneeded deps
* build: add WOA release to list of releases
* Add job count info for sudowoodo
* Add verification of all assets
* Fix linting and add logic to wait before printing out results
* build: add zip manifest for Windows on Arm
* ci: add Windows On Arm testing
(cherry picked from commit 4064e1f4874ff7a37c52c2ad974f92418c7e71c4)
* Fix path to invoke CI on WOA hardware
* Explicitly call 7z.exe to unzip files
* Make sure GCLIENT_EXTRA_ARGS set for WOA builds get prepended on release build
* set proper arch for npm
* Try using Compress-Archive/Expand-Archive
* Revert "Try using Compress-Archive/Expand-Archive"
* disable woa hardware test for now
* fix: fix cache filename of issue comments
* fix: update octokit deprecated API use
* fix: "TypeError: fn is not a function" in notes.js
* chore: retryableFunc does not need to be async
* chore: simplify checkCache() operation param
* build: move zip manifest logic in zip_manifests dir
* build: remove unused get-version.py script
* chore: move all release/sudowoodo related scripts into script/releases
* chore: update paths to zip manifests in CI configs
* build: fix path to ci release build script for arm tests