Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

build: set v8_enable_webassembly=false when lite mode is enabled #48248

Merged

Conversation

TerrorJack
Copy link
Contributor

We should set v8_enable_webassembly=false when lite mode is enabled, since v8_enable_lite_mode implies v8_jitless, and wasm currently doesn't work with jitless. This is automatically handled in gn, but seems to be not the case in gyp. Enabling lite mode without setting v8_enable_webassembly=false leads to "Warning: disabling flag --expose_wasm due to conflicting flags" at runtime.

We should set v8_enable_webassembly=false when lite mode is enabled,
since v8_enable_lite_mode implies v8_jitless, and wasm currently
doesn't work with jitless. This is automatically handled in gn, but
seems to be not the case in gyp. Enabling lite mode without setting
v8_enable_webassembly=false leads to "Warning: disabling flag
--expose_wasm due to conflicting flags" at runtime.
@nodejs-github-bot nodejs-github-bot added build Issues and PRs related to build files or the CI. needs-ci PRs that need a full CI run. labels May 30, 2023
@TerrorJack
Copy link
Contributor Author

The parallel/test-inspector-async-stack-traces-promise-then timeout failure in the macos job is likely a transient test failure that's irrelevant to this patch.

@lpinca lpinca added the request-ci Add this label to start a Jenkins CI on a PR. label May 30, 2023
@github-actions github-actions bot removed the request-ci Add this label to start a Jenkins CI on a PR. label May 30, 2023
@nodejs-github-bot
Copy link
Collaborator

@lpinca lpinca added the commit-queue Add this label to land a pull request using GitHub Actions. label Jun 1, 2023
@nodejs-github-bot nodejs-github-bot removed the commit-queue Add this label to land a pull request using GitHub Actions. label Jun 1, 2023
@nodejs-github-bot nodejs-github-bot merged commit aca3a10 into nodejs:main Jun 1, 2023
@nodejs-github-bot
Copy link
Collaborator

Landed in aca3a10

@TerrorJack TerrorJack deleted the v8-no-wasm-when-lite-mode branch June 1, 2023 18:17
targos pushed a commit that referenced this pull request Jun 4, 2023
We should set v8_enable_webassembly=false when lite mode is enabled,
since v8_enable_lite_mode implies v8_jitless, and wasm currently
doesn't work with jitless. This is automatically handled in gn, but
seems to be not the case in gyp. Enabling lite mode without setting
v8_enable_webassembly=false leads to "Warning: disabling flag
--expose_wasm due to conflicting flags" at runtime.

PR-URL: #48248
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Michaël Zasso <[email protected]>
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Yagiz Nizipli <[email protected]>
@targos targos mentioned this pull request Jun 4, 2023
franciszek-koltuniuk-red pushed a commit to franciszek-koltuniuk-red/node that referenced this pull request Jun 7, 2023
We should set v8_enable_webassembly=false when lite mode is enabled,
since v8_enable_lite_mode implies v8_jitless, and wasm currently
doesn't work with jitless. This is automatically handled in gn, but
seems to be not the case in gyp. Enabling lite mode without setting
v8_enable_webassembly=false leads to "Warning: disabling flag
--expose_wasm due to conflicting flags" at runtime.

PR-URL: nodejs#48248
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Michaël Zasso <[email protected]>
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Yagiz Nizipli <[email protected]>
danielleadams pushed a commit that referenced this pull request Jul 6, 2023
We should set v8_enable_webassembly=false when lite mode is enabled,
since v8_enable_lite_mode implies v8_jitless, and wasm currently
doesn't work with jitless. This is automatically handled in gn, but
seems to be not the case in gyp. Enabling lite mode without setting
v8_enable_webassembly=false leads to "Warning: disabling flag
--expose_wasm due to conflicting flags" at runtime.

PR-URL: #48248
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Michaël Zasso <[email protected]>
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Yagiz Nizipli <[email protected]>
MoLow pushed a commit to MoLow/node that referenced this pull request Jul 6, 2023
We should set v8_enable_webassembly=false when lite mode is enabled,
since v8_enable_lite_mode implies v8_jitless, and wasm currently
doesn't work with jitless. This is automatically handled in gn, but
seems to be not the case in gyp. Enabling lite mode without setting
v8_enable_webassembly=false leads to "Warning: disabling flag
--expose_wasm due to conflicting flags" at runtime.

PR-URL: nodejs#48248
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Michaël Zasso <[email protected]>
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Yagiz Nizipli <[email protected]>
ckerr added a commit to electron/electron that referenced this pull request Jul 19, 2023
…ng_a_new.patch

Xref: nodejs/node#48248

manually sync patch to minor upstream code shear
ckerr added a commit to electron/electron that referenced this pull request Jul 25, 2023
…ng_a_new.patch

Xref: nodejs/node#48248

manually sync patch to minor upstream code shear
ckerr added a commit to electron/electron that referenced this pull request Aug 7, 2023
…ng_a_new.patch

Xref: nodejs/node#48248

manually sync patch to minor upstream code shear
ckerr added a commit to electron/electron that referenced this pull request Aug 8, 2023
* chore: bump node in DEPS to v18.17.0

* chore: update build_modify_js2c_py_to_allow_injection_of_original-fs_and_custom_embedder_js.patch

Xref: nodejs/node#46930

manually sync patch to minor upstream code shear

* chore: update build_ensure_native_module_compilation_fails_if_not_using_a_new.patch

Xref: nodejs/node#48248

manually sync patch to minor upstream code shear

* chore: update fix_expose_the_built-in_electron_module_via_the_esm_loader.patch

Xref: nodejs/node#47824

chore: upstream func throwIfUnsupportedURLProtocol() has been removed, so no need to patch it

* chore: update api_pass_oomdetails_to_oomerrorcallback.patch

Xref: nodejs/node#47695

manually sync patch to minor upstream code shear

* chore: remove fix_prevent_changing_functiontemplateinfo_after_publish.patch

Xref: nodejs/node#46979 (upstreamed patch)

Xref: https://chromium-review.googlesource.com/c/v8/v8/+/2718147 (related)

* chore: update fix_adapt_debugger_tests_for_upstream_v8_changes.patch

Xref: nodejs/node#47274

manually sync patch to minor upstream code shear

some tests moved from sequential to parallel

* chore: remove fix_libc_buffer_overflow_in_string_view_ctor.patch

Xref: fix_libc_buffer_overflow_in_string_view_ctor.patch

patch is no longer needed due to upstream bump to ada 2.2.0

* chore: remove fix_preventing_potential_oob_in_ada_no_scheme_parsing.patch

Xref: nodejs/node#47339

patch is no longer needed due to upstream bump to ada 2.2.0

* chore: rebuild filenames.json

several files removed/added/changed upstream

* chore: update build_add_gn_build_files.patch

upstream dep histogram 0.11.7 moved its include path from src/ to include/

Xref: nodejs/node#47742

* chore: update fix_crypto_tests_to_run_with_bssl.patch

Xref: nodejs/node#47160

BoringSSL doesn't support BIO_s_secmem() (a secure heap variant of
BIO_s_mem()), so use BIO_s_mem() instead.

Related discussion of secure heap support in BoringSSL:
https://boringssl-review.googlesource.com/c/boringssl/+/54309

* fix: ftbfs in node dep ada

* fix: ftbfs in node dep uvwasi

* chore: rebuild patches

* chore: update fix_handle_boringssl_and_openssl_incompatibilities.patch

Upstream used `BIO_s_secmem()`, a secure heap variant of `BIO_s_mem()`.
BoringSSL doesn't support it, so this PR opts for `BIO_s_mem()` instead.

Upstream Node.js change that prompted this:
nodejs/node#47160

Related discussion of BoringSSL support of secure heap:
https://boringssl-review.googlesource.com/c/boringssl/+/54309

* fix: work around Node 18 isURL() regression

* chore: sort script/node-disabled-tests.json alphabetically

* test: add parallel/test-snapshot-argv1 to disabled list

test: add parallel/test-snapshot-namespaced-builtin to disabled list

We don't support that type of snapshotting at the moment.

* chore: disable flaky node test parallel/test-dgram-send-cb-quelches-error

fails upstream in v18.x on my box as well

* ci: ensure spawned node tests have ELECTRON_RUN_AS_NODE set

* fixup! fix: work around Node 18 isURL() regression

fix: infinite loop regression

* fixup! fix: work around Node 18 isURL() regression

* chore: patch fixtures/errors/force_colors.snapshot

The line numbers in the stacktrace from our v8 build don't match what
Node's tests are expecting, so update the stacktrace to match our build.

The specific numbers probably aren't t needed for the force_colors test,
which is trying to see whether or not the lines are greyed out. One option
is to upstream a test change to stop hardcoding the stacktrace.

* fixup! fix: work around Node 18 isURL() regression

fix; pull in upstream bugfix

* fixup! ci: ensure spawned node tests have ELECTRON_RUN_AS_NODE set

chore: do not inject ELECTRON_RUN_AS_NODE in test-assert-colors.js

* chore: disable flaky node test parallel/test-debugger-random-port-with-inspect-port

---------

Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com>
Co-authored-by: Charles Kerr <[email protected]>
Ceres6 pushed a commit to Ceres6/node that referenced this pull request Aug 14, 2023
We should set v8_enable_webassembly=false when lite mode is enabled,
since v8_enable_lite_mode implies v8_jitless, and wasm currently
doesn't work with jitless. This is automatically handled in gn, but
seems to be not the case in gyp. Enabling lite mode without setting
v8_enable_webassembly=false leads to "Warning: disabling flag
--expose_wasm due to conflicting flags" at runtime.

PR-URL: nodejs#48248
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Michaël Zasso <[email protected]>
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Yagiz Nizipli <[email protected]>
Ceres6 pushed a commit to Ceres6/node that referenced this pull request Aug 14, 2023
We should set v8_enable_webassembly=false when lite mode is enabled,
since v8_enable_lite_mode implies v8_jitless, and wasm currently
doesn't work with jitless. This is automatically handled in gn, but
seems to be not the case in gyp. Enabling lite mode without setting
v8_enable_webassembly=false leads to "Warning: disabling flag
--expose_wasm due to conflicting flags" at runtime.

PR-URL: nodejs#48248
Reviewed-By: Ben Noordhuis <[email protected]>
Reviewed-By: Michaël Zasso <[email protected]>
Reviewed-By: Richard Lau <[email protected]>
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Yagiz Nizipli <[email protected]>
win32ss pushed a commit to win32ss/supermium-electron that referenced this pull request Sep 24, 2023
* chore: bump node in DEPS to v18.17.0

* chore: update build_modify_js2c_py_to_allow_injection_of_original-fs_and_custom_embedder_js.patch

Xref: nodejs/node#46930

manually sync patch to minor upstream code shear

* chore: update build_ensure_native_module_compilation_fails_if_not_using_a_new.patch

Xref: nodejs/node#48248

manually sync patch to minor upstream code shear

* chore: update fix_expose_the_built-in_electron_module_via_the_esm_loader.patch

Xref: nodejs/node#47824

chore: upstream func throwIfUnsupportedURLProtocol() has been removed, so no need to patch it

* chore: update api_pass_oomdetails_to_oomerrorcallback.patch

Xref: nodejs/node#47695

manually sync patch to minor upstream code shear

* chore: remove fix_prevent_changing_functiontemplateinfo_after_publish.patch

Xref: nodejs/node#46979 (upstreamed patch)

Xref: https://chromium-review.googlesource.com/c/v8/v8/+/2718147 (related)

* chore: update fix_adapt_debugger_tests_for_upstream_v8_changes.patch

Xref: nodejs/node#47274

manually sync patch to minor upstream code shear

some tests moved from sequential to parallel

* chore: remove fix_libc_buffer_overflow_in_string_view_ctor.patch

Xref: fix_libc_buffer_overflow_in_string_view_ctor.patch

patch is no longer needed due to upstream bump to ada 2.2.0

* chore: remove fix_preventing_potential_oob_in_ada_no_scheme_parsing.patch

Xref: nodejs/node#47339

patch is no longer needed due to upstream bump to ada 2.2.0

* chore: rebuild filenames.json

several files removed/added/changed upstream

* chore: update build_add_gn_build_files.patch

upstream dep histogram 0.11.7 moved its include path from src/ to include/

Xref: nodejs/node#47742

* chore: update fix_crypto_tests_to_run_with_bssl.patch

Xref: nodejs/node#47160

BoringSSL doesn't support BIO_s_secmem() (a secure heap variant of
BIO_s_mem()), so use BIO_s_mem() instead.

Related discussion of secure heap support in BoringSSL:
https://boringssl-review.googlesource.com/c/boringssl/+/54309

* fix: ftbfs in node dep ada

* fix: ftbfs in node dep uvwasi

* chore: rebuild patches

* chore: update fix_handle_boringssl_and_openssl_incompatibilities.patch

Upstream used `BIO_s_secmem()`, a secure heap variant of `BIO_s_mem()`.
BoringSSL doesn't support it, so this PR opts for `BIO_s_mem()` instead.

Upstream Node.js change that prompted this:
nodejs/node#47160

Related discussion of BoringSSL support of secure heap:
https://boringssl-review.googlesource.com/c/boringssl/+/54309

* fix: work around Node 18 isURL() regression

* chore: sort script/node-disabled-tests.json alphabetically

* test: add parallel/test-snapshot-argv1 to disabled list

test: add parallel/test-snapshot-namespaced-builtin to disabled list

We don't support that type of snapshotting at the moment.

* chore: disable flaky node test parallel/test-dgram-send-cb-quelches-error

fails upstream in v18.x on my box as well

* ci: ensure spawned node tests have ELECTRON_RUN_AS_NODE set

* fixup! fix: work around Node 18 isURL() regression

fix: infinite loop regression

* fixup! fix: work around Node 18 isURL() regression

* chore: patch fixtures/errors/force_colors.snapshot

The line numbers in the stacktrace from our v8 build don't match what
Node's tests are expecting, so update the stacktrace to match our build.

The specific numbers probably aren't t needed for the force_colors test,
which is trying to see whether or not the lines are greyed out. One option
is to upstream a test change to stop hardcoding the stacktrace.

* fixup! fix: work around Node 18 isURL() regression

fix; pull in upstream bugfix

* fixup! ci: ensure spawned node tests have ELECTRON_RUN_AS_NODE set

chore: do not inject ELECTRON_RUN_AS_NODE in test-assert-colors.js

* chore: disable flaky node test parallel/test-debugger-random-port-with-inspect-port

---------

Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com>
Co-authored-by: Charles Kerr <[email protected]>
MrHuangJser pushed a commit to MrHuangJser/electron that referenced this pull request Dec 11, 2023
* chore: bump node in DEPS to v18.17.0

* chore: update build_modify_js2c_py_to_allow_injection_of_original-fs_and_custom_embedder_js.patch

Xref: nodejs/node#46930

manually sync patch to minor upstream code shear

* chore: update build_ensure_native_module_compilation_fails_if_not_using_a_new.patch

Xref: nodejs/node#48248

manually sync patch to minor upstream code shear

* chore: update fix_expose_the_built-in_electron_module_via_the_esm_loader.patch

Xref: nodejs/node#47824

chore: upstream func throwIfUnsupportedURLProtocol() has been removed, so no need to patch it

* chore: update api_pass_oomdetails_to_oomerrorcallback.patch

Xref: nodejs/node#47695

manually sync patch to minor upstream code shear

* chore: remove fix_prevent_changing_functiontemplateinfo_after_publish.patch

Xref: nodejs/node#46979 (upstreamed patch)

Xref: https://chromium-review.googlesource.com/c/v8/v8/+/2718147 (related)

* chore: update fix_adapt_debugger_tests_for_upstream_v8_changes.patch

Xref: nodejs/node#47274

manually sync patch to minor upstream code shear

some tests moved from sequential to parallel

* chore: remove fix_libc_buffer_overflow_in_string_view_ctor.patch

Xref: fix_libc_buffer_overflow_in_string_view_ctor.patch

patch is no longer needed due to upstream bump to ada 2.2.0

* chore: remove fix_preventing_potential_oob_in_ada_no_scheme_parsing.patch

Xref: nodejs/node#47339

patch is no longer needed due to upstream bump to ada 2.2.0

* chore: rebuild filenames.json

several files removed/added/changed upstream

* chore: update build_add_gn_build_files.patch

upstream dep histogram 0.11.7 moved its include path from src/ to include/

Xref: nodejs/node#47742

* chore: update fix_crypto_tests_to_run_with_bssl.patch

Xref: nodejs/node#47160

BoringSSL doesn't support BIO_s_secmem() (a secure heap variant of
BIO_s_mem()), so use BIO_s_mem() instead.

Related discussion of secure heap support in BoringSSL:
https://boringssl-review.googlesource.com/c/boringssl/+/54309

* fix: ftbfs in node dep ada

* fix: ftbfs in node dep uvwasi

* chore: rebuild patches

* chore: update fix_handle_boringssl_and_openssl_incompatibilities.patch

Upstream used `BIO_s_secmem()`, a secure heap variant of `BIO_s_mem()`.
BoringSSL doesn't support it, so this PR opts for `BIO_s_mem()` instead.

Upstream Node.js change that prompted this:
nodejs/node#47160

Related discussion of BoringSSL support of secure heap:
https://boringssl-review.googlesource.com/c/boringssl/+/54309

* fix: work around Node 18 isURL() regression

* chore: sort script/node-disabled-tests.json alphabetically

* test: add parallel/test-snapshot-argv1 to disabled list

test: add parallel/test-snapshot-namespaced-builtin to disabled list

We don't support that type of snapshotting at the moment.

* chore: disable flaky node test parallel/test-dgram-send-cb-quelches-error

fails upstream in v18.x on my box as well

* ci: ensure spawned node tests have ELECTRON_RUN_AS_NODE set

* fixup! fix: work around Node 18 isURL() regression

fix: infinite loop regression

* fixup! fix: work around Node 18 isURL() regression

* chore: patch fixtures/errors/force_colors.snapshot

The line numbers in the stacktrace from our v8 build don't match what
Node's tests are expecting, so update the stacktrace to match our build.

The specific numbers probably aren't t needed for the force_colors test,
which is trying to see whether or not the lines are greyed out. One option
is to upstream a test change to stop hardcoding the stacktrace.

* fixup! fix: work around Node 18 isURL() regression

fix; pull in upstream bugfix

* fixup! ci: ensure spawned node tests have ELECTRON_RUN_AS_NODE set

chore: do not inject ELECTRON_RUN_AS_NODE in test-assert-colors.js

* chore: disable flaky node test parallel/test-debugger-random-port-with-inspect-port

---------

Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com>
Co-authored-by: Charles Kerr <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
build Issues and PRs related to build files or the CI. needs-ci PRs that need a full CI run.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

7 participants