[wasm] Bump Emscripten to 5.0.6 (6ea9c28c)#240
Open
pavelsavara wants to merge 4891 commits intodotnet:dotnet/mainfrom
Open
[wasm] Bump Emscripten to 5.0.6 (6ea9c28c)#240pavelsavara wants to merge 4891 commits intodotnet:dotnet/mainfrom
pavelsavara wants to merge 4891 commits intodotnet:dotnet/mainfrom
Conversation
`emscripten_websocket_new` uses `$webSockets` but does not declare it as a dependency. Fixes: emscripten-core#26310
…emscripten-core#26343) When using the `--failing-and-slow-first` flag on the CI, treat failure to parse the previous run log file as a non-fatal error, to help forcibly terminated CI jobs recover.
When EMTEST_AUTOSKIP=1 and current Node.js version is less than 25, then skip the Node.js 25 requiring tests automatically.
) emscripten-core#26297 didn't make it into the 5.0.2 tag. This pull request moves the ChangeLog entry to 5.0.3.
…mscripten-core#26345) This removes the dependency that libc++abi has on the pthread API functions, allowing better support for Wasm Workers. Replaces: emscripten-core#26283 Fixes: emscripten-core#26277
When building -sSINGLE_FILE builds with -sSINGLE_FILE_BINARY_ENCODE enabled, the base64Decode function was being emitted, and Closure was unable to DCE away the dependent base64ReverseLookup variable. So optimize this away manually to save code size. --------- Co-authored-by: juj <clb@h12dsi.clbri.com>
This file seems like it was previously referenced in the docs prior to 6dc95a7 but never actually built or used in any test.
Also, add more assertions about the return value.
) This file should have been removed in 11e404d.
… NFC (emscripten-core#26359) This works for over 100 reftests in test_browser.py
Support multi-threaded applications in memoryprofiler. Fixes emscripten-core#18107.
…n-core#26357) Should fix emscripten-core#26311 in the manner suggested by @sbc100.
…fault. NFC (emscripten-core#26363) These are only ever used by EMSCRIPTEN_TRACING so we know exactly when to include them. I noticed this while reviewing emscripten-core#26175
…#26364) This allows for the complete removal of `libmemoryprofiler.js` since the memory profiler can now just use the same libtrace-based hooks for all events. I noticed this while reviewing emscripten-core#26175
These files should have been removed in emscripten-core#6827 (2018!)
…er (emscripten-core#26370) From what I gather, this header was only ever supported in Firefox, and now Firefox has removed it. https://bugzilla.mozilla.org/show_bug.cgi?id=1598759
…ipten-core#26371) Release notes: - Fixed [GH#27](pongasoft/emscripten-glfw#27): properly handles scale change (it was only working when switching back and forth between 2 scales) - Fixed [GH#29](pongasoft/emscripten-glfw#29): when the canvas does not have a `tabindex` attribute, the library automatically adds one to make sure that the callback, set via `glfwSetWindowFocusCallback`, is called when the canvas is focused/unfocused.
In order to measure a difference with the native base64 API we need to disable the new `-sSINGLE_FILE_BINARY_ENCODE` setting.
…en-core#26368) This file as added back in 1326270 and apparently comes from gnulib. See https://github.com/mirror/gnulib/blob/master/lib/float%2B.h for the latest version.
…#26367) Also more a few other tests into `other`. Followup to emscripten-core#26350 After this change we are down to under 300 files in the top level test dir.
…26360) This is the only file that uses it so we don't need to it in the shared utils.
The code here previously used a pthread_lock. However this is unnecessary since all calls to the `do_tzset` function are already wrapped in `LOCK(lock)` and `UNLOCK(lock)`. This is nice little codesize saving too.
…core#26376) This just documents the current limitations. See emscripten-core#26375
…mscripten-core#24858) This change does not start using the executable launcher by default. That can happen later once we get more feedback on it. For example, its not clear if this will lead to issues with virus/security scanners. Assuming we don't run into any issues I hope to eventually completely remove support for the `.bat` files. There are several reasons to want to do this: 1. Batch files are notoriously painful to work with and mis-understood. 2. Should be faster (no need to launch cmd.exe). 3. Works around several known issues with .bat files including one that is known unsolvable one (see emcc.bat for more details) 4. In the future we could potentially launch clang directly, bypassing the python completely (or at least only invoking python for the link stage). See: emscripten-core#26229 See: emscripten-core#19207
This test broke when binaryen recently got more strict: WebAssembly/binaryen#8403
…6651) This is an automatic change generated by tools/maint/rebaseline_tests.py. The following (1) test expectation files were updated by running the tests with `--rebaseline`: ``` codesize/test_codesize_cxx_lto.json: 120519 => 120725 [+206 bytes / +0.17%] Average change: +0.17% (+0.17% - +0.17%) ``` Co-authored-by: emscripten-bot <emscripten-bot@users.noreply.github.com>
This avoids repeating the same set of decorators all over the place.
Under node we install the `uncaughtException` handler in workers (both in pthread and wasm worker). This means that `uncaughtException` is sent via `postMessage` back to the main thread. See nodejs/node#59617 for why we do this. However, these message was simply being ignored by the main thread in the case of Wasm Workers. With this change we honor the `uncaughtException` message and avoid completely loosing the uncaught exception.
This does not remove anything, just generates a warning if the setting is used. See emscripten-core#26647 and emscripten-core#26648
We dropped support for firefox older than 58 a long time back. We dropped support node older than 18.1 in emscripten-core#26604.
…6656) This is an automatic change generated by tools/maint/rebaseline_tests.py. The following (4) test expectation files were updated by running the tests with `--rebaseline`: ``` codesize/test_minimal_runtime_code_size_hello_webgl2_wasm.json: 13200 => 13200 [+0 bytes / +0.00%] codesize/test_minimal_runtime_code_size_hello_webgl2_wasm2js.json: 18534 => 18537 [+3 bytes / +0.02%] codesize/test_minimal_runtime_code_size_hello_webgl_wasm.json: 12738 => 12738 [+0 bytes / +0.00%] codesize/test_minimal_runtime_code_size_hello_webgl_wasm2js.json: 18060 => 18063 [+3 bytes / +0.02%] Average change: +0.01% (+0.00% - +0.02%) ``` Co-authored-by: emscripten-bot <emscripten-bot@users.noreply.github.com>
These minor cleanups were split out from my larger work on precise wakeups.
Fix the following errors after emscripten-core#26424: ``` - ERROR - [JSC_BAD_JSDOC_ANNOTATION] Parse error. illegal use of unknown JSDoc tag "noreturn"; ignoring it. Place another character before the @ to stop JSCompiler from parsing it as an annotation. - ERROR - [JSC_USED_GLOBAL_THIS] dangerous use of the global this object. ```
We are nearing the end of our budget, so for a few weeks we need to cut back.
…en-core#26658) It turns out that the files in `musl/arch/generic/` and `musl/arch/emscripten/` are supposed to me internal headers even though the `bits/` subdirectories are public. This change means that `syscall_arch.h` is not longer visible outside of the musl build (e.g. when building wasmfs) so I moved the public function declarations to `emscripten/syscalls.h`.
- Test using `pthread_kill` on yourself. - Test using `pthread_kill` from a background thread to the main thread. For some reason this was disallowed before. - Add a stub version of `pthread_kill` along with a test.
…pten-core#26665) After llvm/llvm-project#191068 it is no longer necessary (or correct).
This is useful when debugging tests that are timing out without having to wait the full default of 5 minutes to get the failure. For example I've been debugging a timeout recently using: ``` EMTEST_TIMEOUT=5 ./test/runner posixtest.test_pthread_join_4_1 ``` This waits for 5 seconds rather than 5 minutes.
This special handling for `SIGCANCEL` should not be needed. If we need to do something like `_emscripten_runtime_keepalive_clear` during pthread cancelation it would be best done when the cancellation in processed in `__testcancel`/`__cancel`. The comment & code I'm removing here was added back in emscripten-core#22467 along with testing in the form of `test/pthread/test_pthread_kill.c` and the pthread_kill tests in posixtest. I recently expanded the testing in emscripten-core#26663, and this change doesn't break any of those tests.
Also, add some testing for this. I also ran this test under native clang / glibc on my desktop and got the expected results.
…ripten-core#26673) This fix only applies to the hybrid Wasm Workers + pthread build mode. We can't currently rely on calling `pthread_self()` from a Wasm Worker, even the hybrid mode. We are working on fixing that in emscripten-core#26631. However, we do have a `CURRENT_THREAD_ID` macro that works on both Wasm Workers and pthreads. As it happens, pthread locks under musl do not normally depend on `pthread_self` except for "recursive" which need to access the `pthread_self()->tid` field to track the current owner. This means that I broke the debug + hybrid mode locks in emscripten-core#24607 because I added the recursion check to all locks in debug builds. While I have added a repro case for this its not directly testing the issue, it just heppens to crash due to the current undefined behavior when doing `pthread_self()->tid` from a Wasm Worker. Fixes: emscripten-core#26619
Merges emscripten-core/emscripten 5.0.6 (sha 6ea9c28). Dotnet patches re-applied: - .gitignore: dotnet build outputs - .gitmodules: empty (kept) - embuilder.py: sqlite3 patch dropped (upstream already removed sqlite3 from MINIMAL_TASKS) - package.json: devDependencies stripped (package-lock.json needs npm prune --production regen) - src/jsifier.js renamed by upstream to src/jsifier.mjs; CodeQL [js/bad-code-sanitization] suppression re-applied - tools/js_manipulation.py: CodeQL [SM03905] safe regex retained (auto-merged) - tools/link.py: STB_IMAGE error retained - tools/system_libs.py: libstb_image class disabled (auto-merged) Deletions retained: emrun.py, system/lib/stb_image.c, third_party/stb_image.h, third_party/ply/doc/makedoc.py, test/** (including newly-modified upstream test files). eng/emscripten-revision.txt bumped to 6ea9c28. emscripten-version.txt now reads 5.0.6-git. eng/Versions.props VersionPrefix unchanged (11.0.0). Note: package-lock.json taken from upstream; must be regenerated with npm install --omit=dev to match the dev-dep-free package.json before CI.
Matches the dev-dep-free package.json carried as a dotnet patch. Produced via: rm package-lock.json; npm install --omit=dev --package-lock-only with Node v24.14.0 / npm 11.9.0.
pavelsavara
changed the title
5.0.6 into dotnet/main
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
20 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Contributes to dotnet/runtime#113786
Summary
Bumps
dotnet/emscriptento upstreamemscripten-core/emscriptentag5.0.6at6ea9c28c38cd.Branch contents
https://github.com/pavelsavara/emscripten/commits/dotnet/main-bump-emscripten-5.0.6/
Validation
Microsoft.NETCore.Runtime.Wasm.Emscripten.Transport.11.0.0-alpha.1.26253.2.nupkg(22M)