repo
stringclasses
8 values
issue_number
int64
13k
155k
issue_title
stringlengths
3
139
issue_body
stringlengths
4
68.4k
commit_sha
stringlengths
40
40
files
listlengths
1
300
facebook/react
36,054
[eslint-plugin-react-hooks] Fix typescript typing
<!-- Thanks for submitting a pull request! We appreciate you spending the time to work on these changes. Please provide enough information so that others can review your pull request. The three fields below are mandatory. Before submitting a pull request, please make sure the following is done: 1. Fork [the repository](https://github.com/facebook/react) and create your branch from `main`. 2. Run `yarn` in the repository root. 3. If you've fixed a bug or added code that should be tested, add tests! 4. Ensure the test suite passes (`yarn test`). Tip: `yarn test --watch TestName` is helpful in development. 5. Run `yarn test --prod` to test in the production environment. It supports the same options as `yarn test`. 6. If you need a debugger, run `yarn test --debug --watch TestName`, open `chrome://inspect`, and press "Inspect". 7. Format your code with [prettier](https://github.com/prettier/prettier) (`yarn prettier`). 8. Make sure your code lints (`yarn lint`). Tip: `yarn linc` to only check changed files. 9. Run the [Flow](https://flowtype.org/) type checks (`yarn flow`). 10. If you haven't already, complete the CLA. Learn more about contributing: https://reactjs.org/docs/how-to-contribute.html --> ## Summary <!-- Explain the **motivation** for making this change. What existing problem does the pull request solve? --> ## How did you test this change? <!-- Demonstrate the code is solid. Example: The exact commands you ran and their output, screenshots / videos if the pull request changes the user interface. How exactly did you verify that your PR solves the issue you wanted to solve? If you leave this empty, your PR will very likely be closed. -->
0043ffbd1a015ddc799dbc9df5b6a93631b9942d
[ { "filename": "packages/eslint-plugin-react-hooks/src/index.ts", "patch": "@@ -58,7 +58,7 @@ const recommendedLatestRuleConfigs: Linter.RulesRecord = {\n const plugins = ['react-hooks'];\n \n type ReactHooksFlatConfig = {\n- plugins: {react: any};\n+ plugins: {'react-hooks': any};\n rules: Linter.RulesR...
vercel/next.js
92,210
Fix DashMap read-write self-deadlock in task_cache causing hangs
### What? Fixes a deadlock in the Turbopack backend's `task_cache` DashMap that causes `next` to hang during incremental builds with persistent caching. ### Why? In `get_or_create_persistent_task` and `get_or_create_transient_task`, `self.task_cache.get(&task_type)` returns a `dashmap::Ref` that holds a **read lock** on a DashMap shard. Although the `TaskId` was copied out with `let task_id = *task_id`, the original `Ref` was not dropped until the end of the `if let` block — which includes the call to `ConnectChildOperation::run`. `ConnectChildOperation::run` can trigger `AggregationUpdateQueue::process` → `for_each_task_all` → `prepare_tasks_with_callback`, which calls `task_cache.entry(task_type).or_insert(task_id)` — requiring a **write lock** on a `task_cache` shard. When this targets the same shard that the current thread already holds a read lock on, the thread self-deadlocks because dashmap's `RawRwLock` is not reentrant. Once one thread self-deadlocks, all other tokio worker threads needing that shard pile up behind it, eventually causing a complete system hang where every worker is stuck in `dashmap::lock::RawRwLock::lock_exclusive_slow`. This was observed as a hang during the 7th incremental build with persistent caching of a large application, with a macOS `sample` trace showing all tokio-runtime-worker threads blocked on `_pthread_cond_wait` inside `lock_exclusive_slow` on the `DashMap<Arc<CachedTaskType>, TaskId>`. ### How? Use `.map(|r| *r)` to copy the `TaskId` out and drop the `Ref` (releasing the read lock) before calling `ConnectChildOperation::run`: ```rust // Before: Ref lives until end of `if let` block if let Some(task_id) = self.task_cache.get(&task_type) { let task_id = *task_id; ConnectChildOperation::run(...); // read lock still held! } // After: Ref dropped by .map(), read lock released if let Some(task_id) = self.task_cache.get(&task_type).map(|r| *r) { ConnectChildOperation::run(...); // no lock held } ``` Applied to both `get_or_create_persistent_task` and `get_or_create_transient_task`.
36f2ee17a4303c3b4bce98c42206d3f71aea8068
[ { "filename": "turbopack/crates/turbo-tasks-backend/src/backend/mod.rs", "patch": "@@ -1532,8 +1532,9 @@ impl<B: BackingStorage> TurboTasksBackendInner<B> {\n // Create a single ExecuteContext for both lookup and connect_child\n let mut ctx = self.execute_context(turbo_tasks);\n // F...
nodejs/node
62,510
meta: require DCO signoff in commit message guidelines
Refs: https://github.com/nodejs/core-validate-commit/pull/141
804006bfda8f127fe504a6632f9192dedc81e2b7
[ { "filename": "doc/contributing/pull-requests.md", "patch": "@@ -204,6 +204,8 @@ A good commit message should describe what changed and why.\n Bot generated commits are exempt from this requirement. If a commit has\n multiple authors, the `Signed-off-by` line should be added for each author;\n and ...
electron/electron
50,645
fix: glitchy rendering and maximize behavior with different GTK themes
Backport of #50550 See that PR for details. Notes: Fixed the appearance of maximized windows on GNOME in Wayland, especially when non-default GTK themes like Breeze are set.
9b737636cb21922a17f48334b09135cec2dca2a8
[ { "filename": "shell/browser/ui/views/client_frame_view_linux.cc", "patch": "@@ -163,7 +163,7 @@ int ClientFrameViewLinux::ResizingBorderHitTest(const gfx::Point& point) {\n gfx::Rect ClientFrameViewLinux::GetBoundsForClientView() const {\n gfx::Rect client_bounds = bounds();\n if (!frame_->IsFullscreen...
huggingface/transformers
45,138
CI] Small T5 expectations updated
This is a small T5 expectations update. It is the same for both AMD and NVIDIA A10 GPUs.
7944af60ca6adb740f36cc909f89321a60dfe4c9
[ { "filename": "tests/models/t5/test_modeling_t5.py", "patch": "@@ -1058,7 +1058,7 @@ def test_small_v1_1_integration_test(self):\n loss = model(input_ids.to(torch_device), labels=labels.to(torch_device)).loss\n mtf_score = -(labels.shape[-1] * loss.item())\n \n- EXPECTED_SCORE = -40.1...
golang/go
77,841
runtime: unified the calculate for max objects per span
the freemark is used for mark the status for one span's slots. the size can be simplify by directly use the predefined constant of gc.MaxObjsPerSpan
d36aa731754912a697d8ca3d82079a067e64acf8
[ { "filename": "src/runtime/heapdump.go", "patch": "@@ -480,7 +480,7 @@ func dumproots() {\n \n // Bit vector of free marks.\n // Needs to be as big as the largest number of objects per span.\n-var freemark [pageSize / 8]bool\n+var freemark [gc.MaxObjsPerSpan]bool\n \n func dumpobjs() {\n \t// To protect mhe...
ollama/ollama
15,104
server: preserve raw manifest bytes during pull
pullModelManifest unmarshals the registry response into a Go struct then re-marshals with json.Marshal before writing to disk. When the registry's JSON formatting or field ordering differs from Go's output, the local SHA256 won't match the registry's Ollama-Content-Digest header, causing false "out of date" warnings. Preserve the raw bytes from the registry response and write them directly to disk so the local manifest is byte-for-byte identical to what the registry serves.
6f7b1d9d23f48cd85cfb36b86da2658ebf679fc0
[ { "filename": "server/images.go", "patch": "@@ -578,7 +578,7 @@ func PullModel(ctx context.Context, name string, regOpts *registryOptions, fn fu\n \n \tfn(api.ProgressResponse{Status: \"pulling manifest\"})\n \n-\tmf, err := pullModelManifest(ctx, n, regOpts)\n+\tmf, manifestData, err := pullModelManifest(c...
rust-lang/rust
154,743
Remove an unused `StableHash` impl.
r? @TaKO8Ki
c3d9f996c462b99ad29f20ff064eb49a622d755c
[ { "filename": "compiler/rustc_middle/src/ich/impls_syntax.rs", "patch": "@@ -4,7 +4,7 @@\n use rustc_ast as ast;\n use rustc_data_structures::stable_hasher::{HashStable, StableHasher};\n use rustc_hir as hir;\n-use rustc_span::{SourceFile, Symbol, sym};\n+use rustc_span::{Symbol, sym};\n use smallvec::Small...
vercel/next.js
92,209
turbopack: avoid duplicate matches for slash-ambiguous patterns
### What? - Deduplicate identical `PatternMatch` entries returned by `read_matches` - Preserve the earliest match position when collapsing duplicates so result ordering stays stable - Add a regression test for slash-ambiguous `path.join(...)`-style patterns ### Why? `path.join(...)` can produce multiple slash/no-slash pattern alternatives that resolve to the same real file. `read_matches` was returning each of those alternatives separately, which inflated the "matches N files" warning. In the reported case from #92201, 1,000 real files could be reported as 16,000 matches. ### How? - Make `PatternMatch` hashable - Collect `read_matches` output into a dedupe map keyed by the final match - Keep the lowest alternative index for stable sorting after deduplication - Extend `test_read_matches` to assert the slash-ambiguous pattern only returns unique files Closes NEXT- Fixes #92201
c3e78d6cfe0517256b3ceb10df3c10b644160af4
[ { "filename": "turbopack/crates/turbopack-core/src/resolve/pattern.rs", "patch": "@@ -1,17 +1,17 @@\n use std::{\n- collections::{VecDeque, hash_map::Entry},\n+ collections::{hash_map::Entry, VecDeque},\n mem::take,\n sync::LazyLock,\n };\n \n-use anyhow::{Result, bail};\n+use anyhow::{bail, R...
electron/electron
50,644
fix: glitchy rendering and maximize behavior with different GTK themes
Backport of #50550 See that PR for details. Notes: Fixed the appearance of maximized windows on GNOME in Wayland, especially when non-default GTK themes like Breeze are set.
7590363230fddfe768aa460e5367d0ae20369168
[ { "filename": "shell/browser/ui/views/client_frame_view_linux.cc", "patch": "@@ -163,7 +163,7 @@ int ClientFrameViewLinux::ResizingBorderHitTest(const gfx::Point& point) {\n gfx::Rect ClientFrameViewLinux::GetBoundsForClientView() const {\n gfx::Rect client_bounds = bounds();\n if (!frame_->IsFullscreen...
nodejs/node
62,505
meta: expand memory leak DoS criteria to all DoS
We have dedicated requirements about memory leaks when triaging DoS. These applies in generall to all types of DoS, and many recent reports about DoS attack vectors fail to meet them, resulting in a lot of extra back-and-forth in triaging. Clarify in the threat model by expanding these requirements to all DoS. Drive-by: clarify criteria of documented JavaScript behavior is that they are included in ECMA262. Also use "Node.js application developer" instead of "user" the refer to the party being vulnerable to avoid confusion. <!-- Before submitting a pull request, please read: - the CONTRIBUTING guide at https://github.com/nodejs/node/blob/HEAD/CONTRIBUTING.md - the commit message formatting guidelines at https://github.com/nodejs/node/blob/HEAD/doc/contributing/pull-requests.md#commit-message-guidelines For code changes: 1. Include tests for any bug fixes or new features. 2. Update documentation if relevant. 3. Ensure that `make -j4 test` (UNIX), or `vcbuild test` (Windows) passes. If you believe this PR should be highlighted in the Node.js CHANGELOG please add the `notable-change` label. Developer's Certificate of Origin 1.1 By making a contribution to this project, I certify that: (a) The contribution was created in whole or in part by me and I have the right to submit it under the open source license indicated in the file; or (b) The contribution is based upon previous work that, to the best of my knowledge, is covered under an appropriate open source license and I have the right under that license to submit that work with modifications, whether created in whole or in part by me, under the same open source license (unless I am permitted to submit under a different license), as indicated in the file; or (c) The contribution was provided directly to me by some other person who certified (a), (b) or (c) and I have not modified it. (d) I understand and agree that this project and the contribution are public and that a record of the contribution (including all personal information I submit with it, including my sign-off) is maintained indefinitely and may be redistributed consistent with this project or the open source license(s) involved. -->
a51256b9d8995c821f468f835c51aec8bede427b
[ { "filename": "SECURITY.md", "patch": "@@ -152,28 +152,33 @@ does not trust is considered a vulnerability:\n the correct use of Node.js APIs.\n * The unavailability of the runtime, including the unbounded degradation of its\n performance.\n-* Memory leaks qualify as vulnerabilities when all of the follo...
ollama/ollama
15,102
launch: skip context length warning for MLX models and show model name
We currently just check server context length to do warnings. With the MLX change we should skip that since max context length gets allocated automatically.
442bff6a5a1750ac996648801d885b139e1272ed
[ { "filename": "cmd/launch/launch_test.go", "patch": "@@ -1506,6 +1506,7 @@ func TestConfirmLowContextLength(t *testing.T) {\n \t\tstatusBody string\n \t\tstatusCode int\n \t\tshowParams string // Parameters field returned by /api/show\n+\t\tshowBody string // full JSON body for /api/show (over...
golang/go
77,816
cmd/go: fix toolchain escape via PATH in covermode
When a user has a local Go installation of 1.25.5 but their go.mod specifies 1.26.0, running a test command with coverage flags like go test -covermode=count -coverpkg=./... ./... triggers the GOTOOLCHAIN auto-switch mechanism to use the newer toolchain. However, b.CovData uses the hardcoded 'go' command to invoke the builtin tool covdata. Because GOTOOLCHAIN does not modify the system PATH, exec.LookPath resolves this hardcoded 'go' to the host's older Go version (e.g., 1.25.5). Since covdata is a builtin tool built and run dynamically by the go command, invoking the older go binary to handle the new toolchain's files causes a fatal mismatch. The build fails with an error like: compile: version "go1.26.0" does not match go tool version "go1.25.5..." This CL fixes the issue by replacing the hardcoded 'go' with the absolute path to the current toolchain's go executable, ensuring the correct toolchain context is maintained and preventing PATH escape.
3f679a5196c396643619cf6d29a805e62f75417c
[ { "filename": "src/cmd/go/internal/work/cover.go", "patch": "@@ -24,7 +24,8 @@ import (\n func (b *Builder) CovData(a *Action, cmdargs ...any) ([]byte, error) {\n \tcmdline := str.StringList(cmdargs...)\n \targs := append([]string{}, cfg.BuildToolexec...)\n-\targs = append(args, \"go\", \"tool\", \"covdata\...
huggingface/transformers
45,136
Fix #45127: Auto-fix diverged tie_word_embeddings config on save to prevent silent weight corruption
# What does this PR do? This PR fixes a bug in `PreTrainedModel.save_pretrained()` where `config.tie_word_embeddings` can be inconsistent with the actual weight state, leading to silent model corruption for downstream consumers. ### Problem After PEFT `merge_and_unload()` (typical scenario: Qwen, Llama, Mistral, etc.), `embed_tokens` and [lm_head](file:///d:/transformers/transformers/src/transformers/modeling_utils.py#2859-2987) weights are separated in memory with different values, but `config.tie_word_embeddings` remains `True`. Currently, [save_pretrained()](file:///d:/transformers/transformers/src/transformers/modeling_utils.py#3168-3452) performs **no validation** of `tie_word_embeddings` against the actual weight state — the incorrect config is written to `config.json` as-is. This causes two issues: 1. **Reloading via [from_pretrained](file:///d:/transformers/transformers/src/transformers/modeling_utils.py#3726-4219)**: The load-side safety check ([modeling_utils.py:2535-2547](file:///d:/transformers/transformers/src/transformers/modeling_utils.py#L2535-L2547)) detects the inconsistency and refuses to tie, emitting a warning each time — but the config is still semantically wrong. 2. **Downstream tool consumption** (GGUF converters, quantization scripts, etc.): These tools trust `tie_word_embeddings: true` in `config.json` directly, potentially causing **silent weight corruption** — one tensor overwrites the other, producing completely degraded outputs. ### Fix In [save_pretrained()](file:///d:/transformers/transformers/src/transformers/modeling_utils.py#3168-3452), before writing the config to disk, we detect whether the input/output embeddings have diverged. If so, we automatically set `config.tie_word_embeddings = False` and emit a warning. **Key safety considerations:** - **Only triggers when the output embedding key** (e.g., `lm_head.weight`) **is explicitly declared** in the model's [_tied_weights_keys](file:///d:/transformers/transformers/src/transformers/modeling_utils.py#2394-2507) mapping as tied to the input embedding. This prevents false positives on models like Pop2Piano, which uses `tie_word_embeddings=True` for decoder output scaling but does not declare `lm_head.weight` in its [_tied_weights_keys](file:///d:/transformers/transformers/src/transformers/modeling_utils.py#2394-2507) (it only ties `encoder.embed_tokens` and `decoder.embed_tokens` to [shared](file:///d:/transformers/transformers/tests/utils/test_modeling_utils.py#487-506)). - **Cross-device scenarios** (model parallelism / offloading) are skipped entirely to avoid false positives and potential OOM from implicit device copies. - **T5 family safety analysis:** - **T5**: Scaling is decoupled to an independent `scale_decoder_outputs` field (`configuration_t5.py:82-83`) and `tie_word_embeddings` is forced to `True`. Not affected. - **UMT5**: Config init forcibly overrides `tie_word_embeddings = True` — even if saved as `False`, it's restored on load. Not affected. - **LongT5, Pop2Piano, SwitchTransformers**: Still read `tie_word_embeddings` for scaling in forward, but these are guarded by the [_tied_weights_keys](file:///d:/transformers/transformers/src/transformers/modeling_utils.py#2394-2507) check described above. ### Changes **[modeling_utils.py](file:///d:/transformers/transformers/tests/utils/test_modeling_utils.py):** - Added weight divergence detection + auto-fix logic before config saving in [save_pretrained()](file:///d:/transformers/transformers/src/transformers/modeling_utils.py#3168-3452) - [_tied_weights_keys](file:///d:/transformers/transformers/src/transformers/modeling_utils.py#2394-2507) guard to only auto-fix when the output embedding is declared as tied to the input embedding - Cross-device: skip check (avoid false positives and potential OOM) - `NotImplementedError`: silently ignored (expected for vision/speech backbones) - Other exceptions: logged via `logger.debug` **[test_modeling_utils.py](file:///d:/transformers/transformers/tests/utils/test_modeling_utils.py):** - Added [test_save_pretrained_auto_fixes_diverged_tied_embeddings](file:///d:/transformers/transformers/tests/utils/test_modeling_utils.py#1605-1639): constructs a tied Llama model → simulates weight divergence (PEFT merge) → verifies saved config is corrected + warning is emitted + reloaded weights are preserved correctly <!-- Remove if not applicable --> Fixes # #45127 ## Code Agent Policy The Transformers repo is currently being overwhelmed by a large number of PRs and issue comments written by code agents. We are currently bottlenecked by our ability to review and respond to them. As a result, **we ask that new users do not submit pure code agent PRs** at this time. You may use code agents in drafting or to help you diagnose issues. We'd also ask autonomous "OpenClaw"-like agents not to open any PRs or issues for the moment. PRs that appear to be fully agent-written will probably be closed without review, and we may block users who do this repeatedly or maliciously. This is a rapidly-evolving situation that's causing significant shockwaves in the open-source community. As a result, this policy is likely to be updated regularly in the near future. For more information, please read [`CONTRIBUTING.md`](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md). - [x] I confirm that this is not a pure code agent PR. ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#create-a-pull-request), Pull Request section? - [x] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [x] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR. <!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @ @CyrilVallez @BenjaminBossan Models: - text models: @ArthurZucker @Cyrilvallez - vision models: @yonigozlan @molbap - audio models: @eustlb @ebezzam @vasqu - multimodal models: @zucchini-nlp - graph models: @clefourrier Library: - generate: @zucchini-nlp (visual-language models) or @gante (all others) - continuous batching: @remi-or @ArthurZucker @McPatate - pipelines: @Rocketknight1 - tokenizers: @ArthurZucker and @itazap - trainer: @SunMarc - attention: @vasqu @ArthurZucker @CyrilVallez - model loading (from pretrained, etc): @CyrilVallez - distributed: @3outeille @ArthurZucker - CIs: @ydshieh Integrations: - ray/raytune: @richardliaw, @amogkam - Big Model Inference: @SunMarc - quantization: @SunMarc - kernels: @drbh - peft: @BenjaminBossan @githubnemo Devices/Backends: - AMD ROCm: @ivarflakstad - Intel XPU: @IlyasMoutawwakil - Ascend NPU: @ivarflakstad Documentation: @stevhliu Research projects are not maintained and should be taken as is. -->
75609b5c47f09ceea94d13cd9fda767c2dbd35f2
[ { "filename": "docs/source/en/_toctree.yml", "patch": "@@ -89,6 +89,8 @@\n title: Experts backends\n - local: continuous_batching\n title: Continuous batching\n+ - local: continuous_batching_architecture\n+ title: Continuous batching architecture\n - local: paged_attention\n ...
rust-lang/rust
154,741
Enforce `#![warn(unreachable_pub)]` in compiletest source
The actual public API surface of compiletest is restricted to two specific modules, `cli` and `rustdoc_gui_test`. Other items have no reason to be `pub`. There should be no change to compiletest behaviour.
14c7788a167723fe95dd3bf1080508db0e250140
[ { "filename": "src/tools/compiletest/src/common.rs", "patch": "@@ -14,7 +14,7 @@ use crate::util::{Utf8PathBufExt, add_dylib_path, string_enum};\n \n string_enum! {\n #[derive(Clone, Copy, PartialEq, Debug)]\n- pub enum TestMode {\n+ pub(crate) enum TestMode {\n Pretty => \"pretty\",\n ...
vercel/next.js
92,205
Fix pr-status.js: retry on API errors and report timed-out jobs
### What? Fixes two bugs in `scripts/pr-status.js` that caused the script to silently report zero failures when there were actual CI failures. ### Why? When running `node scripts/pr-status.js 92080`, the script reported "No failed jobs found" despite the PR having real failures. Two root causes: 1. **Transient API errors silently swallowed.** `getAllJobs()` had a bare `catch { break }` that returned an empty array on any GitHub API error (e.g., HTTP 502). Since the GitHub Actions jobs API for large runs frequently returns transient 502s, this caused false "no failures" reports. 2. **`timed_out` and `startup_failure` jobs were invisible.** The script only checked for `conclusion === 'failure'`, but GitHub uses distinct conclusion values like `timed_out` (job exceeded timeout) and `startup_failure` (runner failed to start). These jobs fell through all filters and were silently omitted from reports. ### How? **Retry logic in `getAllJobs()`:** - Retries each paginated API call up to 3 times with 2s/4s backoff - If all retries fail on the first page, throws an error (no silent empty results) - If later pages fail after partial data is collected, warns and returns what was fetched **Broader failure detection with `FAILED_CONCLUSIONS`:** - Added a shared `FAILED_CONCLUSIONS` set: `{'failure', 'timed_out', 'startup_failure'}` - Used in `getFailedJobs()`, `categorizeJobs()`, and flaky test detection - Jobs with non-`failure` conclusions are annotated in the report table (e.g., "(timed_out)") - `getFailedJobs()` now also returns the `conclusion` field so downstream code can distinguish failure types
2a642671d724bf697835bb43341d9426ac897bbb
[ { "filename": "scripts/pr-status.js", "patch": "@@ -216,15 +216,17 @@ function getRunMetadata(runId) {\n )\n }\n \n+const FAILED_CONCLUSIONS = new Set(['failure', 'timed_out', 'startup_failure'])\n+\n function getFailedJobs(runId) {\n // Fetch all jobs first, then filter for failures in JS.\n // We ca...
electron/electron
50,643
fix: prefill native print dialog options on macOS with OOP printing
Backport of #50600 See that PR for details. Notes: Fixed an issue where custom options in `webContents.print()` did not prefill the print dialog on macOS.
d191cac698e434a29f171e847e7fc5f6857daa08
[ { "filename": "patches/chromium/printing.patch", "patch": "@@ -68,7 +68,7 @@ index f91857eb0b6ad385721b8224100de26dfdd7dd8d..45e8766fcb8d46d8edc3bf8d21d3f826\n : PdfRenderSettings::Mode::POSTSCRIPT_LEVEL3;\n }\n diff --git a/chrome/browser/printing/print_view_manager_base.cc b/chrome/brow...
nodejs/node
62,504
test_runner: add passed, attempt, and diagnostic to SuiteContext
null
3a1675c81569ccbda0445b4f316ef9ae63b4d889
[ { "filename": "doc/api/test.md", "patch": "@@ -3875,7 +3875,9 @@ added: v25.0.0\n \n * Type: {number}\n \n-Number of times the test has been attempted.\n+The attempt number of the test. This value is zero-based, so the first attempt is `0`,\n+the second attempt is `1`, and so on. This property is useful in ...
ollama/ollama
15,101
app: fix false "out of date" model warnings
file on disk) against the registry's Ollama-Content-Digest header. These never matched because PullModel re-serializes the manifest JSON before writing, producing different bytes than the registry's original. The fallback comparison (local modified_at vs upstream push time) was also broken: the generated TypeScript Time class discards the actual timestamp value, so Date parsing always produced NaN. Fix by moving the staleness comparison server-side where we have reliable access to both the local manifest file mtime and the upstream push time. The /api/v1/model/upstream endpoint now returns a simple `stale` boolean instead of raw digests for the frontend to compare. Also adds User-Agent to the CORS allowed headers for dev mode.
c2aa0bdb7470e3bd31444c71044db46511ebad12
[ { "filename": "app/ui/app/codegen/gotypes.gen.ts", "patch": "@@ -550,14 +550,12 @@ export class Error {\n }\n }\n export class ModelUpstreamResponse {\n- digest?: string;\n- pushTime: number;\n+ stale: boolean;\n error?: string;\n \n constructor(source: any = {}) {\n if ('string...
golang/go
77,778
testing: fix construction of the testing artifacts path
The existing implementation had a few problems: - It constructs a path which starts with a forward slash, which is then immediately rejected by filepath.Localize() as invalid. - It did not correctly remove the module path from the import path if the test was in the root package of the module: it would do the equivalent of strings.TrimPrefix("example.com/jdoe/loader", "example.com/jdoe/loader/"), which trims nothing, and leaves the full module name in the base. - Tests didn't check for any behaviors related to which artifact path was selected. Removed leading slash, handle tests in the root package by omitting the <pkg> component from the artifact dir path, and add tests. Fixes #77763
9c39392a691a81cef48adde88ab45dc40ed81105
[ { "filename": "src/testing/testing.go", "patch": "@@ -1363,6 +1363,21 @@ func (c *common) makeArtifactDir() (string, error) {\n \t\treturn c.makeTempDir()\n \t}\n \n+\tartifactBase := filepath.Join(artifactDir, c.relativeArtifactBase())\n+\tif err := os.MkdirAll(artifactBase, 0o777); err != nil {\n+\t\tretu...
facebook/react
36,052
Snyk fix 7437f92a28cf1a13f2aafbbf2f46141e
<!-- Thanks for submitting a pull request! We appreciate you spending the time to work on these changes. Please provide enough information so that others can review your pull request. The three fields below are mandatory. Before submitting a pull request, please make sure the following is done: 1. Fork [the repository](https://github.com/facebook/react) and create your branch from `main`. 2. Run `yarn` in the repository root. 3. If you've fixed a bug or added code that should be tested, add tests! 4. Ensure the test suite passes (`yarn test`). Tip: `yarn test --watch TestName` is helpful in development. 5. Run `yarn test --prod` to test in the production environment. It supports the same options as `yarn test`. 6. If you need a debugger, run `yarn test --debug --watch TestName`, open `chrome://inspect`, and press "Inspect". 7. Format your code with [prettier](https://github.com/prettier/prettier) (`yarn prettier`). 8. Make sure your code lints (`yarn lint`). Tip: `yarn linc` to only check changed files. 9. Run the [Flow](https://flowtype.org/) type checks (`yarn flow`). 10. If you haven't already, complete the CLA. Learn more about contributing: https://reactjs.org/docs/how-to-contribute.html --> ## Summary <!-- Explain the **motivation** for making this change. What existing problem does the pull request solve? --> ## How did you test this change? <!-- Demonstrate the code is solid. Example: The exact commands you ran and their output, screenshots / videos if the pull request changes the user interface. How exactly did you verify that your PR solves the issue you wanted to solve? If you leave this empty, your PR will very likely be closed. -->
351e7c4bb23eb60ead7b1d91f6e93763893e5217
[ { "filename": ".claude/instructions.md", "patch": "@@ -0,0 +1,46 @@\n+# React\n+\n+**Scope**: All code EXCEPT `/compiler/` (compiler has its own instructions).\n+\n+## Project Structure\n+\n+| Directory | Purpose |\n+|-----------|---------|\n+| `/packages/` | Publishable packages (react, react-dom, schedule...
electron/electron
50,642
ci: add Datadog metrics to clean-src-cache job
#### Description of Change Enhanced the `clean-src-cache` workflow to capture and report disk space metrics to Datadog. The workflow now: 1. Captures disk space metrics before cleanup (free space and total capacity for both cache volumes) 2. Performs the existing cleanup of files older than 15 days 3. Captures disk space metrics after cleanup 4. Sends all metrics to Datadog via the API, including: - Free space before/after cleanup - Space freed during cleanup - Total capacity - Tagged by volume (cross-instance-cache, win-cache) for easy filtering This enables monitoring of cache cleanup effectiveness and disk usage trends over time. #### Checklist - [x] PR description included - [x] I have built and tested this PR - [ ] `npm test` passes (N/A - workflow configuration change) - [ ] tests are changed or added (N/A - workflow configuration change) - [ ] relevant API documentation is updated (N/A - workflow configuration change) #### Release Notes Notes: none https://claude.ai/code/session_013bpDsZLrFDpWMiARNFH4z9
0e987bb67a30e5f18ed4bd967a961fd4ed2188fb
[ { "filename": ".github/workflows/clean-src-cache.yml", "patch": "@@ -16,6 +16,8 @@ jobs:\n runs-on: electron-arc-centralus-linux-amd64-32core\n permissions:\n contents: read\n+ env:\n+ DD_API_KEY: ${{ secrets.DD_API_KEY }}\n container:\n image: ghcr.io/electron/build:bc2f48b2...
vercel/next.js
92,204
[test] Reduce `elementByCssInstant` timeout to 1ms
Just checking if it's possible. Opus 4.6 claims that the timeout is racing PW's IPC channel. That would be a bug imo. But if it still passes with 1ms timeout, it's less likely a race.
74c4886d8d9ac76f23fd7ceb879ccad6ba8aeb75
[ { "filename": "test/lib/browsers/playwright.ts", "patch": "@@ -437,7 +437,7 @@ export class Playwright<TCurrent = undefined> {\n /** A replacement for the default `browser.elementByCss` that doesn't wait for the page to fire \"load\". */\n elementByCssInstant(selector: string, opts?: ElementByCssOpts) {...
nodejs/node
62,501
test_runner: add `getTestContext()` public API
Exposes `getTestContext()` function to access test context information from within tests and async operations. My use case was exposing a pino logger that writes to test diagnostics, but there can be many other usecases: ```mjs import { prettyFactory } from 'pino-pretty'; const pretty = prettyFactory({ colorize: true }); const asyncId = executionAsyncId(); const t = testResources.get(asyncId); t.diagnostic(pretty(obj).replace(/\n$/, '')); ```
28d1af603354e88bc022d4edf059ebd2f5a9b50a
[ { "filename": "doc/api/test.md", "patch": "@@ -3563,6 +3563,44 @@ Emitted when no more tests are queued for execution in watch mode.\n \n Emitted when one or more tests are restarted due to a file change in watch mode.\n \n+## `getTestContext()`\n+\n+<!-- YAML\n+added: REPLACEME\n+-->\n+\n+* Returns: {TestC...
rust-lang/rust
154,740
Rollup of 6 pull requests
Successful merges: - rust-lang/rust#154444 (rustdoc ICE fix: When collecting `Deref` impls with their targets, skip the negative ones) - rust-lang/rust#154590 (Make #[cfg] suggest any or all on #[cfg(a, b)]) - rust-lang/rust#154691 (core: Update the feature gate on `TryFrom<integer> for bool`) - rust-lang/rust#154697 (rustdoc: fix href of extern crates in search results) - rust-lang/rust#154728 (rustdoc: Improve internal function name) - rust-lang/rust#154732 (fix(std): avoid AT_MINSIGSTKSZ on uclibc targets) Failed merges: - rust-lang/rust#154722 (fix(lints): Improve `ill_formed_attribute_input` with better help message) <!-- homu-ignore:start --> r? @ghost [Create a similar rollup](https://bors.rust-lang.org/queue/rust?prs=154444,154590,154691,154697,154722,154728,154732) <!-- homu-ignore:end -->
53af5aba8ff98338c0747ff3169fe780e6c2ef09
[ { "filename": "library/std/src/sys/pal/unix/stack_overflow.rs", "patch": "@@ -305,7 +305,7 @@ mod imp {\n }\n \n /// Modern kernels on modern hardware can have dynamic signal stack sizes.\n- #[cfg(any(target_os = \"linux\", target_os = \"android\"))]\n+ #[cfg(all(any(target_os = \"linux\", tar...
ollama/ollama
15,100
anthropic: fix KV cache reuse degraded by tool call argument reordering
Use typed structs for tool call arguments instead of map[string]any to preserve JSON key order, which Go maps do not guarantee.
95311f607e180f5781a608343db6d94816ce989a
[ { "filename": "anthropic/anthropic.go", "patch": "@@ -68,7 +68,7 @@ type MessagesRequest struct {\n \tModel string `json:\"model\"`\n \tMaxTokens int `json:\"max_tokens\"`\n \tMessages []MessageParam `json:\"messages\"`\n-\tSystem any `json:\"system,...
facebook/react
36,050
fix(compiler): preserve HTML entities in JSX text by correctly detecting non-empty lines
null
8d7d13a5f123c80ccc4967532d9edba7992176b1
[ { "filename": "compiler/packages/babel-plugin-react-compiler/src/HIR/BuildHIR.ts", "patch": "@@ -3563,7 +3563,7 @@ function trimJsxText(original: string): string | null {\n let lastNonEmptyLine = 0;\n \n for (let i = 0; i < lines.length; i++) {\n- if (lines[i].match(/[^ \\t]/)) {\n+ if (lines[i].t...
electron/electron
50,641
fix: extension service workers not starting beyond first app launch
Backport of #50611 See that PR for details. Notes: Fixed certain DevTools extension panels not showing without a page reload.
e7e9b9781505cde074609e3002ee351a98004688
[ { "filename": "shell/browser/extensions/electron_extension_loader.cc", "patch": "@@ -28,6 +28,7 @@\n #include \"extensions/common/error_utils.h\"\n #include \"extensions/common/file_util.h\"\n #include \"extensions/common/manifest_constants.h\"\n+#include \"extensions/common/manifest_handlers/background_inf...
huggingface/transformers
45,135
Fix model saving corruption for dynamically untied embeddings
# What does this PR do? Fixes an issue where PEFT adapters applied independently to tied embeddings (`embed_tokens` and [lm_head](cci:1://file:///d:/transformers/transformers/src/transformers/modeling_utils.py:2858:4-2985:26)) cause silent model corruption upon reloading via `AutoModelForCausalLM.from_pretrained()`. **Root Cause:** When embeddings are untied dynamically during runtime (e.g., vocabulary resizing and independent PEFT merging), their tensor memory storage diverges. `PreTrainedModel.save_pretrained()` correctly saves both parameter tensors because [remove_tied_weights_from_state_dict()](cci:1://file:///d:/transformers/transformers/src/transformers/modeling_utils.py:391:0-468:21) sees they don't share identical storage. However, the model configuration saves `tie_word_embeddings = True`. Upon reloading, [from_pretrained()](cci:1://file:///d:/transformers/transformers/src/transformers/modeling_utils.py:3696:4-4188:20) sees `tie_word_embeddings=True` and aggressively re-ties the two embeddings by overwriting one parameter with the other, effectively destroying the independent delta weights. **Fix:** Included a check in [save_pretrained()](cci:1://file:///d:/transformers/transformers/src/transformers/modeling_utils.py:3167:4-3469:13): if `config.tie_word_embeddings` is `True` but `input_embeddings.weight.data_ptr() != output_embeddings.weight.data_ptr()`, it automatically flips `model.config.tie_word_embeddings` to `False` before saving the configuration mapping, preventing silent destruction on loading. <!-- Remove if not applicable --> Fixes # #45127 ## Code Agent Policy The Transformers repo is currently being overwhelmed by a large number of PRs and issue comments written by code agents. We are currently bottlenecked by our ability to review and respond to them. As a result, **we ask that new users do not submit pure code agent PRs** at this time. You may use code agents in drafting or to help you diagnose issues. We'd also ask autonomous "OpenClaw"-like agents not to open any PRs or issues for the moment. PRs that appear to be fully agent-written will probably be closed without review, and we may block users who do this repeatedly or maliciously. This is a rapidly-evolving situation that's causing significant shockwaves in the open-source community. As a result, this policy is likely to be updated regularly in the near future. For more information, please read [`CONTRIBUTING.md`](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md). - [x] I confirm that this is not a pure code agent PR. ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#create-a-pull-request), Pull Request section? - [x] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? *(Note: A minimal reproduction script is provided below to easily test and validate the fix, as it requires mocking PEFT adaptation)* Reproduction Script <details><summary>Click to view minimal repro (repro.py)</summary> ```python import torch import gc from transformers import AutoModelForCausalLM, AutoConfig def main(): # Make a tiny dummy model with tied embeddings config = AutoConfig.from_pretrained("Qwen/Qwen1.5-0.5B", trust_remote_code=True) config.hidden_size = 32 config.intermediate_size = 64 config.num_hidden_layers = 2 config.num_attention_heads = 4 config.num_key_value_heads = 4 config.vocab_size = 1000 config.tie_word_embeddings = True # Needs to be small to fit in memory model = AutoModelForCausalLM.from_config(config) print("Initial tie config:", model.config.tie_word_embeddings) print("Are weights tied initially?", id(model.get_input_embeddings().weight) == id(model.get_output_embeddings().weight)) # Suppose PEFT does something and they are no longer tied model.get_output_embeddings().weight = torch.nn.Parameter(model.get_output_embeddings().weight.clone()) model.get_output_embeddings().weight.data += 1.0 # Modify to make them different print("Are weights tied after fake PEFT?", id(model.get_input_embeddings().weight) == id(model.get_output_embeddings().weight)) model.save_pretrained("./test_tied_model") # Reload model_reloaded = AutoModelForCausalLM.from_pretrained("./test_tied_model") print("Reloaded tie config:", model_reloaded.config.tie_word_embeddings) print("Are weights tied after reload?", id(model_reloaded.get_input_embeddings().weight) == id(model_reloaded.get_output_embeddings().weight)) # Check if the output embeddings weight is equal to input embeddings print("Is the modified output weight preserved?", not torch.allclose(model_reloaded.get_output_embeddings().weight, model_reloaded.get_input_embeddings().weight)) if __name__ == "__main__": main() ``` </details> ## Who can review? @BenjaminBossan @githubnemo <!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @ If you know how to use git blame, that is the easiest way, otherwise, here is a rough guide of **who to tag**. Please tag fewer than 3 people. Models: - text models: @ArthurZucker @Cyrilvallez - vision models: @yonigozlan @molbap - audio models: @eustlb @ebezzam @vasqu - multimodal models: @zucchini-nlp - graph models: @clefourrier Library: - generate: @zucchini-nlp (visual-language models) or @gante (all others) - continuous batching: @remi-or @ArthurZucker @McPatate - pipelines: @Rocketknight1 - tokenizers: @ArthurZucker and @itazap - trainer: @SunMarc - attention: @vasqu @ArthurZucker @CyrilVallez - model loading (from pretrained, etc): @CyrilVallez - distributed: @3outeille @ArthurZucker - CIs: @ydshieh Integrations: - ray/raytune: @richardliaw, @amogkam - Big Model Inference: @SunMarc - quantization: @SunMarc - kernels: @drbh - peft: @BenjaminBossan @githubnemo Devices/Backends: - AMD ROCm: @ivarflakstad - Intel XPU: @IlyasMoutawwakil - Ascend NPU: @ivarflakstad Documentation: @stevhliu Research projects are not maintained and should be taken as is. -->
458cc1339ca623d962c236c276151ca0c46b17b5
[ { "filename": "docs/source/en/_toctree.yml", "patch": "@@ -89,6 +89,8 @@\n title: Experts backends\n - local: continuous_batching\n title: Continuous batching\n+ - local: continuous_batching_architecture\n+ title: Continuous batching architecture\n - local: paged_attention\n ...
rust-lang/rust
154,739
ci: update upload-artifact action to v7
Closes https://github.com/rust-lang/rust/issues/154738
8798ba69439d71c4a84635d5dc891e748dec9e4f
[ { "filename": ".github/workflows/ci.yml", "patch": "@@ -254,7 +254,7 @@ jobs:\n df -h\n \n - name: upload artifacts to github\n- uses: actions/upload-artifact@v4\n+ uses: actions/upload-artifact@v7\n with:\n # name is set in previous step\n name: ${{...
vercel/next.js
92,202
Update .gitignore
<!-- Thanks for opening a PR! Your contribution is much appreciated. To make sure your PR is handled as smoothly as possible we request that you follow the checklist sections below. Choose the right checklist for the change(s) that you're making: ## For Contributors ### Improving Documentation - Run `pnpm prettier-fix` to fix formatting issues before opening the PR. - Read the Docs Contribution Guide to ensure your contribution follows the docs guidelines: https://nextjs.org/docs/community/contribution-guide ### Fixing a bug - Related issues linked using `fixes #number` - Tests added. See: https://github.com/vercel/next.js/blob/canary/contributing/core/testing.md#writing-tests-for-nextjs - Errors have a helpful link attached, see https://github.com/vercel/next.js/blob/canary/contributing.md ### Adding a feature - Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR. (A discussion must be opened, see https://github.com/vercel/next.js/discussions/new?category=ideas) - Related issues/discussions are linked using `fixes #number` - e2e tests added (https://github.com/vercel/next.js/blob/canary/contributing/core/testing.md#writing-tests-for-nextjs) - Documentation added - Telemetry added. In case of a feature if it's used or not. - Errors have a helpful link attached, see https://github.com/vercel/next.js/blob/canary/contributing.md ## For Maintainers - Minimal description (aim for explaining to someone not on the team to understand the PR) - When linking to a Slack thread, you might want to share details of the conclusion - Link both the Linear (Fixes NEXT-xxx) and the GitHub issues - Add review comments if necessary to explain to the reviewer the logic behind a change ### What? ### Why? ### How? Closes NEXT- Fixes # -->
8a60957224f18cc15ca2d173f6730430b45ad129
[ { "filename": ".gitignore", "patch": "@@ -1,3 +1,4 @@\n+#test-harsh-ht\n # build output\n dist\n !.github/actions/*/dist", "additions": 1, "deletions": 0 } ]
ollama/ollama
15,098
parsers: qwen3.5 streaming tool-call parsing and add regression test
null
ae874abfd03be7829d418312db9a035fea3799f8
[ { "filename": "model/parsers/qwen35_test.go", "patch": "@@ -220,6 +220,71 @@ SF\n \t}\n }\n \n+func TestQwen35ParserFakeoutPartialToolCallThenThinkCloseAcrossChunks(t *testing.T) {\n+\tparser := ParserForName(\"qwen3.5\")\n+\tif parser == nil {\n+\t\tt.Fatal(\"expected qwen3.5 parser\")\n+\t}\n+\n+\ttools :...
golang/go
77,749
cmd/go: document -p flag in go help testflag
The -p flag controls the number of packages that can be tested in parallel by `go test`. Although the flag is documented in `go help build`, it is not explicitly documented in `go help testflag`. This change adds documentation for -p to improve discoverability and consistency with other test flags. Fixes #77746.
72c53c51ae3c631433ce81a0488ddfa6e8712399
[ { "filename": "src/cmd/go/internal/test/test.go", "patch": "@@ -310,6 +310,11 @@ control the execution of any test:\n \t in parallel as well, according to the setting of the -p flag\n \t (see 'go help build').\n \n+\t-p n\n+\t\tThe number of packages that can be run in parallel.\n+\t\tThe default is G...
electron/electron
50,640
fix: extension service workers not starting beyond first app launch
Backport of #50611 See that PR for details. Notes: Fixed certain DevTools extension panels not showing without a page reload.
0b954156b2a2ed21f746967562c689b503c711d8
[ { "filename": "shell/browser/extensions/electron_extension_loader.cc", "patch": "@@ -28,6 +28,7 @@\n #include \"extensions/common/error_utils.h\"\n #include \"extensions/common/file_util.h\"\n #include \"extensions/common/manifest_constants.h\"\n+#include \"extensions/common/manifest_handlers/background_inf...
huggingface/transformers
45,132
Fix: Remove double softmax in MoE router load-balancing loss (Mixtral, Qwen2MoE, Qwen3VLMoE)
## Summary This PR fixes GitHub issue #45120: "Double softmax in MoE router load-balancing loss". MoE routers in Mixtral, Qwen2MoE, and Qwen3VLMoE were applying softmax inside forward(), then the load_balancing_loss_func applied softmax AGAIN, resulting in softmax(softmax(logits)) which flattened routing probabilities. ## Root Cause Three routers reassigned `router_logits` with softmaxed values, then returned them to load_balancing_loss_func which expected raw logits. ## Solution Separated concepts: keep raw logits in `router_logits`, use `router_probs` for softmaxed values during routing. ## Changes - MixtralTopKRouter: Renamed softmax reassignment to router_probs - Qwen2MoeTopKRouter: Renamed softmax reassignment to router_probs - Qwen3VLMoeTextTopKRouter: Renamed softmax reassignment to router_probs - Added comprehensive unit tests verifying router_logits are raw logits ## Impact - Fine-tuning: Load-balancing loss now receives correct raw logits - Inference: No changes - Backward compatible: Only internal computation affected Fixes #45120
6ed3ada4236e70631ad2d309dac3c2b7c93ae696
[ { "filename": "tests/models/mixtral/test_modeling_mixtral.py", "patch": "@@ -203,16 +203,6 @@ def test_small_model_aux_loss(self):\n \n # This is to mimic torch.testing.assert_not_close\n self.assertNotAlmostEqual(include_padding_result.aux_loss.item(), result.aux_loss.item())\n- # fm...
vercel/next.js
92,199
[test] Skip flaky `cached-navigations` tests
Skips all [`@test.source.file:test/e2e/app-dir/segment-cache/cached-navigations/cached-navigations.test.ts @git.branch:canary -@test.status:skip @ci.pipeline.url:*/attempts/1 @test.is_known_flaky:true `](https://app.datadoghq.com/ci/test/runs?query=test_level%3Atest%20%40test.source.file%3Atest%2Fe2e%2Fapp-dir%2Fsegment-cache%2Fcached-navigations%2Fcached-navigations.test.ts%20%40git.branch%3Acanary%20-%40test.status%3Askip%20%40ci.pipeline.url%3A%2A%2Fattempts%2F1%20%40test.is_known_flaky%3Atrue&agg_m=count&agg_m_source=base&agg_q=%40test.name&agg_q_source=base&agg_t=count&citest_explorer_sort=timestamp%2Cdesc&cols=%40test.status%2Ctimestamp%2C%40duration%2C%40test.type%2C%40test.name%3A619%2C%40test_session.name&currentTab=overview&eventStack=&fromUser=false&index=citest&top_n=15&top_o=top&viz=toplist&x_missing=true&start=1774438780635&end=1775043580635&paused=true)
39d95b131a6139323f46e340119bb5dc3ae3cc73
[ { "filename": "test/e2e/app-dir/segment-cache/cached-navigations/cached-navigations.test.ts", "patch": "@@ -13,7 +13,8 @@ describe('cached navigations', () => {\n return\n }\n \n- it('serves cached static segments instantly on the second navigation', async () => {\n+ // TODO: flaky\n+ it.skip('serv...
rust-lang/rust
154,736
Add a regression test for the duplicated `crate` keyword in path suggestions
The issue has (long) been fixed, but needs a test. Closes rust-lang/rust#115858.
e2461cf2add278e00aa30bb0c5935906a5da638b
[ { "filename": "tests/ui/resolve/path-suggestion-duplicated-crate-keyword.rs", "patch": "@@ -0,0 +1,20 @@\n+//! Regression test for <https://github.com/rust-lang/rust/issues/115858>.\n+//!\n+//! The compiler used to suggest `crate::crate::unix::linux::system::Y`\n+//! (duplicating the `crate` keyword) instea...
ollama/ollama
15,093
ci: harden cuda include path handling
On windows we can get multiple include dirs, so find where the headers are then copy from that location.
07625ea633be5fc959d90a5bb01d562339f70da1
[ { "filename": "CMakeLists.txt", "patch": "@@ -284,41 +284,54 @@ if(MLX_ENGINE)\n # The Go mlxrunner sets CUDA_PATH to OLLAMA_INSTALL_DIR so MLX finds them at\n # $CUDA_PATH/include/*.h via NVRTC --include-path.\n if(CUDAToolkit_FOUND)\n- set(_cuda_inc \"${CUDAToolkit_INCLUDE_DIRS}\")\n- ...
facebook/react
36,043
Bump flatted from 3.2.7 to 3.4.1
Bumps [flatted](https://github.com/WebReflection/flatted) from 3.2.7 to 3.4.1. <details> <summary>Commits</summary> <ul> <li><a href="https://github.com/WebReflection/flatted/commit/2a02dce7c641dec31194c67663f9b0b12e62da20"><code>2a02dce</code></a> 3.4.1</li> <li><a href="https://github.com/WebReflection/flatted/commit/fba4e8f2e113665da275b19cd0f695f3d98e9416"><code>fba4e8f</code></a> Merge pull request <a href="https://redirect.github.com/WebReflection/flatted/issues/89">#89</a> from WebReflection/python-fix</li> <li><a href="https://github.com/WebReflection/flatted/commit/5fe86485e6df7f7f34a07a2a85498bd3e17384e7"><code>5fe8648</code></a> added &quot;when in Rome&quot; also a test for PHP</li> <li><a href="https://github.com/WebReflection/flatted/commit/53517adbefe724fe472b2f9ebcdb01910d0ae3f0"><code>53517ad</code></a> some minor improvement</li> <li><a href="https://github.com/WebReflection/flatted/commit/b3e2a0c387bf446435fec45ad7f05299f012346f"><code>b3e2a0c</code></a> Fixing recursion issue in Python too</li> <li><a href="https://github.com/WebReflection/flatted/commit/c4b46dbcbf782326e54ea1b65d3ebb1dc7a23fad"><code>c4b46db</code></a> Add SECURITY.md for security policy and reporting</li> <li><a href="https://github.com/WebReflection/flatted/commit/f86d071e0f70de5a7d8200198824a3f07fc9c988"><code>f86d071</code></a> Create dependabot.yml for version updates</li> <li><a href="https://github.com/WebReflection/flatted/commit/d3418c718160eae69dbc0405dce75f7849019e1e"><code>d3418c7</code></a> 3.4.0</li> <li><a href="https://github.com/WebReflection/flatted/commit/7eb65d857e1a40de11c47461cdbc8541449f0606"><code>7eb65d8</code></a> Merge pull request <a href="https://redirect.github.com/WebReflection/flatted/issues/88">#88</a> from WebReflection/avoid-recusrion</li> <li><a href="https://github.com/WebReflection/flatted/commit/7774aae45d3775c842abe9d071fd009171a5fc0c"><code>7774aae</code></a> Avoid recursion on parse due possible shenanigans</li> <li>Additional commits viewable in <a href="https://github.com/WebReflection/flatted/compare/v3.2.7...v3.4.1">compare view</a></li> </ul> </details> <br /> [![Dependabot compatibility score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=flatted&package-manager=npm_and_yarn&previous-version=3.2.7&new-version=3.4.1)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores) Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting `@dependabot rebase`. [//]: # (dependabot-automerge-start) [//]: # (dependabot-automerge-end) --- <details> <summary>Dependabot commands and options</summary> <br /> You can trigger Dependabot actions by commenting on this PR: - `@dependabot rebase` will rebase this PR - `@dependabot recreate` will recreate this PR, overwriting any edits that have been made to it - `@dependabot show <dependency name> ignore conditions` will show all of the ignore conditions of the specified dependency - `@dependabot ignore this major version` will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself) - `@dependabot ignore this minor version` will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself) - `@dependabot ignore this dependency` will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself) You can disable automated security fix PRs for this repo from the [Security Alerts page](https://github.com/facebook/react/network/alerts). </details>
6f06f2bf3403e194b0fb781e225f57d1ac211b28
[ { "filename": "yarn.lock", "patch": "@@ -9229,15 +9229,10 @@ flat-cache@^4.0.0:\n flatted \"^3.2.9\"\n keyv \"^4.5.4\"\n \n-flatted@^3.1.0:\n- version \"3.2.7\"\n- resolved \"https://registry.yarnpkg.com/flatted/-/flatted-3.2.7.tgz#609f39207cb614b89d0765b477cb2d437fbf9787\"\n- integrity sha512-5n...
huggingface/transformers
45,129
fix(config): annotate PreTrainedConfig.dtype as Any to fix pydantic schema generation (#45070)
## Problem Fixes #45070. `PreTrainedConfig.dtype` was annotated as `Union[str, "torch.dtype"] | None`. Since `torch` is only imported under `TYPE_CHECKING`, pydantic's schema builder encounters the `"torch.dtype"` forward reference at runtime and fails with: ``` PydanticUndefinedAnnotation: name 'torch' is not defined ``` This breaks any user code that includes `PreTrainedConfig` as a field in a pydantic model (e.g. vLLM speculator configs, config validators). The repro from the issue: ```python from pydantic import BaseModel, ConfigDict, Field from transformers import PretrainedConfig class MyModelConfig(BaseModel): model_config = ConfigDict(arbitrary_types_allowed=True) sub_config: PretrainedConfig = Field(...) MyModelConfig.model_rebuild(force=True) # PydanticUndefinedAnnotation ``` ## Fix Change `dtype: Union[str, "torch.dtype"] | None = None` to `dtype: Any = None` as suggested by @zucchini-nlp in the issue. `Any` is semantically correct for pydantic's purposes — the field accepts string or `torch.dtype` values at runtime. The forward reference is now only needed for static type checkers (which still see the correct type through `TYPE_CHECKING`-gated imports in user code). The runtime behaviour is unchanged. ## Test Plan - [ ] `MyModelConfig.model_rebuild(force=True)` no longer raises `PydanticUndefinedAnnotation` - [ ] `PreTrainedConfig` can be used as a pydantic field with `arbitrary_types_allowed=True` - [ ] Existing dtype handling in `__post_init__` still works (assigns `torch.dtype` objects at runtime) - [ ] Existing model loading tests pass 🤖 Generated with [Claude Code](https://claude.com/claude-code)
e7dfc6f782b516453b5e9a978292409b93d5d777
[ { "filename": "src/transformers/configuration_utils.py", "patch": "@@ -224,7 +224,11 @@ class PreTrainedConfig(PushToHubMixin, RotaryEmbeddingConfigMixin):\n # Common attributes for all models\n output_hidden_states: bool | None = False\n return_dict: bool | None = True\n- dtype: Union[str, \...
golang/go
77,745
doc/go1.26: document ServeMux trailing slash redirect status change
This CL documents a behavior change in net/http ServeMux introduced in Go 1.26. In Go 1.25, trailing slash redirects returned HTTP 301. In Go 1.26, they now return HTTP 307. This change is not currently mentioned in the Go 1.26 release notes under net/http. Verified behavior: go version go1.25.7 windows/amd64 → trailing slash redirect returns 301 go version go1.26.0 windows/amd64 → trailing slash redirect returns 307 Fixes #77741
f4a642f14c42416974105d14241143dcbff7f789
[ { "filename": "src/cmd/vendor/golang.org/x/tools/go/analysis/passes/modernize/doc.go", "patch": "@@ -91,8 +91,6 @@ The any analyzer suggests replacing uses of the empty interface type,\n `interface{}`, with the `any` alias, which was introduced in Go 1.18.\n This is a purely stylistic change that makes code...
vercel/next.js
92,198
[test] Skip flaky `prefetch-layout-sharing` suite
1 out of 9 tests fails: [`@test.source.file:test/e2e/app-dir/segment-cache/prefetch-layout-sharing/prefetch-layout-sharing.test.ts @git.branch:canary -@test_session.name:*development* -@test.status:skip @ci.pipeline.url:*/attempts/1 `](https://app.datadoghq.com/ci/test/runs?query=test_level%3Atest%20%40test.source.file%3Atest%2Fe2e%2Fapp-dir%2Fsegment-cache%2Fprefetch-layout-sharing%2Fprefetch-layout-sharing.test.ts%20%40git.branch%3Acanary%20-%40test_session.name%3A%2Adevelopment%2A%20-%40test.status%3Askip%20%40ci.pipeline.url%3A%2A%2Fattempts%2F1&agg_m=count&agg_m_source=base&agg_q=%40test.name&agg_q_source=base&agg_t=count&citest_explorer_sort=timestamp%2Cdesc&cols=%40test.status%2Ctimestamp%2C%40duration%2C%40test.type%2C%40test.name%3A619%2C%40test_session.name&currentTab=overview&eventStack=&fromUser=true&index=citest&top_n=10&top_o=top&viz=stream&x_missing=true&start=1774438780635&end=1775043580635&paused=true)
4cc947a18a789489654a6479f60dead272781a02
[ { "filename": "test/e2e/app-dir/segment-cache/prefetch-layout-sharing/prefetch-layout-sharing.test.ts", "patch": "@@ -3,7 +3,8 @@ import { Playwright as NextBrowser } from '../../../../lib/next-webdriver'\n import type * as Playwright from 'playwright'\n import { createRouterAct } from 'router-act'\n \n-des...
electron/electron
50,639
fix: extension service workers not starting beyond first app launch
Backport of #50611 See that PR for details. Notes: Fixed certain DevTools extension panels not showing without a page reload.
0e6ff205fd6fa9ca2c07083b44082d82267da5a9
[ { "filename": "shell/browser/extensions/electron_extension_loader.cc", "patch": "@@ -28,6 +28,7 @@\n #include \"extensions/common/error_utils.h\"\n #include \"extensions/common/file_util.h\"\n #include \"extensions/common/manifest_constants.h\"\n+#include \"extensions/common/manifest_handlers/background_inf...
rust-lang/rust
154,735
Use interior mutability for `StableHashingContext::caching_source_map`.
This lets a lot of `&mut hcx` parameters change to `&hcx`, for symmetry with `to_hash_stable_key`, and avoids some unnecessary cloning of `StableHashingContext`. Details in individual commits.
79b99e682951737086a811a8a725c520c2ccceaa
[ { "filename": "compiler/rustc_middle/src/ich/hcx.rs", "patch": "@@ -11,7 +11,6 @@ use rustc_span::{BytePos, CachingSourceMapView, DUMMY_SP, Pos, SourceFile, Span,\n \n // Very often, we are hashing something that does not need the `CachingSourceMapView`, so we\n // initialize it lazily.\n-#[derive(Clone)]\n...
ollama/ollama
15,083
ci: include mlx jit headers on linux
null
17882660010331a615abc38855985f24fc461abf
[ { "filename": "CMakeLists.txt", "patch": "@@ -246,13 +246,21 @@ if(MLX_ENGINE)\n COMPONENT MLX)\n endif()\n \n- # Install CCCL headers for NVRTC JIT compilation at runtime.\n+ # Install headers for NVRTC JIT compilation at runtime.\n # MLX's own install rules use the default compon...
huggingface/transformers
45,124
[Qwen3.5 MoE] Add _tp_plan to ForConditionalGeneration
# What does this PR do? Adds `_tp_plan = {"lm_head": "colwise_gather_output"}` to `Qwen3_5MoeForConditionalGeneration` (the VL wrapper class). The text-only `Qwen3_5MoeForCausalLM` already had `_tp_plan`, but the VL variant was missing it. This meant that when using `tp_plan="auto"`, the `lm_head` on the VL model was not sharded — each GPU held a full copy and the all-gather behavior (`colwise_gather_output`) was not applied, which could produce incorrect logits under tensor parallelism. **Change:** Applied in `modular_qwen3_5_moe.py` (source of truth) and regenerated `modeling_qwen3_5_moe.py`. **Already in place (no changes needed):** - `base_model_tp_plan` on `Qwen3_5MoeTextConfig` covers full attention (q/k/v/o_proj, q/k_norm), MoE experts, and shared experts. **Out of scope (future work):** - Linear attention (GatedDeltaNet) TP — blocked on `causal_conv1d` DTensor support. - Vision block TP — pending path resolution investigation. Fixes #45125 ## Code Agent Policy - [x] I confirm that this is not a pure code agent PR. ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#create-a-pull-request), Pull Request section? - [x] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? @3outeille @ArthurZucker (distributed / model loading)
c26c592264962ed54f9d525661f3f55d85c61425
[ { "filename": "src/transformers/models/qwen3_5_moe/modeling_qwen3_5_moe.py", "patch": "@@ -2088,6 +2088,7 @@ class Qwen3_5MoeForConditionalGeneration(Qwen3_5MoePreTrainedModel, GenerationMi\n # Reference: fix gemma3 grad acc #37208\n accepts_loss_kwargs = False\n config: Qwen3_5MoeConfig\n+ _...
facebook/react
36,041
Bump undici from 6.23.0 to 6.24.0 in /compiler
Bumps [undici](https://github.com/nodejs/undici) from 6.23.0 to 6.24.0. <details> <summary>Release notes</summary> <p><em>Sourced from <a href="https://github.com/nodejs/undici/releases">undici's releases</a>.</em></p> <blockquote> <h2>v6.24.0</h2> <h1>Undici v6.24.0 Security Release Notes (LTS)</h1> <p>This release backports fixes for security vulnerabilities affecting the v6 line.</p> <h2>Upgrade guidance</h2> <p>All users on v6 should upgrade to <strong>v6.24.0</strong> or later.</p> <h2>Fixed advisories</h2> <ul> <li> <p><a href="https://github.com/nodejs/undici/security/advisories/GHSA-2mjp-6q6p-2qxm">GHSA-2mjp-6q6p-2qxm</a> / CVE-2026-1525 (Medium)<br /> Inconsistent interpretation of HTTP requests (request/response smuggling class issue).</p> </li> <li> <p><a href="https://github.com/nodejs/undici/security/advisories/GHSA-f269-vfmq-vjvj">GHSA-f269-vfmq-vjvj</a> / CVE-2026-1528 (High)<br /> Malicious WebSocket 64-bit frame length handling could crash the client.</p> </li> <li> <p><a href="https://github.com/nodejs/undici/security/advisories/GHSA-4992-7rv2-5pvq">GHSA-4992-7rv2-5pvq</a> / CVE-2026-1527 (Medium)<br /> CRLF injection via the <code>upgrade</code> option.</p> </li> <li> <p><a href="https://github.com/nodejs/undici/security/advisories/GHSA-v9p9-hfj2-hcw8">GHSA-v9p9-hfj2-hcw8</a> / CVE-2026-2229 (High)<br /> Unhandled exception from invalid <code>server_max_window_bits</code> in WebSocket permessage-deflate negotiation.</p> </li> <li> <p><a href="https://github.com/nodejs/undici/security/advisories/GHSA-vrm6-8vpv-qv8q">GHSA-vrm6-8vpv-qv8q</a> / CVE-2026-1526 (High)<br /> Unbounded memory consumption in WebSocket permessage-deflate decompression.</p> </li> </ul> <h2>Not applicable to v6</h2> <ul> <li><a href="https://github.com/nodejs/undici/security/advisories/GHSA-phc3-fgpg-7m6h">GHSA-phc3-fgpg-7m6h</a> / CVE-2026-2581 affects <code>&gt;= 7.17.0 &lt; 7.24.0</code> only.</li> </ul> <h2>Affected and patched ranges (v6)</h2> <ul> <li>CVE-2026-1525: affected <code>&lt; 6.24.0</code>, patched <code>6.24.0</code></li> <li>CVE-2026-1528: affected <code>&gt;= 6.0.0 &lt; 6.24.0</code>, patched <code>6.24.0</code></li> <li>CVE-2026-1527: affected <code>&lt; 6.24.0</code>, patched <code>6.24.0</code></li> <li>CVE-2026-2229: affected <code>&lt; 6.24.0</code>, patched <code>6.24.0</code></li> <li>CVE-2026-1526: affected <code>&lt; 6.24.0</code>, patched <code>6.24.0</code></li> </ul> <h2>References</h2> <ul> <li>GitHub Security Advisories: <a href="https://github.com/nodejs/undici/security/advisories">https://github.com/nodejs/undici/security/advisories</a></li> <li>NVD CVE-2026-1525: <a href="https://nvd.nist.gov/vuln/detail/CVE-2026-1525">https://nvd.nist.gov/vuln/detail/CVE-2026-1525</a></li> <li>NVD CVE-2026-1528: <a href="https://nvd.nist.gov/vuln/detail/CVE-2026-1528">https://nvd.nist.gov/vuln/detail/CVE-2026-1528</a></li> <li>NVD CVE-2026-1527: <a href="https://nvd.nist.gov/vuln/detail/CVE-2026-1527">https://nvd.nist.gov/vuln/detail/CVE-2026-1527</a></li> <li>NVD CVE-2026-2229: <a href="https://nvd.nist.gov/vuln/detail/CVE-2026-2229">https://nvd.nist.gov/vuln/detail/CVE-2026-2229</a></li> <li>NVD CVE-2026-1526: <a href="https://nvd.nist.gov/vuln/detail/CVE-2026-1526">https://nvd.nist.gov/vuln/detail/CVE-2026-1526</a></li> </ul> </blockquote> </details> <details> <summary>Commits</summary> <ul> <li><a href="https://github.com/nodejs/undici/commit/8873c947271faf1ebc455bdc6158ecbc022ecfa9"><code>8873c94</code></a> Bumped v6.24.0</li> <li><a href="https://github.com/nodejs/undici/commit/411bd01a42e7917009bbf686f7628b99d67bbce9"><code>411bd01</code></a> test(websocket): use node:assert for Node 18 compatibility</li> <li><a href="https://github.com/nodejs/undici/commit/844bf59699d778944f78a24ae819c0e8f295766e"><code>844bf59</code></a> test: fix http2 lint regressions in backport</li> <li><a href="https://github.com/nodejs/undici/commit/a444e4f13e8958b4e1ac42bc0d53ace7fba0a9c1"><code>a444e4f</code></a> test: stabilize h2 and tls-cert-leak under current test runner</li> <li><a href="https://github.com/nodejs/undici/commit/dc032a1050d5489b8ce9b4c22aafba98a942f87b"><code>dc032a1</code></a> fix: h2 CI (<a href="https://redirect.github.com/nodejs/undici/issues/4395">#4395</a>)</li> <li><a href="https://github.com/nodejs/undici/commit/4cd3f4b3a2ef910ba728c47ae78294d956410450"><code>4cd3f4b</code></a> test: increase bitness in <code>test/fixtures/*.pem</code> (<a href="https://redirect.github.com/nodejs/undici/issues/3659">#3659</a>)</li> <li><a href="https://github.com/nodejs/undici/commit/7df6442194b7a54e9ac734335e6e0a56a9bc6666"><code>7df6442</code></a> fix: adapt websocket frame-limit handling for v6 parser</li> <li><a href="https://github.com/nodejs/undici/commit/4e0179ae643e6f4380f24cc3683c1b1ca2afb094"><code>4e0179a</code></a> fix: reject duplicate content-length and host headers</li> <li><a href="https://github.com/nodejs/undici/commit/5a97f0893b53ba7d1d5549d3df7e55d9c2673f89"><code>5a97f08</code></a> Fix websocket 64-bit length overflow</li> <li><a href="https://github.com/nodejs/undici/commit/e43e898603dd5e0c14a75b08b83257598d664a39"><code>e43e898</code></a> fix: validate upgrade header to prevent CRLF injection</li> <li>Additional commits viewable in <a href="https://github.com/nodejs/undici/compare/v6.23.0...v6.24.0">compare view</a></li> </ul> </details> <br /> [![Dependabot compatibility score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=undici&package-manager=npm_and_yarn&previous-version=6.23.0&new-version=6.24.0)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores) Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting `@dependabot rebase`. [//]: # (dependabot-automerge-start) [//]: # (dependabot-automerge-end) --- <details> <summary>Dependabot commands and options</summary> <br /> You can trigger Dependabot actions by commenting on this PR: - `@dependabot rebase` will rebase this PR - `@dependabot recreate` will recreate this PR, overwriting any edits that have been made to it - `@dependabot show <dependency name> ignore conditions` will show all of the ignore conditions of the specified dependency - `@dependabot ignore this major version` will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself) - `@dependabot ignore this minor version` will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself) - `@dependabot ignore this dependency` will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself) You can disable automated security fix PRs for this repo from the [Security Alerts page](https://github.com/facebook/react/network/alerts). </details>
6b6d43eaa67710f24ec257a84ff8b58342e42770
[ { "filename": "compiler/yarn.lock", "patch": "@@ -11080,9 +11080,9 @@ undici-types@~6.19.2:\n integrity sha512-ve2KP6f/JnbPBFyobGHuerC9g1FYGn/F8n1LWTwNxCEzd6IfqTwUQcNXgEtmmQ6DlRrC1hrSrBnCZPokRrDHjw==\n \n undici@^6.19.5:\n- version \"6.23.0\"\n- resolved \"https://registry.yarnpkg.com/undici/-/undici-6....
golang/go
77,714
crypto/tls: support RA-TLS challenge extension in ClientHello
### ⚠️ Not ready This PR should only be considered once the real IANA value is assigned for the RA-TLS extension. See the TODO comment on `extensionRATLS` in `src/crypto/tls/common.go` I open it so that early adopters of RA-TLS can look at / use the fork in the mean time. ### Summary This adds server-side support for the Remote Attestation TLS (RA-TLS) challenge extension, as defined in the [RFC-9334](https://www.rfc-editor.org/rfc/rfc9334.html). When a TLS client includes the RA-TLS extension in its ClientHello, the server can read the challenge bytes and use them to generate a platform-specific attestation evidence (e.g. an Intel TDX) that is bound to the TLS handshake. ### Changes | File | Change | |------|--------| | `src/crypto/tls/common.go` | Added `extensionRATLS` constant (`0xffbb`) and `RATLSChallenge []byte` field to `ClientHelloInfo` | | `src/crypto/tls/handshake_messages.go` | Added `raTLSChallenge` field to `clientHelloMsg` and unmarshal logic with 8–64 byte length validation | | `src/crypto/tls/handshake_server.go` | Propagated `raTLSChallenge` from `clientHelloMsg` into `ClientHelloInfo` | | `src/crypto/tls/handshake_messages_test.go` | Added `TestRATLSChallengeExtension` (8 subtests) | ### Design decisions - **Server-side only (unmarshal, no marshal):** The Go TLS client does not need to send this extension; it is consumed by the server to produce attestation evidence via `GetConfigForClient` or `GetCertificate` callbacks. - **Length validation:** The challenge must be 8–64 bytes. Empty or out-of-range payloads are rejected during unmarshal (handshake fails). - **Extension code `0xffbb`:** Temporary private-use value. Will be replaced with the IANA-assigned code once the draft is finalized. - **Optional:** When the extension is absent, `ClientHelloInfo.RATLSChallenge` is `nil` and the handshake proceeds normally. ### Test coverage `TestRATLSChallengeExtension` covers: - Valid challenges at boundary sizes (8, 32, 64 bytes) - Rejection of undersized (0, 7 bytes) and oversized (65 bytes) payloads - ClientHello with no RA-TLS extension (field stays nil) - End-to-end propagation into `ClientHelloInfo.RATLSChallenge` ### References - RFC: https://www.rfc-editor.org/rfc/rfc9334.html - Consumer module: https://github.com/Privasys/caddy-ra-tls-module
5c01c40bdda1fa2b644819be3784956aca3d14d5
[ { "filename": ".github/workflows/build.yml", "patch": "@@ -0,0 +1,62 @@\n+name: Build & Test RA-TLS Go Fork\r\n+\r\n+on:\r\n+ push:\r\n+ branches: [ratls]\r\n+ tags: ['privasys-*']\r\n+ pull_request:\r\n+ branches: [ratls]\r\n+\r\n+permissions:\r\n+ contents: write\r\n+\r\n+jobs:\r\n+ build:\r\...
vercel/next.js
92,177
disable bmi2 in qfilter
Fixes #91708 set the legacy_x86_64_support feature on qfilter to support older intel cpus
95eb4d22639417110c736d0d17a35ac7a3e69821
[ { "filename": "turbopack/crates/turbo-persistence/Cargo.toml", "patch": "@@ -26,7 +26,8 @@ memmap2 = \"0.9.5\"\n nohash-hasher = { workspace = true }\n parking_lot = { workspace = true }\n pot = \"3.0.0\"\n-qfilter = { version = \"0.3.0-alpha.2\", features = [\"serde\"] }\n+# See https://github.com/vercel/n...
electron/electron
50,638
build: allow clearing src & cross mnt cache via dispatch
#### Description of Change As in title. Once-a-day seems to work fine but we're running into issues on occasion even in light of that and it'd be nice to do it on demand. #### Release Notes Notes: none
a98f1ff5307700efc19827b5b3463cc80cedd64e
[ { "filename": ".github/workflows/clean-src-cache.yml", "patch": "@@ -7,6 +7,7 @@ name: Clean Source Cache\n on:\n schedule:\n - cron: \"0 0 * * *\"\n+ workflow_dispatch:\n \n permissions: {}\n ", "additions": 1, "deletions": 0 } ]
rust-lang/rust
154,732
fix(std): avoid AT_MINSIGSTKSZ on uclibc targets
Closes https://github.com/rust-lang/rust/issues/154679 `AT_MINSIGSTKSZ` is not defined on uclibc. r? @tgross35
b0c4b6ef461ea4c7c75ac8f7c259247f24b4b5c7
[ { "filename": "library/std/src/sys/pal/unix/stack_overflow.rs", "patch": "@@ -305,7 +305,7 @@ mod imp {\n }\n \n /// Modern kernels on modern hardware can have dynamic signal stack sizes.\n- #[cfg(any(target_os = \"linux\", target_os = \"android\"))]\n+ #[cfg(all(any(target_os = \"linux\", tar...
ollama/ollama
15,082
tui: update chat title
<img width="1006" height="796" alt="CleanShot 2026-03-26 at 15 56 17@2x" src="https://github.com/user-attachments/assets/68d374ad-b35e-485e-989f-81864cbbcac8" />
870a24283616c1173acb7e598520fb0c5bf18539
[ { "filename": "cmd/tui/tui_test.go", "patch": "@@ -56,7 +56,7 @@ func launcherTestState() *launch.LauncherState {\n \n func TestMenuRendersPinnedItemsAndMore(t *testing.T) {\n \tview := newModel(launcherTestState()).View()\n-\tfor _, want := range []string{\"Run a model\", \"Launch Claude Code\", \"Launch C...
huggingface/transformers
45,123
Fix PP test_ocr_queries
The `test_ocr_queries` assertion value was wrong, even at the initial commit! I'm not sure how tests passed at the time but they're failing now in the CI. This PR fixes the target value!
05bc1ef89f0600ab91e6e2655c9ab20b57940f28
[ { "filename": "tests/models/pp_chart2table/test_processing_pp_chart2table.py", "patch": "@@ -40,7 +40,7 @@ def test_ocr_queries(self):\n add_generation_prompt=True,\n )\n inputs = processor(images=image_input, text=inputs, return_tensors=\"pt\")\n- self.assertEqual(inputs[...
facebook/react
36,035
Bump flatted from 3.2.7 to 3.4.1 in /compiler
Bumps [flatted](https://github.com/WebReflection/flatted) from 3.2.7 to 3.4.1. <details> <summary>Commits</summary> <ul> <li><a href="https://github.com/WebReflection/flatted/commit/2a02dce7c641dec31194c67663f9b0b12e62da20"><code>2a02dce</code></a> 3.4.1</li> <li><a href="https://github.com/WebReflection/flatted/commit/fba4e8f2e113665da275b19cd0f695f3d98e9416"><code>fba4e8f</code></a> Merge pull request <a href="https://redirect.github.com/WebReflection/flatted/issues/89">#89</a> from WebReflection/python-fix</li> <li><a href="https://github.com/WebReflection/flatted/commit/5fe86485e6df7f7f34a07a2a85498bd3e17384e7"><code>5fe8648</code></a> added &quot;when in Rome&quot; also a test for PHP</li> <li><a href="https://github.com/WebReflection/flatted/commit/53517adbefe724fe472b2f9ebcdb01910d0ae3f0"><code>53517ad</code></a> some minor improvement</li> <li><a href="https://github.com/WebReflection/flatted/commit/b3e2a0c387bf446435fec45ad7f05299f012346f"><code>b3e2a0c</code></a> Fixing recursion issue in Python too</li> <li><a href="https://github.com/WebReflection/flatted/commit/c4b46dbcbf782326e54ea1b65d3ebb1dc7a23fad"><code>c4b46db</code></a> Add SECURITY.md for security policy and reporting</li> <li><a href="https://github.com/WebReflection/flatted/commit/f86d071e0f70de5a7d8200198824a3f07fc9c988"><code>f86d071</code></a> Create dependabot.yml for version updates</li> <li><a href="https://github.com/WebReflection/flatted/commit/d3418c718160eae69dbc0405dce75f7849019e1e"><code>d3418c7</code></a> 3.4.0</li> <li><a href="https://github.com/WebReflection/flatted/commit/7eb65d857e1a40de11c47461cdbc8541449f0606"><code>7eb65d8</code></a> Merge pull request <a href="https://redirect.github.com/WebReflection/flatted/issues/88">#88</a> from WebReflection/avoid-recusrion</li> <li><a href="https://github.com/WebReflection/flatted/commit/7774aae45d3775c842abe9d071fd009171a5fc0c"><code>7774aae</code></a> Avoid recursion on parse due possible shenanigans</li> <li>Additional commits viewable in <a href="https://github.com/WebReflection/flatted/compare/v3.2.7...v3.4.1">compare view</a></li> </ul> </details> <br /> [![Dependabot compatibility score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=flatted&package-manager=npm_and_yarn&previous-version=3.2.7&new-version=3.4.1)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores) Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting `@dependabot rebase`. [//]: # (dependabot-automerge-start) [//]: # (dependabot-automerge-end) --- <details> <summary>Dependabot commands and options</summary> <br /> You can trigger Dependabot actions by commenting on this PR: - `@dependabot rebase` will rebase this PR - `@dependabot recreate` will recreate this PR, overwriting any edits that have been made to it - `@dependabot show <dependency name> ignore conditions` will show all of the ignore conditions of the specified dependency - `@dependabot ignore this major version` will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself) - `@dependabot ignore this minor version` will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself) - `@dependabot ignore this dependency` will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself) You can disable automated security fix PRs for this repo from the [Security Alerts page](https://github.com/facebook/react/network/alerts). </details>
80560961df021a5515337eb62340567672080a27
[ { "filename": "compiler/yarn.lock", "patch": "@@ -5970,15 +5970,10 @@ flat@^5.0.2:\n resolved \"https://registry.npmjs.org/flat/-/flat-5.0.2.tgz\"\n integrity sha512-b6suED+5/3rTpUBdG1gupIl8MPFCAMA0QXwmljLhvCUKcUvdE4gWky9zpuGCcXHOsz4J9wPGNWq6OKpmIzz3hQ==\n \n-flatted@^3.1.0:\n- version \"3.2.7\"\n- re...
golang/go
77,705
cmd/compile: fix typo in comment (nor->not)
Fix typo: change "nor" to "not" in comment in call.go Fixes #77668
bf2889c967e770032361dd64f0798a5632931cd2
[ { "filename": "src/cmd/compile/internal/types2/call.go", "patch": "@@ -388,7 +388,7 @@ func (check *Checker) genericExprList(elist []syntax.Expr) (resList []*operand,\n \t}\n \n \t// Before Go 1.21, uninstantiated or partially instantiated argument functions are\n-\t// nor permitted. Checker.funcInst must i...
rust-lang/rust
154,731
llvm: Fix array ABI test to not check equality implementation
LLVM has moved memcmp expansion in the pipeline, resulting in the bcmp call being expanded into loads and register comparisons, which breaks the test. See llvm/llvm-project#77370 Based on history, I believe the test actually intended validate that these arrays were being passed as pointer arguments, which can be done more directly. r? @durin42 cc @scottmcm (author of the test being modified) @rustbot label llvm-main
d6be991fb4da0acac3d9fb2938f9ad00db999c4f
[ { "filename": "tests/codegen-llvm/array-equality.rs", "patch": "@@ -1,3 +1,6 @@\n+//@ revisions: llvm-current llvm-next\n+//@[llvm-current] ignore-llvm-version: 23-99\n+//@[llvm-next] min-llvm-version: 23\n //@ compile-flags: -Copt-level=3 -Z merge-functions=disabled\n //@ only-x86_64\n #![crate_type = \"li...
vercel/next.js
92,176
Improve revalidateTag JSDoc to include guidance about required second parameter
### What? Improved the JSDoc comment on `revalidateTag` to explain the required second `profile` parameter, linking to `cacheLife` profile docs and mentioning `updateTag` as an alternative for immediate expiration. ### Why? The TypeScript error "Expected 2 arguments, but got 1" at `revalidateTag` call sites gives no actionable guidance. When users hover over the function in their IDE, the existing JSDoc only says "purge cached data on-demand" — nothing about what the second argument should be or what alternatives exist. ### How? Added a prose paragraph to the JSDoc describing the `profile` parameter (a `cacheLife` profile like `"max"` or a `{ expire }` object) and pointing users to `updateTag` for immediate expiration in Server Actions. Follows the established JSDoc conventions in the file (no `@param`, `@example`, or `@see` tags). ## PR checklist (Fixing a bug) - Errors have a helpful link attached: links to `cacheLife` reference, `updateTag` docs, and `revalidateTag` docs
8481df518bb6730a2b62ea6724c7ccb719530aa6
[ { "filename": "test/rspack-dev-tests-manifest.json", "patch": "@@ -1193,11 +1193,14 @@\n },\n \"test/development/app-dir/server-hmr/server-hmr.test.ts\": {\n \"passed\": [\n+ \"server-hmr CLI/config conflict warning should warn when CLI flag conflicts with config\",\n+ \"server-hmr metadat...
electron/electron
50,635
ci: fix pulling previous object checksums
#### Description of Change <!-- Thank you for your Pull Request. Please provide a description above and review the requirements below. Contributors guide: https://github.com/electron/electron/blob/main/CONTRIBUTING.md NOTE: PRS submitted without this template will be automatically closed. --> Follow-up to #50390. There are two issues that this is fixing: * When `archive: false` is used for `actions/upload-artifact`, the name of the artifact ends up being the filename, not the name supplied in the `name` input, so it needs to be looked up using that name. * The parent commit is not guaranteed to have the artifact we're looking for: * Docs only PRs don't trigger a build and won't upload the artifact * On release branches we limit concurrency so a given commit may not get a full build run if another PR is merged shortly after it Removes `dawidd6/action-download-artifact` which in hindsight I shouldn't have added, and it can't handle the non-archived artifacts. #### Checklist <!-- Remove items that do not apply. For completed items, change [ ] to [x]. --> - [x] PR description included #### Release Notes Notes: none <!-- Please add a one-line description for app developers to read in the release notes, or 'none' if no notes relevant to app developers. Examples and help on special cases: https://github.com/electron/clerk/blob/main/README.md#examples -->
707ae177998cee2d670440cc4894443c4dc61d22
[ { "filename": ".github/actions/build-electron/action.yml", "patch": "@@ -51,25 +51,12 @@ runs:\n shell: bash\n if: ${{ (github.event_name == 'push' || github.event_name == 'pull_request') && inputs.is-asan != 'true' }}\n env:\n- GH_TOKEN: ${{ github.token }}\n+ GITHUB_TOKEN: ...
ollama/ollama
15,080
launch: hide cline integration
Hide Cline integration for now due to changes of default running methods
f33a8ec9a37357e9cc1f5fc106bdb629387d285e
[ { "filename": "cmd/launch/registry.go", "patch": "@@ -33,7 +33,7 @@ type IntegrationInfo struct {\n \tDescription string\n }\n \n-var launcherIntegrationOrder = []string{\"opencode\", \"droid\", \"pi\", \"cline\"}\n+var launcherIntegrationOrder = []string{\"opencode\", \"droid\", \"pi\"}\n \n var integratio...
golang/go
77,702
path/filepath: clarify Dir documentation for Windows
Clarify that on Windows Dir C: returns C:. Fixes #77696
3760e7225e90b121d7305beeb7614588c042d523
[ { "filename": "src/path/filepath/path.go", "patch": "@@ -463,6 +463,8 @@ func Base(path string) string {\n // If the path is empty, Dir returns \".\".\n // If the path consists entirely of separators, Dir returns a single separator.\n // The returned path does not end in a separator unless it is the root di...
vercel/next.js
92,174
[backport] Fix CSS HMR on Safari
## Summary - Backport of #92123 to `next-16-2` release branch - Cherry-picked the CSS HMR Safari fix and regenerated turbopack-tests snapshot goldens for the release branch ## Test plan - [x] `UPDATE=1 cargo nextest r -p turbopack-tests` — all 297 tests pass - [x] `pnpm build-all` — clean build, no issues with `generated-native.d.ts`
83064d31b0c75bdd2ff6b12b0b5ceac6a1bc46be
[ { "filename": "packages/next/src/build/swc/generated-native.d.ts", "patch": "@@ -153,8 +153,8 @@ export declare function minify(\n export declare function minifySync(input: Buffer, opts: Buffer): TransformOutput\n export interface NapiEndpointConfig {}\n export interface NapiAssetPath {\n- path: RcStr\n- ...
electron/electron
50,632
ci: correct contributing link and add link to ai tool policy
#### Description of Change This PR fixes the link to our contributing guidelines and adds the link to our ai tool policy that came out since the label was created. #### Checklist <!-- Remove items that do not apply. For completed items, change [ ] to [x]. --> - [x] PR description included - [x] [PR release notes](https://github.com/electron/clerk/blob/main/README.md) describe the change in a way relevant to app developers, and are [capitalized, punctuated, and past tense](https://github.com/electron/clerk/blob/main/README.md#examples). #### Release Notes Notes: none
a551f5483a47c58ddf1d1543f790058ba097db82
[ { "filename": ".github/workflows/pull-request-labeled.yml", "patch": "@@ -72,7 +72,7 @@ jobs:\n \n Hello @${{ github.event.pull_request.user.login }}. Due to the high amount of AI spam PRs we receive, if a PR is detected to be majority AI-generated without disclosure and untested, we will automati...
rust-lang/rust
154,728
rustdoc: Improve internal function name
This functions name totally contradicted what it did. And the only places it was used immediatly `!`ed it. Push the `!` into the function, and rename it to make sense. Should hopefully result in less head-scratching next time someone looks at this. r? @GuillaumeGomez
e469da4f842d71f1d80155676f9f1913d8a9ccd8
[ { "filename": "src/librustdoc/clean/cfg.rs", "patch": "@@ -41,15 +41,15 @@ fn is_simple_cfg(cfg: &CfgEntry) -> bool {\n }\n }\n \n-/// Returns `false` if is `Any`, otherwise returns `true`.\n-fn is_all_cfg(cfg: &CfgEntry) -> bool {\n+/// Returns `true` if is [`CfgEntry::Any`], otherwise returns `false`....
facebook/react
36,034
Fix: use Object.is for dependency comparison to handle NaN correctly
## Summary Previously, the React Compiler generated code using \!==\ for dependency comparison, which doesn't match React's \Object.is\ semantics. This caused NaN in dependency arrays to always trigger re-evaluation since \NaN !== NaN\ is true, but \Object.is(NaN, NaN)\ is true. ## Problem When a dependency value is NaN, the compiled code would incorrectly detect a change on every render: \\\js // Generated code (before fix): if ($[0] !== nanValue) { // Always true because NaN !== NaN // Re-computes every time } \\\ ## Solution Changed the generated code to use \!Object.is()\ for dependency comparison: \\\js // Generated code (after fix): if (!Object.is($[0], nanValue)) { // Correctly returns false when both are NaN // Only re-computes when value actually changes } \\\ ## Test Plan - Existing compiler tests should pass - This matches React's documented equality semantics - Manual verification with the playground repro from #35854 Fixes #35854
bbad60d0dc35be1026b883395567602fd8feea00
[ { "filename": "compiler/packages/babel-plugin-react-compiler/src/ReactiveScopes/CodegenReactiveFunction.ts", "patch": "@@ -574,14 +574,22 @@ function codegenReactiveScope(\n \n for (const dep of [...scope.dependencies].sort(compareScopeDependency)) {\n const index = cx.nextCacheIndex;\n- const comp...
huggingface/transformers
45,122
:rotating_light: [`LightGlue`] Remove remote code execution
Native model integration should never have anything related to remote code execution: That just means that we would need to add that model natively, not via remote - it can and will not be maintained by us. Additionally added a new linter rule to check for simple usage of this via `trust_remote_code=...` NOTE: The breaking part that we effectively deprecate a subset of the lightglue models (which should not have been added this way); see [lightglue_disk](https://huggingface.co/ETH-CVG/lightglue_disk) and this [hub PR](https://huggingface.co/ETH-CVG/lightglue_disk/discussions/2)
fd7d1f56d289f98345d7000833b270ce06132e23
[ { "filename": "Makefile", "patch": "@@ -33,6 +33,7 @@ check-code-quality:\n check-repository-consistency:\n \t@python utils/checkers.py \\\n \t\timports,\\\n+\t\timport_complexity,\\\n \t\tcopies,\\\n \t\tmodular_conversion,\\\n \t\tdoc_toc,\\\n@@ -58,6 +59,7 @@ check-repo:\n \t\tinit_isort,\\\n \t\tauto_ma...
ollama/ollama
15,076
launch: hide vs code
Hide VS Code integration for now while leaving it accessible through `ollama launch vscode`
8cc7a88e1e9fbccab22a26fa78ff9f73a95f8df9
[ { "filename": "cmd/launch/registry.go", "patch": "@@ -33,7 +33,7 @@ type IntegrationInfo struct {\n \tDescription string\n }\n \n-var launcherIntegrationOrder = []string{\"vscode\", \"opencode\", \"droid\", \"pi\", \"cline\"}\n+var launcherIntegrationOrder = []string{\"opencode\", \"droid\", \"pi\", \"cline...
golang/go
77,700
os: update examples to use errors.Is instead of os.IsExist/IsNotExist
## Summary Updated two example functions in `src/os/example_test.go` to use the modern `errors.Is(err, fs.ErrExist)` and `errors.Is(err, fs.ErrNotExist)` instead of the deprecated `os.IsExist(err)` and `os.IsNotExist(err)`. ## Changes - `ExampleMkdir()`: Changed `!os.IsExist(err)` to `!errors.Is(err, fs.ErrExist)` - `ExampleUserConfigDir()`: Changed `!os.IsNotExist(err)` to `!errors.Is(err, fs.ErrNotExist)` The `errors` and `io/fs` packages were already imported in the file. Fixes #77643
9818fc0402d3e3f988588a986c7ea776b4f52e21
[ { "filename": "src/os/example_test.go", "patch": "@@ -246,7 +246,7 @@ func ExampleWriteFile() {\n \n func ExampleMkdir() {\n \terr := os.Mkdir(\"testdir\", 0750)\n-\tif err != nil && !os.IsExist(err) {\n+\tif err != nil && !errors.Is(err, fs.ErrExist) {\n \t\tlog.Fatal(err)\n \t}\n \terr = os.WriteFile(\"te...
electron/electron
50,630
test: add tests for `navigationHistory.goToIndex()`
Backport of #50607 See that PR for details. Notes: none.
b2cf6aad7f437193dac83b44f7f21bfaef452644
[ { "filename": "spec/api-web-contents-spec.ts", "patch": "@@ -802,6 +802,65 @@ describe('webContents module', () => {\n });\n });\n \n+ describe('navigationHistory.goToIndex(index) API', () => {\n+ beforeEach(async () => {\n+ await w.loadURL(urlPage1);\n+ await w.loadURL(urlPa...
rust-lang/rust
154,727
Rollup of 20 pull requests
Successful merges: - rust-lang/rust#153105 (Compute the result of a projection type with region errors) - rust-lang/rust#153532 (Attributes containing rustc) - rust-lang/rust#153960 (Make `layout_of` cycles fatal errors) - rust-lang/rust#154527 (Emit pre-expansion feature gate warnings for negative impls and specialization) - rust-lang/rust#154666 (Remove `StableHashContext` impls) - rust-lang/rust#154669 (Introduce #[diagnostic::on_move] on `Arc`) - rust-lang/rust#154710 (opaque_generic_const_args -> generic_const_args) - rust-lang/rust#154712 (Revert "`-Znext-solver` Remove the forced ambiguity hack from search graph") - rust-lang/rust#153614 (`FindParamInClause` handle edge-cases) - rust-lang/rust#154213 (tidy-alphabetical: fix line number in error message) - rust-lang/rust#154425 (Migrate transmute tests) - rust-lang/rust#154442 (Export `derive` at the crate root: `core::derive` and `std::derive`) - rust-lang/rust#154469 (mGCA: Lower spans for literal const args) - rust-lang/rust#154578 (Rename `probe_ty_var` to `try_resolve_ty_var`) - rust-lang/rust#154615 (Moving issues) - rust-lang/rust#154644 (rustdoc: seperate methods and associated functions in sidebar) - rust-lang/rust#154660 (Avoid creating async return opaques for foreign async fns) - rust-lang/rust#154671 (Add a test for a past ICE when calling a const fn of an unresolved type with the wrong number of args) - rust-lang/rust#154680 ([rustdoc] Replace `DocContext` with `TyCtxt` wherever possible) - rust-lang/rust#154709 (Revert `Ty` type alias in `rustc_type_ir`) <!-- homu-ignore:start --> r? @ghost [Create a similar rollup](https://bors.rust-lang.org/queue/rust?prs=153105,153532,153960,154527,154666,154669,154710,154712,153614,154213,154425,154442,154469,154578,154615,154644,154660,154671,154680,154709) <!-- homu-ignore:end -->
f3ba2c41e63ac58ae8aafbe7725db6b5e6021d00
[ { "filename": "src/librustdoc/clean/inline.rs", "patch": "@@ -137,7 +137,7 @@ pub(crate) fn try_inline(\n })\n }\n Res::Def(DefKind::Macro(kinds), did) => {\n- let mac = build_macro(cx, did, name, kinds);\n+ let mac = build_macro(cx.tcx, did, name, kinds);\n...
huggingface/transformers
45,121
fix: remove unsafe exec() in serve.py
## Summary Fix critical severity security issue in `src/transformers/cli/serve.py`. ## Vulnerability | Field | Value | |-------|-------| | **ID** | V-007 | | **Severity** | CRITICAL | | **Scanner** | multi_agent_ai | | **Rule** | `V-007` | | **File** | `src/transformers/cli/serve.py:754` | **Description**: The /load_model endpoint allows loading arbitrary models without authentication or validation. Attackers can load malicious pickle files containing arbitrary Python code that executes during deseri... ## Changes - `src/transformers/cli/serve.py` ## Verification - [x] Build passes - [x] Scanner re-scan confirms fix - [x] Code review passed --- *Automated security fix by [OrbisAI Security](https://orbisappsec.com)*
6a6f77ec341c1a698232b5ee1b03ea9d8d32e8d1
[ { "filename": "src/transformers/cli/serve.py", "patch": "@@ -728,14 +728,48 @@ def responses(request: dict):\n async def audio_transcriptions(request: Request):\n # Parses the multipart/form-data request into the request format used by other endpoints\n async with request.for...
ollama/ollama
15,073
launch/vscode: prefer known vs code paths over `code` on PATH
When launching VS Code, findBinary() previously checked for code on PATH first. This could silently launch Cursor or another VS Code fork if their CLI was registered as code. This PR flips the lookup order and adds validation: 1. Check platform-specific paths first — `/Applications/Visual Studio Code.app` on macOS, `Microsoft VS Code\bin\code.cmd` on Windows, `/usr/bin/code` or `/snap/bin/code` on Linux 2. Fall back to `code` on PATH only if it's actually vscode — resolve the symlink and check that the real path points to a vscode installation (e.g. contains `Visual Studio Code` or `Microsoft VS Code`) 3. Handle wrapper scripts — if `code` is a small shell script (not a symlink), read its content and check for vscode references
07baff1ce3590ebfc7e0e2a913ebc90203776aca
[ { "filename": "cmd/launch/vscode.go", "patch": "@@ -25,9 +25,7 @@ type VSCode struct{}\n func (v *VSCode) String() string { return \"Visual Studio Code\" }\n \n // findBinary returns the path/command to launch VS Code, or \"\" if not found.\n-// It checks platform-specific locations first, then falls back t...
golang/go
77,696
path/filepath: clarify Dir and Clean documentation for Windows
Fixes #77314 Updates documentation. The current documentation is misleading because it suggests that "C:" would return "C:\" but it actually returns "C:."
1dd401c925ef52fb618659cc3215325f9b448e4d
[ { "filename": "src/path/filepath/path.go", "patch": "@@ -466,6 +466,7 @@ func Base(path string) string {\n // If the path consists entirely of separators, Dir returns a single separator.\n // The returned path does not end in a separator unless it is the root directory,\n // except for Windows volume paths ...
electron/electron
50,615
fix: don't force kFitToPrintableArea scaling when custom margins are set
#### Description of Change Fixes https://github.com/electron/electron/issues/50481 When silent printing with non-default margins (custom, no margins, or printable area margins), the kFitToPrintableArea scaling option causes double-marginalization: the custom margins define the content area, then the scaling additionally fits content to the printer's printable area. Only apply kFitToPrintableArea when using default margins in silent mode. For non-default margins, use the same scaling as non-silent prints. #### Checklist <!-- Remove items that do not apply. For completed items, change [ ] to [x]. --> - [x] PR description included - [x] I have built and tested this PR - [x] `npm test` passes - [x] [PR release notes](https://github.com/electron/clerk/blob/main/README.md) describe the change in a way relevant to app developers, and are [capitalized, punctuated, and past tense](https://github.com/electron/clerk/blob/main/README.md#examples). #### Release Notes Notes: Fixed an issue where margins did not look as expected when printing in silent mode.
b695415f52150c8084373ca4684749e15f6a9d6e
[ { "filename": "patches/chromium/printing.patch", "patch": "@@ -675,7 +675,7 @@ index ac2f719be566020d9f41364560c12e6d6d0fe3d8..16d758a6936f66148a196761cfb875f6\n PrintingFailed(int32 cookie, PrintFailureReason reason);\n \n diff --git a/components/printing/renderer/print_render_frame_helper.cc b/compone...
facebook/react
36,026
Enable Fragment Ref flags across builds
null
5673298c62799488040bdb48cebadd579d252bd8
[ { "filename": "packages/shared/ReactFeatureFlags.js", "patch": "@@ -145,7 +145,7 @@ export const enableInfiniteRenderLoopDetection: boolean = false;\n \n export const enableFragmentRefs: boolean = true;\n export const enableFragmentRefsScrollIntoView: boolean = true;\n-export const enableFragmentRefsInstanc...
ollama/ollama
15,068
fix: pin 10 unpinned action(s),extract 1 unsafe expression(s) to env vars
## Fix: CI/CD Security Vulnerabilities in GitHub Actions Hi! [Runner Guard](https://github.com/Vigilant-LLC/runner-guard), an open-source CI/CD security scanner by [Vigilant Cyber Security](https://www.vigilantdefense.com), identified security vulnerabilities in this repository's GitHub Actions workflows. This PR applies automated fixes where possible and reports additional findings for your review. ### Fixes applied (in this PR) | Rule | Severity | File | Description | |------|----------|------|-------------| | RGS-007 | high | `.github/workflows/latest.yaml` | Pinned 1 third-party action(s) to commit SHA | | RGS-007 | high | `.github/workflows/release.yaml` | Pinned 8 third-party action(s) to commit SHA | | RGS-007 | high | `.github/workflows/test.yaml` | Pinned 1 third-party action(s) to commit SHA | | RGS-002 | high | `.github/workflows/test.yaml` | Extracted 1 unsafe expression(s) to env vars | ### Advisory: additional findings (manual review recommended) | Rule | Severity | File | Description | | RGS-003 | high | `.github/workflows/release.yaml` | Filename Injection via Git Diff or File Listing | ### Why this matters GitHub Actions workflows that use untrusted input in `run:` blocks, expose secrets inline, or use unpinned third-party actions are vulnerable to code injection, credential theft, and supply chain attacks. These are the same vulnerability classes exploited in the [tj-actions/changed-files incident](https://www.vigilantdefense.com/resources/runner-guard) and subsequent supply chain attacks, which compromised CI secrets across thousands of repositories. ### How to verify Review the diff — each change is mechanical and preserves workflow behavior: - **Expression extraction** (RGS-002/008/014): Moves `${{ }}` expressions from `run:` blocks into `env:` mappings, preventing shell injection - **SHA pinning** (RGS-007): Pins third-party actions to immutable commit SHAs (original version tag preserved as comment) - **Debug env removal** (RGS-015): Removes `ACTIONS_RUNNER_DEBUG`/`ACTIONS_STEP_DEBUG` which leak secrets in workflow logs Run `brew install Vigilant-LLC/tap/runner-guard && runner-guard scan .` or install from the [repo](https://github.com/Vigilant-LLC/runner-guard) to verify. --- Found by [Runner Guard](https://github.com/Vigilant-LLC/runner-guard) | Built by [Vigilant Cyber Security](https://www.vigilantdefense.com) | [Learn more](https://www.vigilantdefense.com/resources/runner-guard) If this PR is not welcome, just close it -- we won't send another.
410f0894e320c2804d979d7ef7cd9bd360e4b261
[ { "filename": ".github/workflows/latest.yaml", "patch": "@@ -11,7 +11,7 @@ jobs:\n steps:\n - uses: actions/checkout@v4\n - name: Login to Docker Hub\n- uses: docker/login-action@v3\n+ uses: docker/login-action@c94ce9fb468520275223c153574b00df6fe4bcc9 # v3\n with:\n ...
huggingface/transformers
45,119
Fix: Preserve PreTrainedConfig __init__ signatures for type checkers (fixes #45071)
## Summary This PR fixes GitHub issue #45071: "v5.4.0 breaks PretrainedConfig type checking". The regression prevents type checkers (mypy, pyright) from validating `PretrainedConfig` subclass instantiation with valid parameters. ## Root Cause PR #44953 ("Config kwargs") introduced `wrap_init_to_accept_kwargs()` to accept extra kwargs for backward compatibility. However, the wrapper replaced the dataclass-generated `__init__` signature with a generic `(*args, **kwargs: Any)` signature. Type checkers rely on function signatures to validate parameters, so they lost information about actual config parameters. ## Solution This PR uses Python's standard `__signature__` attribute technique (documented in PEP 362) to preserve the original `__init__` signature on the wrapped function. This allows: - Type checkers to see and validate the original parameters - Runtime to continue accepting extra kwargs for backward compatibility - Zero breaking changes to existing code This technique is used in production Python libraries including Pydantic, Click, and FastAPI. ## Changes ### Modified: `src/transformers/configuration_utils.py` - Line 18: Added `import inspect` - Line 78: Capture original signature before wrapping - Lines 112-113: Restore signature on wrapped function ### Added: `tests/utils/test_configuration_utils.py` - New test `test_pretrained_config_signature_preserved()` validates signature preservation, parameter validation, backward compatibility, and prevents generic signatures ## Testing All tests pass with real `BertConfig` class - backward compatibility with extra kwargs verified. ## Impact - **Type Checking**: mypy, pyright, and other type checkers can now properly validate `PretrainedConfig` subclass instantiation - **Runtime**: No changes - extra kwargs still accepted for backward compatibility - **Performance**: Negligible - only affects type checker introspection - **Breaking Changes**: None - fully backward compatible Fixes #45071
d802749d1ac8e416b423a6c59a200697c0622d6c
[ { "filename": "src/transformers/configuration_utils.py", "patch": "@@ -15,6 +15,7 @@\n \"\"\"Configuration base class and utilities.\"\"\"\n \n import copy\n+import inspect\n import json\n import math\n import os\n@@ -74,6 +75,7 @@\n # copied from huggingface_hub.dataclasses.strict when `accept_kwargs=True`...
rust-lang/rust
154,726
Rollup of 4 pull requests
Successful merges: - rust-lang/rust#154444 (rustdoc ICE fix: When collecting `Deref` impls with their targets, skip the negative ones) - rust-lang/rust#154590 (Make #[cfg] suggest any or all on #[cfg(a, b)]) - rust-lang/rust#154691 (core: Update the feature gate on `TryFrom<integer> for bool`) - rust-lang/rust#154697 (rustdoc: fix href of extern crates in search results) Failed merges: - rust-lang/rust#154722 (fix(lints): Improve `ill_formed_attribute_input` with better help message) <!-- homu-ignore:start --> r? @ghost [Create a similar rollup](https://bors.rust-lang.org/queue/rust?prs=154444,154590,154691,154697,154722) <!-- homu-ignore:end -->
8c76d2b0de818017d6b3f7a7dd609cb593ec2bed
[ { "filename": "src/librustdoc/html/render/search_index.rs", "patch": "@@ -13,6 +13,7 @@ use ::serde::{Deserialize, Serialize};\n use rustc_ast::join_path_syms;\n use rustc_data_structures::fx::{FxHashMap, FxHashSet, FxIndexMap};\n use rustc_data_structures::thin_vec::ThinVec;\n+use rustc_hir::def_id::LOCAL_...
nodejs/node
62,500
buffer: fix 32-bit truncation in SlowCopy for buffers larger than 4 GiB
## Summary `SlowCopy` in `src/node_buffer.cc` extracted `targetStart`, `sourceStart`, and `to_copy` via `.As<Uint32>()->Value()`, silently truncating values that exceed `2^32 - 1` (4 GiB). This caused `Buffer.prototype.copy()` to copy the wrong number of bytes or write to the wrong position when dealing with buffers larger than 4 GiB. Root cause was introduced in #54087 ("buffer: use native copy impl"), which replaced the original `ParseArrayIndex`-based extraction (using `size_t`) with `Uint32Value` (32-bit). ## Changes - `src/node_buffer.cc`: Use `IntegerValue()` (returns `int64_t`) in `SlowCopy` instead of `.As<Uint32>()->Value()`, and widen `CopyImpl`'s parameter types from `uint32_t` to `size_t`. - `FastCopy` is left unchanged — V8's Fast API guarantees it is only invoked when all arguments fit in `uint32_t`; for larger values V8 falls back to `SlowCopy` automatically. - `test/parallel/test-buffer-copy-large.js`: New test that allocates a buffer slightly larger than 4 GiB and verifies `Buffer.prototype.copy()` copies all bytes correctly. The test skips gracefully on systems with insufficient memory. ## Test ```js // Before fix: only 5 bytes were copied (2^32 + 5 mod 2^32 = 5). // After fix: all FOUR_GB + 5 bytes are copied correctly. const src = Buffer.alloc(2 ** 32 + 5, 0xbb); const dst = Buffer.alloc(src.length, 0xaa); src.copy(dst); assert.strictEqual(dst[dst.length - 1], 0xbb); ``` Fixes #55422
4ce7086b2433ba48ca1f843f334754246e056245
[ { "filename": ".clang-format", "patch": "@@ -0,0 +1,111 @@\n+---\n+Language: Cpp\n+# BasedOnStyle: Google\n+AccessModifierOffset: -1\n+AlignAfterOpenBracket: Align\n+AlignConsecutiveAssignments: false\n+AlignConsecutiveDeclarations: false\n+AlignEscapedNewlines: Right\n+AlignOperands: true\n+Align...
electron/electron
50,614
fix: remove menu update debug log
Backport of #50608 See that PR for details. Notes: Removed "representedObject is not a WeakPtrToElectronMenuModelAsNSObject" logging when interacting with macOS menus.
22fb426acf923e1dd4d080ae795350c6759c0868
[ { "filename": "shell/browser/ui/cocoa/electron_menu_controller.mm", "patch": "@@ -478,7 +478,6 @@ - (void)applyStateToMenuItem:(NSMenuItem*)item {\n \n if (![represented\n isKindOfClass:[WeakPtrToElectronMenuModelAsNSObject class]]) {\n- NSLog(@\"representedObject is not a WeakPtrToElectronMe...
golang/go
77,694
all: test: remove unneeded loop variables
This CL follows from the abandoned CL 722961. Alan Donovan asked if modernize would catch these changes; it does in part. Here is how modernize was used: 1. Build the Go compiler from latest master 2. Clone x/tools and build modernize with the newest compiler 3. Then, do: PATH=PATH_TO_BIN:$PATH go list ./... | xargs env PATH=PATH_TO_BIN:$PATH GOMOD=off MODERNIZE -forvar -fix -test -debug fpstv 4. For assurance, move into a package directory (i.e, cmd), and redo Step 3. From the obtained result, it seems modernize did not remove the loop variable when: - the range notation was not used - the loop variable was not directly under the for directive. Which does happens here: # git ls-files | xargs -I {} perl -nE 'print "$ARGV:$.:$_" if /\s+(\w+) := \1$/' {} | grep _test [...] cmd/compile/internal/types2/api_test.go:2423: i := i md/go/internal/modload/query_test.go:182: tt := tt md/go/internal/web/url_test.go:17: tc := tc cmd/go/internal/web/url_test.go:49: tc := tc cmd/internal/par/queue_test.go:54: i := i runtime/syscall_windows_test.go:781: arglen := arglen Link: https://go-review.googlesource.com/c/go/+/722961/comments/e8d47866_fc399fa1 Co-authored-by: Plamerdi Makela <plamerdi447@gmail.com> Fixes #76411
77c3e11fc21a4ede70b733b567a1690edb944dc1
[ { "filename": "src/cmd/api/api_test.go", "patch": "@@ -232,7 +232,6 @@ var warmupCache = sync.OnceFunc(func() {\n \t// Warm up the import cache in parallel.\n \tvar wg sync.WaitGroup\n \tfor _, context := range contexts {\n-\t\tcontext := context\n \t\twg.Add(1)\n \t\tgo func() {\n \t\t\tdefer wg.Done()", ...
ollama/ollama
15,065
mlx: fix KV cache snapshot memory leak
mlx.Copy shares the backing buffer with its source (via copy_shared_buffer) rather than allocating independent storage. When used to snapshot a slice of the KV cache, the snapshot array holds the entire original cache buffer alive through the shared data pointer — even after eval detaches the computation graph. Replace Copy with Contiguous in Snapshot and Split. Contiguous allocates a compact buffer when the source buffer is significantly larger than the logical slice (Contiguous::eval checks buffer_size > nbytes + 16384), which is always the case for KV cache slices.
d432aabe6953f7345bbd73f97f1b9f260b6953e1
[ { "filename": "x/mlxrunner/cache/cache.go", "patch": "@@ -109,8 +109,8 @@ func (c *KVCache) Snapshot(fromOffset int) Snapshot {\n \n \tkSlice := c.keys.Slice(mlx.Slice(), mlx.Slice(), mlx.Slice(from, to), mlx.Slice())\n \tvSlice := c.values.Slice(mlx.Slice(), mlx.Slice(), mlx.Slice(from, to), mlx.Slice())\n...
rust-lang/rust
154,725
[beta] backports
- stdarch subtree update rust-lang/rust#153336 (partial) - aarch64: fix UB in non-power-of-two reads and writes rust-lang/stdarch#2042 - add neon load/store assembly test rust-lang/rust#154094 - don't drop arguments' temporaries in `dbg!` rust-lang/rust#154074 - Init self_decl with a correct vis rust-lang/rust#154313 - Update LLVM to 22.1.2 rust-lang/rust#154344 - [perf] Revert FastISel patch rust-lang/rust#154511 - core: Destabilize beta-stable `RangeInclusiveIter::remainder` rust-lang/rust#154459 - Revert "Fix: On wasm targets, call `panic_in_cleanup` if panic occurs in cleanup" rust-lang/rust#154700 - core: Update the feature gate on `TryFrom<integer> for bool` rust-lang/rust#154691 r? cuviper
940fc4d3b3c42525450bb180a71149aba78c7940
[ { "filename": "library/core/src/convert/num.rs", "patch": "@@ -332,7 +332,7 @@ macro_rules! impl_try_from_both_bounded {\n /// Implement `TryFrom<integer>` for `bool`\n macro_rules! impl_try_from_integer_for_bool {\n ($($int:ty)+) => {$(\n- #[stable(feature = \"try_from\", since = \"1.34.0\")]\n+...
vercel/next.js
92,169
fix: propagate scroll={false} through server-patch retries
## Summary `scroll={false}` on `<Link>` is ignored when middleware rewrites to a deeper internal path and `prefetch={false}` is set. The page scrolls to top on every search-param navigation. ## Root Cause When a navigation with `scroll={false}` uses an optimistic route match, `spawnDynamicRequests` fetches dynamic data from the server. If the server response creates a tree mismatch (the page segment key includes search params that differ from the optimistic tree), `dispatchRetryDueToTreeMismatch` dispatches a `SERVER_PATCH` action. The server-patch reducer hardcodes `ScrollBehavior.Default` ([source](https://github.com/vercel/next.js/blob/canary/packages/next/src/client/components/router-reducer/reducers/server-patch-reducer.ts#L45)), dropping the original `scroll={false}`. The retry creates new cache nodes with `scrollRef = { current: true }` that are never neutralized, causing `handlePotentialScroll` to scroll to top. ## Fix Thread `scrollBehavior` from the original navigation through `spawnDynamicRequests` -> `finishNavigationTask` -> `dispatchRetryDueToTreeMismatch` -> `ServerPatchAction`, and use `action.scrollBehavior` in the server-patch reducer instead of the hardcoded default. ## Reproduction **Repo:** https://github.com/damusnet/next-middleware-rewrite-scroll-bug | | Commit | Deployment | Scroll behavior | |---|---|---|---| | **Bug** | [`7f483d0`](https://github.com/damusnet/next-middleware-rewrite-scroll-bug/commit/7f483d0) | [Preview](https://next-scroll-repro-2k8ta88n7-damien-varrons-projects.vercel.app/) | Scrolls to top :x: | | **Fix** | [`fa2b386`](https://github.com/damusnet/next-middleware-rewrite-scroll-bug/commit/fa2b386) | [Production](https://next-scroll-repro.vercel.app/) | Scroll preserved :white_check_mark: | Both deployments include console logging (`[scroll-debug]`) showing RSC fetches and scroll resets. ### Conditions required to trigger: 1. Middleware rewrites a short URL to a deeper internal path (common in i18n/multi-tenant apps) 2. `<Link scroll={false} prefetch={false} href="?param=value">` 3. Deployed on Vercel (ISR) — does not reproduce on local `next start` ## Changes | File | Change | |---|---| | `router-reducer-types.ts` | Add `scrollBehavior` to `ServerPatchAction` | | `ppr-navigations.ts` | Thread `scrollBehavior` through `spawnDynamicRequests` -> `finishNavigationTask` -> `dispatchRetryDueToTreeMismatch` | | `server-patch-reducer.ts` | Use `action.scrollBehavior` instead of `ScrollBehavior.Default` | | `navigation.ts` | Pass `scrollBehavior` to `spawnDynamicRequests` | | `restore-reducer.ts` | Pass `ScrollBehavior.Default` for history traversals |
cb52f4ea239f2582382788d2dfdc43d444eb5373
[ { "filename": ".agents/skills/README.md", "patch": "@@ -0,0 +1,115 @@\n+# Skills Authoring Guide\n+\n+Skills are on-demand context files that Claude loads when relevant. They extend `AGENTS.md` with deep-dive workflows, code templates, and verification steps.\n+\n+## When to Create a Skill\n+\n+Create a ski...
nodejs/node
62,499
crypto: unify asymmetric key import through KeyObjectHandle::Init
Consolidate all asymmetric key import paths (DER/PEM, JWK, raw) into a single KeyObjectHandle::Init() entry point with a uniform signature. Remove the per-type C++ init methods (InitECRaw, InitEDRaw, InitPqcRaw, InitJwk, InitECPrivateRaw) and their JS-side callers, replacing them with shared C++ and JS helpers. createPublicKey, createPrivateKey, sign, verify, and other operations that accept key material now handle JWK and raw formats directly in C++, removing redundant JS-to-C++ key handle round-trips. This also makes JWK import error for keys when their private and public components don't match, this aligns with the behaviour present in other runtimes, most notably browser webcrypto APIs. It was already present in our WebCryptoAPI for some keys but not all, this makes it all as well as `node:crypto`, this behaviour was useful for bypassing the inability to import "raw" private keys but that affordance is also coming with 26.x through its own recognized key format (#62455).
e8621565734120a74dde5e909105c23a69598d29
[ { "filename": "lib/internal/crypto/aes.js", "patch": "@@ -10,6 +10,8 @@ const {\n AESCipherJob,\n KeyObjectHandle,\n kCryptoJobAsync,\n+ kKeyFormatJWK,\n+ kKeyTypeSecret,\n kKeyVariantAES_CTR_128,\n kKeyVariantAES_CBC_128,\n kKeyVariantAES_GCM_128,\n@@ -30,7 +32,6 @@ const {\n const {\n hasA...
electron/electron
50,613
fix: remove menu update debug log
Backport of #50608 See that PR for details. Notes: Removed "representedObject is not a WeakPtrToElectronMenuModelAsNSObject" logging when interacting with macOS menus.
411eeb3dc1edc579379a84d5d864bb092b934008
[ { "filename": "shell/browser/ui/cocoa/electron_menu_controller.mm", "patch": "@@ -478,7 +478,6 @@ - (void)applyStateToMenuItem:(NSMenuItem*)item {\n \n if (![represented\n isKindOfClass:[WeakPtrToElectronMenuModelAsNSObject class]]) {\n- NSLog(@\"representedObject is not a WeakPtrToElectronMe...
huggingface/transformers
45,116
Add full GGUF loading support for GPT‑OSS (fixes #43366)
# What does this PR do? This PR adds full GGUF loading support for GPT‑OSS models (20B/120B). It allows Transformers (and consequently vLLM) to directly load GPT‑OSS GGUF files without falling back to a wrong architecture. The changes include: - Architecture registration in GGUF mappings. - A custom `GptOssTensorProcessor` to handle MoE expert splitting and gate/up interleaving. - Reconstruction of nested `rope_scaling` (YaRN) from flat GGUF metadata. - Tests: fast registration test + slow integration test using a real 20B GGUF file. Fixes #43366, supersedes #43757. Related vLLM issue: https://github.com/vllm-project/vllm/issues/22353 ## Code Agent Policy - [x] I confirm that this is not a pure code agent PR. ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). – *Not applicable, it adds a feature.* - [x] Did you read the contributor guideline, Pull Request section? – Yes. - [x] Was this discussed/approved via a Github issue or the forum? Please add a link to it if that's the case. – Issue #43366, discussion in comments. - [x] Did you make sure to update the documentation with your changes? – Yes! - [x] Did you write any new necessary tests? – Yes, `GptOssGgufLoadingTest` in `tests/models/gpt_oss/test_modeling_gpt_oss.py`. ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Tagging: - @SunMarc (original issue tagger, quantization) - @CyrilVallez (model loading) - @ArthurZucker (tokenizers, model structure)
6e2e655db30a7a68a101b0256b4d5a2fb22f8733
[ { "filename": "src/transformers/image_processing_backends.py", "patch": "@@ -529,7 +529,7 @@ def resize(\n self,\n image: np.ndarray,\n size: SizeDict,\n- resample: Union[\"PILImageResampling\", \"tvF.InterpolationMode\", int] | None = None,\n+ resample: \"PILImageResam...
golang/go
77,686
builtin: update new documentation for Go 1.26 expression syntax
The built-in new function now accepts an expression as its operand, not just a type. Update the doc comment to reflect this change introduced in Go 1.26. The documentation now describes both behaviors: new(T) allocates a zero value of type T new(expr) allocates with the expression value as initial value Fixes #77584
df66c79ea8c8d91c2d2d4e951b87b8b7c3022757
[ { "filename": "src/builtin/builtin.go", "patch": "@@ -122,6 +122,10 @@ type Type int\n // invocation.\n type Type1 int\n \n+// TypeOrExpr is here for the purposes of documentation only. It is a stand-in\n+// for either a Go type or an expression.\n+type TypeOrExpr int\n+\n // IntegerType is here for the pur...
ollama/ollama
15,064
doc: update vscode doc
null
be04444b5c93458d8cb3bc66e32fc1929b4e7c0d
[ { "filename": "docs/integrations/vscode.mdx", "patch": "@@ -8,24 +8,34 @@ VS Code includes built-in AI chat through GitHub Copilot Chat. Ollama models can\n ![VS Code with Ollama](/images/vscode.png)\n \n \n+## Prerequisites\n \n-## Install\n+- Ollama v0.18.3+\n+- [VS Code 1.113+](https://code.visualstudio....
vercel/next.js
92,167
Add rust-fingerprint task and SCCACHE passthrough env
## What Replace duplicated Rust input globs across every turbo task with a single rust-fingerprint task that computes a TURBO_HASH stamp file. All Rust build/check tasks now depend on this stamp instead of repeating the full glob list. Also add `SCCACHE_*` to `globalPassThroughEnv` so `sccache` env vars are available to turbo tasks. ## Why `SCCACHE_*` and `RUSTC_WRAPPER` env vars are required to build under sccache, but do not affect the build in a material way. This simplifies the repeated constructs that are unfortunately difficult to factor out, but is provably the same as listing all the files each time (the rust-fingerprint file's content is its `TURBO_HASH`).
28e0e0ccab008f732e0872349962d5b746b455ce
[ { "filename": "packages/next-swc/package.json", "patch": "@@ -15,6 +15,7 @@\n \"build-native-no-plugin-release\": \"napi build --platform -p next-napi-bindings --cargo-cwd ../../ --cargo-name next_napi_bindings --release --features image-webp,tracing/release_max_level_trace --js false native\",\n \"...
rust-lang/rust
154,718
Rollup of 21 pull requests
Successful merges: - rust-lang/rust#153105 (Compute the result of a projection type with region errors) - rust-lang/rust#153286 (various fixes for scalable vectors) - rust-lang/rust#153532 (Attributes containing rustc) - rust-lang/rust#153960 (Make `layout_of` cycles fatal errors) - rust-lang/rust#154527 (Emit pre-expansion feature gate warnings for negative impls and specialization) - rust-lang/rust#154666 (Remove `StableHashContext` impls) - rust-lang/rust#154669 (Introduce #[diagnostic::on_move] on `Arc`) - rust-lang/rust#154710 (opaque_generic_const_args -> generic_const_args) - rust-lang/rust#154712 (Revert "`-Znext-solver` Remove the forced ambiguity hack from search graph") - rust-lang/rust#153614 (`FindParamInClause` handle edge-cases) - rust-lang/rust#154213 (tidy-alphabetical: fix line number in error message) - rust-lang/rust#154425 (Migrate transmute tests) - rust-lang/rust#154442 (Export `derive` at the crate root: `core::derive` and `std::derive`) - rust-lang/rust#154469 (mGCA: Lower spans for literal const args) - rust-lang/rust#154578 (Rename `probe_ty_var` to `try_resolve_ty_var`) - rust-lang/rust#154615 (Moving issues) - rust-lang/rust#154644 (rustdoc: seperate methods and associated functions in sidebar) - rust-lang/rust#154660 (Avoid creating async return opaques for foreign async fns) - rust-lang/rust#154671 (Add a test for a past ICE when calling a const fn of an unresolved type with the wrong number of args) - rust-lang/rust#154680 ([rustdoc] Replace `DocContext` with `TyCtxt` wherever possible) - rust-lang/rust#154709 (Revert `Ty` type alias in `rustc_type_ir`) <!-- homu-ignore:start --> r? @ghost [Create a similar rollup](https://bors.rust-lang.org/queue/rust?prs=153105,153286,153532,153960,154527,154666,154669,154710,154712,153614,154213,154425,154442,154469,154578,154615,154644,154660,154671,154680,154709) <!-- homu-ignore:end -->
f3ba2c41e63ac58ae8aafbe7725db6b5e6021d00
[ { "filename": "src/librustdoc/clean/inline.rs", "patch": "@@ -137,7 +137,7 @@ pub(crate) fn try_inline(\n })\n }\n Res::Def(DefKind::Macro(kinds), did) => {\n- let mac = build_macro(cx, did, name, kinds);\n+ let mac = build_macro(cx.tcx, did, name, kinds);\n...
electron/electron
50,611
fix: extension service workers not starting beyond first app launch
#### Description of Change Fixes https://github.com/electron/electron/issues/41613 ##### Before this patch When an extension was loaded into an app for the very first time, its service worker would start. When the same extension was loaded on subsequent app launches, the service worker would no longer start. ##### With this patch This PR makes it so that extension service workers are also started on subsequent app launches. #### Checklist <!-- Remove items that do not apply. For completed items, change [ ] to [x]. --> - [x] PR description included - [x] I have built and tested this PR - [x] `npm test` passes - [x] tests are [changed or added](https://github.com/electron/electron/blob/main/docs/development/testing.md) - [x] [PR release notes](https://github.com/electron/clerk/blob/main/README.md) describe the change in a way relevant to app developers, and are [capitalized, punctuated, and past tense](https://github.com/electron/clerk/blob/main/README.md#examples). #### Release Notes Notes: Fixed certain DevTools extension panels not showing without a page reload.
4a58f1b634d8855f83959e350955a59c92dab369
[ { "filename": "shell/browser/extensions/electron_extension_loader.cc", "patch": "@@ -28,6 +28,7 @@\n #include \"extensions/common/error_utils.h\"\n #include \"extensions/common/file_util.h\"\n #include \"extensions/common/manifest_constants.h\"\n+#include \"extensions/common/manifest_handlers/background_inf...
facebook/react
36,024
[Flight] Clear chunk reason after successful module initialization
When `requireModule` triggers a reentrant `readChunk` on the same module chunk, the reentrant call can fail and set `chunk.reason` to an error. After the outer `requireModule` succeeds, the chunk transitions to initialized but retains the stale error as `reason`. When the Flight response stream later closes, it iterates all chunks and expects `reason` on initialized chunks to be a `FlightStreamController`. Since the stale `reason` is an `Error` object instead, calling `chunk.reason.error()` crashes with `TypeError: chunk.reason.error is not a function`. The reentrancy can occur when module evaluation synchronously triggers `readChunk` on the same chunk — for example, when code called during evaluation tries to resolve the client reference for the module that is currently being initialized. In Fizz SSR, `captureOwnerStack()` can trigger this because it constructs component stacks that resolve lazy client references via `readChunk`. The reentrant `requireModule` call returns the module's namespace object, but since the module is still being evaluated, accessing the export binding throws a TDZ (Temporal Dead Zone) `ReferenceError`. This sets the chunk to the errored state, and the `ReferenceError` becomes the stale `chunk.reason` after the outer call succeeds. This scenario is triggered in Next.js when a client module calls an instrumented API like `Math.random()` in module scope, which synchronously invokes `captureOwnerStack()`.
b43ee2a3d98863a2ea6dcf63d1b66e0b93a59cd2
[ { "filename": "packages/react-client/src/ReactFlightClient.js", "patch": "@@ -1040,6 +1040,8 @@ function initializeModelChunk<T>(chunk: ResolvedModelChunk<T>): void {\n // Initialize any debug info and block the initializing chunk on any\n // unresolved entries.\n initializeDebugChunk(response, ...
huggingface/transformers
45,112
[CB] Add warmup feature
This PR adds a warmup phase before generation starts, turned on by default. It allows for better diagnostics and a more representative user experience than without warmup, where the cost of wamup is payed during the first request rather than in a phase before generation starts. Warmup is skipped if cuda graphs and compile are both turned off.
f29aca3f70b14ce9dd6ded52dfb8af0e89efff15
[ { "filename": "docs/source/en/_toctree.yml", "patch": "@@ -89,6 +89,8 @@\n title: Experts backends\n - local: continuous_batching\n title: Continuous batching\n+ - local: continuous_batching_architecture\n+ title: Continuous batching architecture\n - local: paged_attention\n ...
nodejs/node
62,497
Revert "deps: update libuv to 1.52.1"
This reverts commit 5bebd7eaea287ccf187b02c8d6afa863cf326a58. The update appears to be causing persistent flakiness in test-http(s), test-net-*, and test-tls-* tests on Windows. The flakiness manifests as an abort apparently caused by a race condition introduced in connection cleanup on process teardown. Root cause still needs to be fully determined. Reverting the update will hopefully get us back to a stable state while the underlying issue is investigated and resolved.
8d72d50b13c5e67b716a0f9ee5ef7fce5369b3fe
[ { "filename": "deps/uv/.clang-tidy", "patch": "@@ -1,47 +0,0 @@\n----\n-# Configuration file for clang-tidy\n-# This configuration is tailored for the libuv C library project\n-\n-# Use default checks with minimal necessary disables\n-Checks:\n- # don't suggest reordering struct definitions\n- - -clang-an...
golang/go
77,683
cmd/fix: clarify -any analyzer behavior in documentation
This change clarifies the documentation for the any analyzer used by go fix. It adds a note explaining that the -any flag enables only the any analyzer and does not affect other analyzers in the fix suite. This makes the behavior of the flag explicit and helps avoid confusion when reading the documentation. This is a documentation-only change and does not modify analyzer behavior. fixes: #77679
f9e337d1e857c0ba66f6723b4147c9c08f93208a
[ { "filename": "api/next/73450.txt", "patch": "@@ -0,0 +1 @@\n+pkg net/url, method (*URL) Clone() *URL #73450", "additions": 1, "deletions": 0 }, { "filename": "doc/next/6-stdlib/99-minor/net/url/73450.md", "patch": "@@ -0,0 +1 @@\n+The new [URL.Clone] method creates a deep copy of a URL....
rust-lang/rust
154,716
Rollup of 21 pull requests
Successful merges: - rust-lang/rust#153105 (Compute the result of a projection type with region errors) - rust-lang/rust#153286 (various fixes for scalable vectors) - rust-lang/rust#153532 (Attributes containing rustc) - rust-lang/rust#153960 (Make `layout_of` cycles fatal errors) - rust-lang/rust#154527 (Emit pre-expansion feature gate warnings for negative impls and specialization) - rust-lang/rust#154666 (Remove `StableHashContext` impls) - rust-lang/rust#154669 (Introduce #[diagnostic::on_move] on `Arc`) - rust-lang/rust#154710 (opaque_generic_const_args -> generic_const_args) - rust-lang/rust#154712 (Revert "`-Znext-solver` Remove the forced ambiguity hack from search graph") - rust-lang/rust#154713 (Stop compiling when we get resolving crate failure) - rust-lang/rust#154213 (tidy-alphabetical: fix line number in error message) - rust-lang/rust#154425 (Migrate transmute tests) - rust-lang/rust#154442 (Export `derive` at the crate root: `core::derive` and `std::derive`) - rust-lang/rust#154469 (mGCA: Lower spans for literal const args) - rust-lang/rust#154578 (Rename `probe_ty_var` to `try_resolve_ty_var`) - rust-lang/rust#154615 (Moving issues) - rust-lang/rust#154644 (rustdoc: seperate methods and associated functions in sidebar) - rust-lang/rust#154660 (Avoid creating async return opaques for foreign async fns) - rust-lang/rust#154671 (Add a test for a past ICE when calling a const fn of an unresolved type with the wrong number of args) - rust-lang/rust#154680 ([rustdoc] Replace `DocContext` with `TyCtxt` wherever possible) - rust-lang/rust#154709 (Revert `Ty` type alias in `rustc_type_ir`) <!-- homu-ignore:start --> r? @ghost [Create a similar rollup](https://bors.rust-lang.org/queue/rust?prs=153105,153286,153532,153960,154527,154666,154669,154710,154712,154713,154213,154425,154442,154469,154578,154615,154644,154660,154671,154680,154709) <!-- homu-ignore:end -->
f3ba2c41e63ac58ae8aafbe7725db6b5e6021d00
[ { "filename": "src/librustdoc/clean/inline.rs", "patch": "@@ -137,7 +137,7 @@ pub(crate) fn try_inline(\n })\n }\n Res::Def(DefKind::Macro(kinds), did) => {\n- let mac = build_macro(cx, did, name, kinds);\n+ let mac = build_macro(cx.tcx, did, name, kinds);\n...
electron/electron
50,608
fix: remove menu update debug log
#### Description of Change Removes a diagnostic log: ``` representedObject is not a WeakPtrToElectronMenuModelAsNSObject ``` This log was introduced alongside a guard in macOS menu refresh. The guard is part of the system's expected behavior, preventing us from processing menu items we don't own. In other words, the guard itself is benign. Keep the guard, remove the log. Fixes #50389 #### Checklist - [x] PR description included - [x] I have built and tested this PR - [x] `npm test` passes - [x] [PR release notes](https://github.com/electron/clerk/blob/main/README.md) describe the change in a way relevant to app developers, and are [capitalized, punctuated, and past tense](https://github.com/electron/clerk/blob/main/README.md#examples). #### Release Notes Notes: Removed "representedObject is not a WeakPtrToElectronMenuModelAsNSObject" logging when interacting with macOS menus.
5689abff4d2a9a292b57bf2fe6c72e308fd2afc2
[ { "filename": "shell/browser/ui/cocoa/electron_menu_controller.mm", "patch": "@@ -478,7 +478,6 @@ - (void)applyStateToMenuItem:(NSMenuItem*)item {\n \n if (![represented\n isKindOfClass:[WeakPtrToElectronMenuModelAsNSObject class]]) {\n- NSLog(@\"representedObject is not a WeakPtrToElectronMe...
vercel/next.js
92,165
#64441 bug: fix not scrolling top issue
# Fix scroll restoration on route change with sticky headers **Actual problem:** The scroll restoration currently runs before layout is fully applied. In cases involving sticky elements and short pages, this results in no visible scroll change. **Tests:** I added an integration test under `test/integration/scroll-restoration-on-navigation` to cover this scenario. Due to local environment limitations (Playwright dependencies on my OS), I wasn’t able to run it locally, but the behavior has been verified manually in a real browser and it is **working as expected** now. Happy to adjust the test if CI surfaces any issues or if there’s a preferred testing approach. Covers the behavior described in Next.js GitHub Issue #64441
f8a22e0c4fb7d084306dbbfa1718a468d366ac6d
[ { "filename": ".config/eslintignore.mjs", "patch": "@@ -47,6 +47,7 @@ export default globalIgnores([\n 'examples/with-typescript-graphql/lib/gql/',\n 'test/development/basic/hmr/components/parse-error.js',\n 'test/development/mcp-server/fixtures/default-template/app/build-error/page.tsx',\n+ 'test/de...
facebook/react
36,023
Create index.html
<!-- Thanks for submitting a pull request! We appreciate you spending the time to work on these changes. Please provide enough information so that others can review your pull request. The three fields below are mandatory. Before submitting a pull request, please make sure the following is done: 1. Fork [the repository](https://github.com/facebook/react) and create your branch from `main`. 2. Run `yarn` in the repository root. 3. If you've fixed a bug or added code that should be tested, add tests! 4. Ensure the test suite passes (`yarn test`). Tip: `yarn test --watch TestName` is helpful in development. 5. Run `yarn test --prod` to test in the production environment. It supports the same options as `yarn test`. 6. If you need a debugger, run `yarn test --debug --watch TestName`, open `chrome://inspect`, and press "Inspect". 7. Format your code with [prettier](https://github.com/prettier/prettier) (`yarn prettier`). 8. Make sure your code lints (`yarn lint`). Tip: `yarn linc` to only check changed files. 9. Run the [Flow](https://flowtype.org/) type checks (`yarn flow`). 10. If you haven't already, complete the CLA. Learn more about contributing: https://reactjs.org/docs/how-to-contribute.html --> ## Summary <!-- Explain the **motivation** for making this change. What existing problem does the pull request solve? --> ## How did you test this change? <!-- Demonstrate the code is solid. Example: The exact commands you ran and their output, screenshots / videos if the pull request changes the user interface. How exactly did you verify that your PR solves the issue you wanted to solve? If you leave this empty, your PR will very likely be closed. -->
abb92660d683958c5e76604fb77f26be5aa305dd
[ { "filename": "examples/todo-list/index.html", "patch": "@@ -0,0 +1,414 @@\n+<!DOCTYPE html>\n+<html lang=\"zh-CN\">\n+<head>\n+ <meta charset=\"UTF-8\" />\n+ <meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\" />\n+ <title>React Todo List Example</title>\n+ <style>\n+ * {\n+ ...
huggingface/transformers
45,111
Fix double softmax in MoE router load-balancing loss
## Summary - Several MoE routers applied softmax inside `forward()` but returned the result as `router_logits`. The `load_balancing_loss_func` then applied softmax **again**, computing the aux loss on `softmax(softmax(logits))` which flattens the distribution toward uniform, rendering the load-balancing loss ineffective. - Fix: use a separate `router_probs` variable for the softmaxed values used in top-k routing, keeping `router_logits` as raw logits so the loss function's single softmax is correct. ### Source modular files fixed (3): - `mixtral/modular_mixtral.py` — `MixtralTopKRouter` - `qwen2_moe/modular_qwen2_moe.py` — `Qwen2MoeTopKRouter` - `qwen3_vl_moe/modular_qwen3_vl_moe.py` — `Qwen3VLMoeTextTopKRouter` ### Downstream models regenerated by `make fix-repo` (10): mixtral, minimax, qwen2_moe, olmoe, flex_olmo, qwen3_moe, qwen3_next, qwen3_omni_moe, qwen3_vl_moe, qwen3_5_moe ## Test plan - [x] Verified with debug script that `router_logits` now returns raw logits (row sums ≠ 1.0) instead of probabilities - [x] Confirmed `load_balancing_loss_func` applies softmax exactly once - [ ] Run MoE model tests: `pytest tests/models/mixtral/ tests/models/qwen2_moe/ tests/models/qwen3_vl_moe/ -v` 🤖 Generated with [Claude Code](https://claude.com/claude-code)
fb4d936ca1782d8b4311a1a1d1dba0489ae8f7d5
[ { "filename": "src/transformers/models/flex_olmo/modeling_flex_olmo.py", "patch": "@@ -300,11 +300,11 @@ def __init__(self, config):\n def forward(self, hidden_states):\n hidden_states = hidden_states.reshape(-1, self.hidden_dim)\n router_logits = F.linear(hidden_states, self.weight) # ...