id
int64 393k
2.82B
| repo
stringclasses 68
values | title
stringlengths 1
936
| body
stringlengths 0
256k
⌀ | labels
stringlengths 2
508
| priority
stringclasses 3
values | severity
stringclasses 3
values |
|---|---|---|---|---|---|---|
533,084,582 | pytorch | Master Task: JIT/C++ Parity With Pytorch Python API | ## 🚀 JIT/C++ Parity With Pytorch Python API
There are a number of reasons why pytorch ops might work in python but not in JIT / C++.
Tracking issues:
- [Ops bound in Python Layer for Legacy Reasons](https://github.com/pytorch/pytorch/issues/30780)
- [Tensor Properties are not natively declarable](https://github.com/pytorch/pytorch/issues/30775)
- [Fast path functions are not natively declarable](https://github.com/pytorch/pytorch/issues/30774)
- [Tensor Options create Schema Mismatch for Ops With dtype, layout, device tuple](https://github.com/pytorch/pytorch/issues/30763)
- [Requires_grad argument is implicitly added to schema of factory functions](https://github.com/pytorch/pytorch/issues/30761)
- [Ops bound in torch/tensor.py and torch/functional.py are python-only](https://github.com/pytorch/pytorch/issues/30786)
- [[jit] Python arg parser / TorchScript incompatibilities](https://github.com/pytorch/pytorch/issues/33887)
cc @suo | oncall: jit,triaged | low | Minor |
533,144,226 | rust | rustdoc should versionsort types in trait implementations | In reading the [documentation for `std::num::Wrapping`](https://doc.rust-lang.org/std/num/struct.Wrapping.html), I noticed (in the left sidebar) a list of trait implementations for `Wrapping` of various numeric types, which appeared in this order:
- `Wrapping<i128>`
- `Wrapping<i16>`
- `Wrapping<i32>`
- `Wrapping<i64>`
- `Wrapping<i8>`
- `Wrapping<isize>`
This seems to sort numbers as strings, rather than by numeric value.
I would propose that everywhere rustdoc sorts identifiers of any kind, it should always use the versionsort algorithm, which sorts numbers by numeric value rather than as strings.
The commit at https://github.com/rust-lang/rustfmt/pull/3764 contains an implementation of the versionsort comparison algorithm. I would suggest putting that (and its tests) into a common implementation for rustdoc and other future callers like rustfmt. | T-rustdoc,C-enhancement,A-docs | low | Minor |
533,156,045 | pytorch | Pytorch openmp thread number tuning option for CPU trainning | ## 🚀 Feature
Allow an option to user so that user can choose the openmp thread either be 1 or max thread number.
## Motivation
We found that the tuning of paralel_for() makes the openmp thread number changing across the layers and we suspect this is the reason why it causes the performance difference. On the optimized DLRM version, , we have found that the performance runs best when all the operations uses max thread number.
In the CPU trainning scenario where we assume all cpus dedicated for training, changing the thread number for some ops may hurt the locality, and also the transient low thread number may not really help the CPU utilization because usually we want the working thread spining a while to expect something coming soon.
## Additional context
We reported DLRM perf regression (https://github.com/pytorch/pytorch/issues/28198) due to PR https://github.com/pytorch/pytorch/pull/26963, which choose number of openmp threads for parallel_for according to the grain size and task size. When the task size is X times bigger than grain size, the number of openmp threads is set to Min(X, max thread number). The max thread number is specified by the user OMP_NUM_THREAD (or default as physical core#). Priori to the change, the number of openmp threads is either 1 or max thread number.
To address the regression, a further tuning was made on PR28770 (https://github.com/pytorch/pytorch/pull/28770). The PR identified THFloatTensor_fill (used by cpu_zero) not using all threads. It reported the THxxx uses TH_OMP_OVERHEAD_THRESHOLD 100K which is about 3x higher than at::internal::GRAIN_SIZE. The PR changed TH_OMP_OVERHEAD_THRESHOLD to be the same as at::internal::GRAIN_SIZE and reported the performance is largely back (back from 3505.85ms/iter to 3453.23ms/iter), since it allows more threads being used for THFloatTensor_fill.
But the tuning doesn't solve all the regression. When we further optimize the DLRM to the level of ~36ms/iter (yes ~100x better than original version on SKX 8180), we observed that a changing number of openmp threads among different layers hurt performance. The tuning of openmp threads for
at::parallel_for() impacts lots of operations, including these implemented on
TensorIterator::for_each() and cpu_kernel_vec() since these build on top of at::parallel_for(). For example, index_put() uses TensorIterator::for_each() and relu() uses cpu_kernel_vec().
We found that the tuning of paralel_for() makes the openmp thread number changing across the layers and we suspect this is the reason why it causes the performance difference. On the optimized DLRM version, we have found that the performance runs best when all the operations uses max thread number. As an experiment, we change the cpu_kernel_vec() and cpu_index_kernel() interface so that it effectively launch max thread number (PR 29705 https://github.com/pytorch/pytorch/pull/29705), and we found the performance regression is fully back (38.25 ms/iter vs. 36.44 ms/iter). This is consistent to our performance tuning BKM for using openmp in training scenario.
With various tuning, we doubt whether it makes sense for Pytorch to tune the openmp thread number per the grain size in the trainning scenario where we assume all cpus dedicated for training. Changing the thread number for some ops may hurt the locality, and also the transient low thread number may not really help the CPU utilization because usually we want the free working thread spining a while to expect something coming soon. Maybe we allow an option to user so that user can choose the openmp thread either be 1 or max thread number.
cc @VitalyFedyunin @ngimel @mruberry | module: performance,module: cpu,triaged,enhancement,module: multithreading | low | Major |
533,173,458 | pytorch | Sending sparse tensors over RPC not yet supported | ## 🐛 Bug
## To Reproduce
Add the following test in `dist_autograd_test.py`
```
@dist_init
def test_sparse_tensors(self):
t1 = torch.rand((3, 3), requires_grad=True).to_sparse()
t2 = torch.rand((3, 3), requires_grad=True).to_sparse()
rpc.rpc_sync("worker{}".format(self._next_rank()), torch.add, args=(t1, t2))
```
This fails with the following stack trace (need to catch exceptions in ProcessGroupAgent::enqueueSend to see this):
```
sparse tensors do not have storage (storage at caffe2/aten/src/ATen/SparseTensorImpl.cpp:87)
> frame #0: <unknown function>
> frame #1: <unknown function>
> frame #2: std::function<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > ()>::operator()() const
> frame #3: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)
> frame #4: at::SparseTensorImpl::storage() const
> frame #5: <unknown function>
> frame #6: torch::jit::Pickler::pushStorageOfTensor(at::Tensor const&)
> frame #7: torch::jit::Pickler::pushLiteralTensor(c10::IValue const&)
> frame #8: torch::jit::Pickler::pushTensor(c10::IValue const&)
> frame #9: torch::jit::Pickler::pushIValueImpl(c10::IValue const&)
> frame #10: torch::jit::Pickler::pushIValue(c10::IValue const&)
> frame #11: <unknown function>
> frame #12: <unknown function>
> frame #13: std::function<void (c10::IValue const&)>::operator()(c10::IValue const&) const
> frame #14: torch::jit::Pickler::pushSpecializedList(c10::IValue const&, torch::jit::PicklerClass, std::function<void (c10::IValue const&)> const&)
> frame #15: torch::jit::Pickler::pushIValueImpl(c10::IValue const&)
> frame #16: torch::jit::Pickler::pushIValue(c10::IValue const&)
> frame #17: torch::distributed::rpc::wireSerialize[abi:cxx11](std::vector<char, std::allocator<char> > const&, std::vector<at::Tensor, std::allocator<at::Tensor> > const&)
> frame #18: <unknown function>
> frame #19: <unknown function>
> frame #20: <unknown function>
> frame #21: <unknown function>
> frame #22: <unknown function>
> frame #23: <unknown function>
> frame #24: std::function<void ()>::operator()() const
> frame #25: c10::ThreadPool::main_loop(unsigned long)
> frame #26: <unknown function>
> frame #27: <unknown function>
> frame #28: <unknown function>
> frame #29: <unknown function>
> frame #30: <unknown function>
> frame #31: <unknown function>
> frame #32: <unknown function>
> frame #33: <unknown function>
> frame #34: clone
```
cc @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @gqchen @aazzolini @rohan-varma @xush6528 | module: serialization,triaged,module: rpc | low | Critical |
533,242,138 | pytorch | Tensorboard graph failing to nest modules | ## 🐛 Bug
See [this discussion](https://discuss.pytorch.org/t/tensorboard-graph-failing-to-nest-modules/62993). Basically tensorboard logging no longer works as expected for me (since version version 1.2 I think, as function `graph` changed a lot at this point). It doesn't nest modules anymore and just puts all leaf parameters directly under the main module. For torchvision's resnet34, it looks like that:
## To Reproduce
Steps to reproduce the behavior:
1. Install pytorch 1.3 with tensorboard 2.0 (though I don't think tensorboard version matters much)
1. `writer = SumarryWriter()`
1. `writer.add_graph(resnet34(), torch.rand(8, 3, 224, 224))`
## Expected behavior
Graph is successfully added and looks like that (obtained from an environment with pytorch 1.1):
## Environment
```bash
PyTorch version: 1.3.1
Is debug build: No
CUDA used to build PyTorch: 10.0.130
OS: CentOS Linux 7 (Core)
GCC version: (GCC) 7.3.0
CMake version: Could not collect
Python version: 3.7
Is CUDA available: Yes
CUDA runtime version: Could not collect
GPU models and configuration:
GPU 0: Tesla P100-PCIE-16GB
GPU 1: Tesla P100-PCIE-16GB
Nvidia driver version: 410.48
cuDNN version: Probably one of the following:
/usr/local/cuda-8.0/targets/x86_64-linux/lib/libcudnn.so.6
/usr/local/cuda-8.0/targets/x86_64-linux/lib/libcudnn.so.7
/usr/local/cuda-9.0/targets/x86_64-linux/lib/libcudnn.so.7
Versions of relevant libraries:
[pip3] numpy==1.16.1
[conda] captum 0.1.0 0 pytorch
[conda] cuda100 1.0 0 pytorch
[conda] mkl 2019.4 243
[conda] mkl-service 2.3.0 py37he904b0f_0
[conda] mkl_fft 1.0.14 py37h516909a_1 conda-forge
[conda] mkl_random 1.1.0 py37hb3f55d8_0 conda-forge
[conda] pytorch 1.3.1 py3.7_cuda10.0.130_cudnn7.6.3_0 pytorch
[conda] pytorch-lightning 0.5.3.2 pypi_0 pypi
[conda] torchsummary 1.5.1 pypi_0 pypi
[conda] torchvision 0.4.2 pypi_0 pypi
```
| oncall: visualization | low | Critical |
533,351,782 | angular | FormGroup not dirty when a nested FormControl changes from disabled to enabled, or vice versa. | <!--🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅
Oh hi there! 😄
To expedite issue processing please search open and closed issues before submitting a new one.
Existing issues often contain information about workarounds, resolution, or progress updates.
🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅🔅-->
# 🐞 bug report
### Affected Package
@angular/forms
### Is this a regression?
No
### Description
A [FormGroup](https://angular.io/api/forms/FormGroup) is not considered [dirty](https://angular.io/api/forms/AbstractControl#dirty) when a nested [FormControl](https://angular.io/api/forms/FormControl) is changed from disabled to enabled, or vice versa. The [pristine](https://angular.io/api/forms/AbstractControl#pristine), [dirty](https://angular.io/api/forms/AbstractControl#dirty), [touched](https://angular.io/api/forms/AbstractControl#touched), and [untouched](https://angular.io/api/forms/AbstractControl#untouched) properties remain unchanged when a control changes from [enabled](https://angular.io/api/forms/AbstractControl#enable) to [disabled](https://angular.io/api/forms/AbstractControl#disable), or vice versa.
## 🔬 Minimal Reproduction
https://stackblitz.com/edit/formbuilder-6cjren
## 🌍 Your Environment
**Angular Version:**
<pre><code>8.0.0</code></pre>
**Anything else relevant?**
Posted more detail on StackOverflow: https://stackoverflow.com/questions/59184512/get-all-angular-form-controls-that-have-changed-from-enabled-to-disabled-or-vic
| area: forms,P3 | low | Critical |
533,363,813 | scrcpy | Put a small "Donate" badge at the top of README.MD | https://blog.rom1v.com/2018/03/introducing-scrcpy/#comment-65 | donation | low | Minor |
533,428,415 | rust | thread_local vs __thread: implementation and docs | The docs for [localkey say](https://doc.rust-lang.org/std/thread/struct.LocalKey.html):
> This key uses the fastest possible implementation available to it for the target platform. It is instantiated with the thread_local! macro and the primary method is the with method.
This is sort of true, but sort of not at the same time. I'm not sure if this means the docs should be clarified, or if this is an implementation bug.
Consider this, for example: https://godbolt.org/z/WEvrPn
simply accessing a value in TLS is much, much heavier in Rust than in C. | I-slow,T-libs-api,A-docs,A-thread-locals,C-bug | low | Critical |
533,468,759 | rust | Tracking issue for the `convert::FloatToInt` trait | PR https://github.com/rust-lang/rust/pull/66841 adds this supporting trait for the `{f32,f64}::approx_unchecked_to` method (tracking issue: https://github.com/rust-lang/rust/issues/67058) to the `core::convert` and `std::convert` modules:
```rust
mod private {
pub trait Sealed {}
}
pub trait FloatToInt<Int>: private::Sealed + Sized {
#[doc(hidden)]
unsafe fn approx_unchecked(self) -> Int;
}
// Macro-generated for all combinations of primitive integer
// and primitive floating point types:
impl FloatToInt<$Int> for $Float {…}
```
The name and location (and usage of the sealed-trait pattern) are chosen to allow later adding additional methods with different conversion semantics between those types, as proposed in https://internals.rust-lang.org/t/pre-rfc-add-explicitly-named-numeric-conversion-apis/11395 | T-libs-api,B-unstable,C-tracking-issue,A-floating-point,Libs-Tracked,S-tracking-perma-unstable | low | Major |
533,498,704 | create-react-app | Possible fix for webpackHotDevClient disconnecting from server. | Hello I would like to submit a fix for the issue people are having with the websocket in webpackHotDevClient.js disconnecting from the server. The websocket is timing out after some time, so we simply ping the server periodically to prevent it from timing out. Here is the code I've been using.
```
// Ping the server to prevent the client from disconnecting
var ping = function() {
if (connection && (connection.readyState === connection.OPEN)) {
connection.send("ping");
setTimeout(ping, 3000);
}
};
connection.onopen = function() {
if (typeof console !== 'undefined' && typeof console.info === 'function') {
console.info(
'The development server has connected.'
);
}
ping();
};
```
I have a branch ready here. I can submit a PR.
https://github.com/vibelife/create-react-app/blob/master/packages/react-dev-utils/webpackHotDevClient.js#L85 | issue: bug | low | Major |
533,518,931 | rust | `rustc_parser` incorrectly parses groups with `Delimiter::None` | Such groups can currently be created by proc macros.
Reproduction, proc macro crate:
```rust
#[proc_macro]
pub fn add_mul(_: TokenStream) -> TokenStream {
// ⟪ 2 + 2 ⟫ * 2
vec![
TokenTree::from(Group::new(Delimiter::None, "2 + 2".parse().unwrap())),
TokenTree::from(Punct::new('*', Spacing::Alone)),
TokenTree::from(Literal::u8_unsuffixed(2)),
].into_iter().collect()
}
```
User crate:
```rust
fn main() {
let x = add_mul!();
assert_eq!(x, 8); // FAIL: the result is 6 != 8
}
```
This is not a huge issue right now because proc macros have no reasons to created `Delimiter::None` delimiters themselves (but they can still get them from input and return them if they return their input partially or fully, see the example below).
However, this will became a huge issue when interpolated tokens like `$e: expr` migrate from AST pieces to token streams, because in the token stream model they are represented exactly like groups with `Delimiter::None` delimiters (proc macros already see them in this way). | A-parser,A-macros,T-lang,C-bug | low | Major |
533,519,108 | pytorch | Torch getting stuck transfering model to GPU in multiple GPU setting | While using torch multiprocessing pool to paralelize the run of multiple equal experiences in multiple GPUs I ran into the fact that a lot of times the processes block in the transfer of the model from CPU to the assigned GPU. I don't understand the reason of the blocking.
I've tried manually deleting the model at the end of each run and collecting the not used GPU memory to no avail. I would like to understand why.
I'm running this on python 3.7 in a linux system with pytorch 1.2.0 and CUDA 10.0.130 with 4 NVIDIA 1080.
My main code is the following. VGG16 is just a big network. I can add the code of it too if you feel it's necessary.
```
import torch
from torch.multiprocessing import Process, Manager, Pool
from VGG16 import VGG16
def run_experiment(name, manager_dict, manager_lock):
device_id = None
manager_lock.acquire()
for device_id in manager_dict.keys():
if device_id == "n_processes":
continue
if manager_dict[device_id] < manager_dict["n_processes"]:
manager_dict[device_id] += 1
device_id = device_id
break
manager_lock.release()
print(manager_dict, flush=True)
device = torch.device(device_id)
print(str(name)+ "selected device: " + str(device_id), flush=True)
torch.cuda.ipc_collect()
torch.cuda.empty_cache()
torch.cuda.synchronize(device)
print(str(name), "Emptied cuda")
model = VGG16()
print(str(name), "Initialized model")
model = model.to(device)
print(str(name), "LOADED MODEL to the device")
manager_dict[device_id] -= 1
model_cpu = model.to(torch.device("cpu"))
del model
print(name, "Should have returned", flush=True)
#return 1
return model_cpu.state_dict()
def normal_experiments():
n_processes_per_gpu = 2
manager = Manager()
manager_dict = manager.dict()
manager_dict["n_processes"] = n_processes_per_gpu
n_gpu = torch.cuda.device_count()
print("Running on %d"%n_gpu, flush=True)
for i in range(n_gpu):
manager_dict[i] = 0
manager_lock = manager.Lock()
with Pool(processes=n_processes_per_gpu*n_gpu) as pool:
#We call run experiment with seeds with its type of arguments
iterable_arguments = [(str(i), manager_dict, manager_lock) for i in range(30)]
res = pool.starmap(run_experiment, iterable_arguments)
print("Done")
def main():
torch.multiprocessing.set_start_method('spawn')
normal_experiments()
if __name__ == "__main__":
main()
```
cc @ngimel | needs reproduction,module: multiprocessing,module: cuda,triaged | low | Minor |
533,535,274 | rust | Clever recursive types + deref = seemingly infinite compile time | The following takes an unknown, potentially infinite amount of time greater than 10 minutes to compile:
```
struct S<T>(T);
impl<T> std::ops::Deref for S<T> {
type Target = S<(T,T)>;
fn deref(&self) -> &Self::Target {
self.thing()
}
}
```
As @rkruppe pointed out on Discord,
> Most places in rustc are supposed to have limits to stop them from going in circles forever (or just for a very long time), so at first glance I'd consider this a bug | E-needs-test,T-compiler,C-bug,I-hang | low | Critical |
533,567,571 | godot | Errors when instancing a sprite that contains a viewport | **Godot version:** commit 9002b045c (Godot 3.2 beta)
**OS/device including version:** Linux Mint 19
**Issue description:**
Instancing a Sprite3D with a ViewportTexture that points to a viewport which descends from it generates many errors and doesn't render correctly.
**Minimal reproduction project:**
[viewport_wierdness.zip](https://github.com/godotengine/godot/files/3928812/viewport_wierdness.zip)
| bug,topic:core,topic:rendering,confirmed | low | Critical |
533,569,447 | kubernetes | Pod.Status.PodIP not updated during postStart lifecycle hook | <!-- Please use this template while reporting a bug and provide as much info as possible. Not doing so may result in your bug not being addressed in a timely manner. Thanks!
If the matter is security related, please disclose it privately via https://kubernetes.io/security/
-->
**What happened**:
When using a postStart lifecycle hook, the Pod.Status.PodIP is not updated while the hook is running, even though the pod has been assigned an IP via the CNI plugin and networking is set up.
**What you expected to happen**:
The Pod.Status.PodIP should be set as soon as it's available.
**How to reproduce it (as minimally and precisely as possible)**:
The following manifest should repro the issue. Apply it and then watch the pod status.
```
apiVersion: v1
kind: Pod
metadata:
name: test-calico-post-start
spec:
containers:
- name: test-calico
image: byrnedo/alpine-curl:latest
command:
- sh
- -c
- "sleep 400000"
lifecycle:
postStart:
exec:
command: [ "/bin/sh","-c","sleep 120 && result=$(curl http://www.google.com 1> /var/log/output.log 2> /var/log/output-error.log ); echo $result"]
```
The postStart lifecycle hook here contains a sleep for 120 seconds. During the first 120 seconds, the pod status will be ContainerCreating and the podIP will not be set. Only after the hook completes will the podIP be set.
**Anything else we need to know?**:
Why is this important? It prevents Calico network policy from operating correctly during the postStart hook since we cannot learn the pod's IP address. (Note that when Calico is also the CNI plug-in, we do learn it, but Calico is designed to run in "policy only mode" on top of other CNI plug-ins, like the AWS VPC-CNI plug-in). C.f. https://github.com/projectcalico/libcalico-go/issues/1125 and https://github.com/projectcalico/felix/issues/2008 for users who are having problems here.
**Environment**:
- Kubernetes version (use `kubectl version`): v1.16.3
- Cloud provider or hardware configuration: AWS
- OS (e.g: `cat /etc/os-release`): Ubuntu 16.04
- Kernel (e.g. `uname -a`): 4.4.0-1098-aws
- Install tools: kubeadm
- Network plugin and version (if this is a network-related bug): amazon-vpc-cni-k8s v1.5 with Calico v3.8.1
- Others:
| kind/bug,sig/network,sig/node,help wanted,triage/needs-information,needs-triage | high | Critical |
533,595,372 | rust | Warn if type-name in parameter-position | I was surprised that one can confuse himself like in the cooked-up example below: Specifying a concrete type for a parameter on an `impl` does what one expects. This might lead to the expectation that you can specify a concrete type for a parameter on an `fn`, which is not the case:
```rust
struct Foobar<T>(T);
trait Foo {
fn bar<H>(&self, b: H) where H: std::fmt::Display ;
}
// `String` is a type...
impl Foo for Foobar<String> {
// `String` is not a type, it's a type-parameter named `String`.
// This is almost guaranteed to be wrong!
fn bar<String>(&self, _b: String) where String: std::fmt::Display {}
}
```
Here, `String` does not refer to the type `String` but is still a generic parameter, now named `String` instead of `H`.
If we don't specify the `Display`-bound, the compiler even helpfully suggests to add the requirement `String: Display`, which we can do right away. We can use `fn bar<u32>(...)` and the compiler will warn that `u32` should be upper-cased.
Maybe the compiler could warn if a parameter is given the name of a type that is in scope? Something to the tune of
> The type-parameter named `String` in `fn bar<String>(...)` has the same name as the type `std::string::String` in the local scope. Be aware that this declares a type-parameter named `String`. It does not specify a concrete type and might lead to confusion. The original definition is `fn bar<H>(...)` | C-enhancement,A-lints,A-diagnostics,T-lang,T-compiler | low | Minor |
533,610,852 | rust | "undefined symbol" when loading dylib compiled on >1.36.0 | `undefined symbol: _ZN42_$LT$$RF$T$u20$as$u20$core..fmt..Debug$GT$3fmt17hd0ba07f053fd4c7fE`
My project structure:
```
parent
main.rs # binary, loads libchild.so via libloading
lib.rs # dylib, shared code
child
lib.rs # dylib, dynamically linked to libparent.so
```
This compiles fine, but fails to load `libchild.so` when compiled on anything newer than 1.36.0. If the `parent` library is made non-dylib, it works fine, however this is not a viable solution for my project as it causes `libchild.so` to contain all the shared code, and my project has many `libchild.so`-equivalent crates.
Demo case: https://github.com/GinjaNinja32/rust-issues/tree/undefined-symbol-core-fmt-debug (has 70 dependency crates, mostly indirect; I'm not sure exactly where it goes wrong or how to make it smaller)
Bisected to merge dbebcee8d07b77eae3725988879001e6205c6e47, #59752.
```
rustc 1.39.0 (4560ea788 2019-11-04)
binary: rustc
commit-hash: 4560ea788cb760f0a34127156c78e2552949f734
commit-date: 2019-11-04
host: x86_64-unknown-linux-gnu
release: 1.39.0
LLVM version: 9.0
``` | A-linkage,E-needs-test,T-compiler,C-bug | low | Critical |
533,611,864 | create-react-app | Since updating to 3.3.0, running out of memory | ### Describe the bug
After upgrading to 3.3.0 and TS 3.7.3 (was previously on 3.2.0 and TS 3.7.2), I am unable to run `yarn start`
I receive this error: `FATAL ERROR: Ineffective mark-compacts near heap limit Allocation failed - JavaScript heap out of memory`
Reverting to react-scripts v3.2.0 resolves this issue - Typescript being at 3.7.3 or 3.7.2. has no effect, so I've narrowed it to react-scripts.
### Did you try recovering your dependencies?
Yes, blew away node_modules and re-ran `yarn`, tried without a lock file, etc. No dice.
### Environment
Windows 10, VSCode (also occurs on my macbook) | issue: bug,issue: needs investigation | high | Critical |
533,625,867 | flutter | Repository disappears from context after a call to google_sign_in package | If I comment out this code from `LoginEvent` then it works.
```dart
GoogleSignInAccount _account = await _googleSignIn.signIn();
GoogleSignInAuthentication _auth = await _account.authentication;
String _token = _auth.accessToken;
```
Also If I create a new variable before `_googleSignIn.signIn()` which has assigned `getStorage` result it works either.
**Valid `LoginEvent` class:**
```
class LoginEvent extends AuthEvent {
BuildContext context;
LoginEvent({@required this.context});
Future<AuthState> run() async {
var _storage = getStorage(context: context);
GoogleSignInAccount _account = await _googleSignIn.signIn();
GoogleSignInAuthentication _auth = await _account.authentication;
String _token = _auth.accessToken;
try {
await _storage.save(_token);
} on FlutterError catch (error) {
return AuthState(
isAuthorized: false,
inProgress: false,
isLoaded: true,
errorMessage: error.message.toString(),
);
}
return AuthState(
isAuthorized: _token != null && _token.isNotEmpty,
inProgress: false,
isLoaded: true,
errorMessage: null,
);
}
}
```
It looks like the `context` rebuilds after `_googleSignIn` call.
I have these additional packages in my `pubspec.yaml`:
```
google_sign_in: ^4.1.4
flutter_secure_storage: ^3.3.1+1
flutter_bloc: ^3.2.0
```
**Full Error message:**
```
I/flutter (20947): Looking up a deactivated widget's ancestor is unsafe.
I/flutter (20947): At this point the state of the widget's element tree is no longer stable.
I/flutter (20947): To safely refer to a widget's ancestor in its dispose() method, save a reference to the ancestor by calling dependOnInheritedWidgetOfExactType() in the widget's didChangeDependencies() method.
E/AccessibilityBridge(20947): VirtualView node must not be the root node.
```
**To reproduce the problem you have to click "Login" button and try to login by Google account.**
<details>
<summary>Invalid code sample</summary>
```dart
import 'package:flutter/material.dart';
import 'package:bloc/bloc.dart';
import 'package:flutter_bloc/flutter_bloc.dart';
import 'package:flutter_secure_storage/flutter_secure_storage.dart';
import 'package:google_sign_in/google_sign_in.dart';
GoogleSignIn _googleSignIn = GoogleSignIn(
scopes: [
'email',
],
);
void main() {
return runApp(MyApp());
}
class MyApp extends StatelessWidget {
@override
Widget build(BuildContext context) {
return RepositoryProvider<UserStorage>(
create: (context) => UserStorage(),
child: BlocProvider<AuthBloc>(
create: (context) => AuthBloc(context: context),
child: BlocBuilder<AuthBloc, AuthState>(
builder: (BuildContext context, AuthState state) {
return MaterialApp(
title: 'Test App',
theme: ThemeData(
primarySwatch: Colors.blue,
),
home: homePage(state),
builder: (BuildContext context, Widget widget) {
if (state.errorMessage != null &&
state.errorMessage.isNotEmpty) {
print(state.errorMessage);
return buildError(context, state.errorMessage);
}
return widget;
},
);
},
),
),
);
}
Widget homePage(AuthState state) {
if (!state.isLoaded) {
return CircularProgressIndicator();
}
if (state.isAuthorized && !state.inProgress) {
return MyHomePage(state.isAuthorized);
} else if (!state.isAuthorized && state.inProgress) {
return Center(
child: CircularProgressIndicator(),
);
} else {
return MyHomePage(state.isAuthorized);
}
}
}
class MyHomePage extends StatelessWidget {
final bool isLoggedIn;
MyHomePage(this.isLoggedIn, {Key key}) : super(key: key);
@override
Widget build(BuildContext context) {
return Scaffold(
body: Center(
child: Column(
mainAxisAlignment: MainAxisAlignment.center,
children: <Widget>[
FlatButton(
color: Colors.blue,
onPressed: () => isLoggedIn
? context.bloc<AuthBloc>().onLogout(context: context)
: context.bloc<AuthBloc>().onLogin(context: context),
child: Text(isLoggedIn ? 'Logout' : 'Login'),
),
Text(
isLoggedIn ? 'Loggedin' : 'Not loggedin',
),
],
),
),
);
}
}
Widget buildError(BuildContext context, String errorMessage) {
return Scaffold(
backgroundColor: Colors.indigo,
body: Padding(
padding: EdgeInsets.all(20.0),
child: ListView(
children: <Widget>[
Icon(Icons.mood_bad, color: Colors.white, size: 80.0),
SizedBox(height: 20.0),
Text(
errorMessage,
style: TextStyle(color: Colors.white),
),
],
),
),
);
}
class AuthState {
final bool inProgress;
final bool isAuthorized;
final bool isLoaded;
final String errorMessage;
const AuthState({
@required this.inProgress,
@required this.isAuthorized,
@required this.isLoaded,
@required this.errorMessage,
});
factory AuthState.initial() => AuthState(
inProgress: false,
isAuthorized: false,
isLoaded: false,
errorMessage: null,
);
}
class AuthBloc extends Bloc<AuthEvent, AuthState> {
@override
AuthState get initialState => AuthState.initial();
AuthBloc({@required BuildContext context}) {
_onCreate(context: context);
}
void _onCreate({@required BuildContext context}) => add(
CreateEvent(context: context),
);
void onLogin({@required BuildContext context}) => add(
LoginEvent(context: context),
);
void onLogout({@required BuildContext context}) => add(
LogoutEvent(context: context),
);
@override
Stream<AuthState> mapEventToState(AuthEvent event) async* {
if (event is CreateEvent) {
yield await event.run();
} else if (event is LoginEvent) {
yield AuthState(
inProgress: true,
isAuthorized: false,
isLoaded: true,
errorMessage: null,
);
yield await event.run();
} else if (event is LogoutEvent) {
yield AuthState(
inProgress: true,
isAuthorized: false,
isLoaded: true,
errorMessage: null,
);
yield await event.run();
}
}
}
abstract class AuthEvent {
Future<AuthState> run();
UserStorage getStorage({@required BuildContext context}) {
return context.repository<UserStorage>();
}
}
class CreateEvent extends AuthEvent {
BuildContext context;
CreateEvent({@required this.context});
Future<AuthState> run() async {
String _token;
_token = await getStorage(context: context).token;
return AuthState(
isAuthorized: _token != null && _token.isNotEmpty,
inProgress: false,
isLoaded: true,
errorMessage: null,
);
}
}
class LoginEvent extends AuthEvent {
BuildContext context;
LoginEvent({@required this.context});
Future<AuthState> run() async {
GoogleSignInAccount _account = await _googleSignIn.signIn();
GoogleSignInAuthentication _auth = await _account.authentication;
String _token = _auth.accessToken;
try {
await getStorage(context: context).save(_token);
} on FlutterError catch (error) {
return AuthState(
isAuthorized: false,
inProgress: false,
isLoaded: true,
errorMessage: error.message.toString(),
);
}
return AuthState(
isAuthorized: _token != null && _token.isNotEmpty,
inProgress: false,
isLoaded: true,
errorMessage: null,
);
}
}
class LogoutEvent extends AuthEvent {
final BuildContext context;
LogoutEvent({@required this.context});
Future<AuthState> run() async {
while (Navigator.of(context).canPop()) {
Navigator.of(context).pop();
}
Navigator.of(context).pushNamedAndRemoveUntil(
Navigator.defaultRouteName, (Route<dynamic> route) => false);
await _googleSignIn.signOut();
getStorage(context: context).remove();
return AuthState(
inProgress: false,
isAuthorized: false,
isLoaded: true,
errorMessage: null,
);
}
}
class UserStorage {
static const _token_key = 'token';
Future<String> get token async {
final _storage = FlutterSecureStorage();
return await _storage.read(key: _token_key);
}
Future<void> save(String token) async {
final _storage = FlutterSecureStorage();
await _storage.write(key: _token_key, value: token);
}
Future<void> remove() async {
final _storage = FlutterSecureStorage();
await _storage.deleteAll();
}
}
```
</details>
With `flutter_facebook_login` (v. 3.0.0) I have the same problem. It seems to me that there is something wrong with context after launching WebView or similar things.
<details>
<summary>flutter doctor -v</summary>
> [√] Flutter (Channel stable, v1.12.13+hotfix.8, on Microsoft Windows [Version 10.0.19559.1000], locale pl-PL)
> • Flutter version 1.12.13+hotfix.8 at C:\src\flutter
> • Framework revision 0b8abb4724 (2 weeks ago), 2020-02-11 11:44:36 -0800
> • Engine revision e1e6ced81d
> • Dart version 2.7.0
>
>
> [√] Android toolchain - develop for Android devices (Android SDK version 29.0.2)
> • Android SDK at C:\Users\owcza\AppData\Local\Android\sdk
> • Android NDK location not configured (optional; useful for native profiling support)
> • Platform android-29, build-tools 29.0.2
> • Java binary at: C:\Program Files\Android\Android Studio\jre\bin\java
> • Java version OpenJDK Runtime Environment (build 1.8.0_212-release-1586-b04)
> • All Android licenses accepted.
>
> [√] Android Studio (version 3.6)
> • Android Studio at C:\Program Files\Android\Android Studio
> • Flutter plugin version 43.0.2
> • Dart plugin version 192.7761
> • Java version OpenJDK Runtime Environment (build 1.8.0_212-release-1586-b04)
>
> [√] Connected device (1 available)
> • LG H930 • LGH930106dd0e5 • android-arm64 • Android 9 (API 28)
>
> • No issues found!
</details>
[flutter run --verbose with 3rd party packages.txt](https://github.com/flutter/flutter/files/4257859/flutter.run.--verbose.with.3rd.party.packages.txt) | platform-android,a: quality,p: google_sign_in,package,P2,team-android,triaged-android | low | Critical |
533,653,751 | flutter | Validate whether performance of fastReassemble can be improved with location objects | The fastReassemble method is currently (via https://github.com/flutter/flutter/pull/45649) implemented by walking the Element tree and comparing the runtimeType of the each Element's widget and the string value it is provided in. Besides the fact that we walk the entire tree O(n), runtime type is also fairly slow.
We might be able to improve the performance of this method by instead using the _Location objects created by the track-widget-creation transformer. This gives each widget constructor a `_location get location` field which contains information about the source location where the widget is constructed: https://github.com/flutter/flutter/blob/master/packages/flutter/lib/src/widgets/widget_inspector.dart#L2840
If fastReassemble method can provide the location for the particular widget that is updated, and the runtime could track elements by the location and provide O(1) | tool,c: proposal,P3,team-tool,triaged-tool | low | Major |
533,653,965 | flutter | Integration with Fuchsia introspection tools | We should make sure that any platform-specific developer introspection tools expose details of the internals of the running applications. For example, a memory viewer should not just report the Flutter engine heap size, it should expose the memory usage of individual isolates, or even lower-level, exposing Dart objects on the Dart heap to the heap viewer of the system-level memory tools. Similarly for CPU, battery usage, etc.
This is a very vague bug, intended to capture a medium-term goal. We should file sub-bugs to track individual specific features that need to be exposed on specific tools.
cc @iskakaushik @cbracken @eseidelGoogle | c: new feature,engine,c: performance,platform-fuchsia,P2,team-engine,triaged-engine | low | Critical |
533,655,803 | flutter | Enable incremental installation by default on iOS | Per discussion in https://github.com/FirebaseExtended/flutterfire/issues/349, we can improve build times for `flutter create` devs by enabling incremental installation in the template Podfile.
```
install! 'cocoapods',
generate_multiple_pod_projects: true,
incremental_installation: true
```
Benefits:
- Subsequent builds are much faster when using plugins like cloud_firestore.
Drawbacks:
- Requires CocoaPods 1.7.0 (if this is an issue we could use cocoapods-binary but then that would have to be installed as well)
- Possibly slightly longer initial build times | platform-ios,tool,P3,a: plugins,team-ios,triaged-ios | low | Major |
533,666,543 | godot | Godot script editor does not respect Mac keyboard bindings | **Godot version:** 312
**OS/device including version:** macOS 10.14.6
**Issue description:**
Mac allows users to override key bindings for customization. One such popular customization is to remap the home and end keys to behave like every other OS. For example
`~/Library/KeyBindings/DefaultKeyBinding.dict`
```
{
/* Remap Home / End */
"\UF729" = "moveToBeginningOfLine:"; /* Home */
"\UF72B" = "moveToEndOfLine:"; /* End */
"$\UF729" = "moveToBeginningOfLineAndModifySelection:"; /* Shift + Home */
"$\UF72B" = "moveToEndOfLineAndModifySelection:"; /* Shift + End */
}
```
**Steps to reproduce:**
Use this custom Mac keybinding file (note you have to log out and back in) and try to use the home and end key in Godot. Godot will treat home and end as if it were the Mac default.
**Minimal reproduction project:**
N/A | bug,platform:macos,topic:editor | low | Minor |
533,710,593 | rust | Unit-like structs in const generic arguments | ```rust
#![feature(const_generics)]
#[derive(PartialEq, Eq)]
struct Unit;
trait Const<const U: Unit> {}
impl Const<Unit> for () {} // wrong number of const arguments
```
`Const<{Unit}>` works. See also #66615.
<!-- TRIAGEBOT_START -->
<!-- TRIAGEBOT_ASSIGN_START -->
<!-- TRIAGEBOT_ASSIGN_DATA_START$${"user":null}$$TRIAGEBOT_ASSIGN_DATA_END -->
<!-- TRIAGEBOT_ASSIGN_END -->
<!-- TRIAGEBOT_END --> | T-compiler,C-feature-request,C-bug,A-const-generics | low | Minor |
533,764,915 | neovim | allow options private to a window (not copied with the buffer) | <!-- Before reporting: search existing issues and check the FAQ. -->
- `nvim --version`: NVIM v0.4.3
- `vim -u DEFAULTS` (version: ) behaves differently?
- Operating system/version: MacOS
- Terminal name/version:
- `$TERM`: xterm-256color
### Steps to reproduce using `nvim -u NORC`
Suppose there are two files test.vim and test2.vim in the same directory:
test.vim
```vim
let id1 = nvim_open_win(bufadd('test.vim'), 1, {'relative':'win', 'width':20, 'height':5, 'row':15, 'col':15})
let id2 = nvim_open_win(bufadd('test2.vim'), 0, {'relative':'win', 'width':20, 'height':5, 'row':5, 'col':5})
echom id1 id2
call nvim_win_set_option(id2, 'cursorline', v:true)
call nvim_win_set_option(id2, 'winhighlight', 'Normal:Error')
call nvim_win_close(id1, 1)
call nvim_win_close(id2, 1)
edit test2.vim
```
test2.vim, the content can be anything.
```vim
let id1 = nvim_open_win(bufadd('test.vim'), 1, {'relative':'win', 'width':20, 'height':5, 'row':15, 'col':15})
let id2 = nvim_open_win(bufadd('test2.vim'), 0, {'relative':'win', 'width':20, 'height':5, 'row':5, 'col':5})
echom id1 id2
call nvim_win_set_option(id2, 'cursorline', v:true)
call nvim_win_set_option(id2, 'winhighlight', 'Normal:Error')
call nvim_win_close(id1, 1)
call nvim_win_close(id2, 1)
edit test2.vim
```
steps:
1. nvim -u NORC test.vim
2. `so %`
### Actual behaviour
1. the cursorline appears
2. the color of the window is the same as `Error`
### Expected behaviour
1. no cursorline
2. the color of the window does not change
| enhancement,api,core,options | medium | Critical |
533,814,667 | ant-design | Instead of false, use value `null` to stand for feature disabling | - [x] I have searched the [issues](https://github.com/ant-design/ant-design/issues) of this repository and believe that this is not a duplicate.
### What problem does this feature solve?
When using Table, people set rowSelection to `null` to disable the whole rowSelection feature, but if they want to disable pagination, they should set it to `false`, which is really confusing
Sometimes `false` itself can be a meaningful boolean value, but generally `null` won't
Also, in OpenAPI@3, fields can keep a [`nullable`](https://github.com/OAI/OpenAPI-Specification/blob/master/versions/3.0.2.md#schemaNullable) property to specify that the value may totally be `null`, so I think `null` is more clear for this situation
What's more, only [nullish](https://developer.mozilla.org/en-US/docs/Glossary/nullish) value will fail in optional chaining and nullish coalescing operator, if still using `false`, maybe there will be some unexpected bugs someday 🤔
### What does the proposed API look like?
- rowSelection: null
- pagination: null
<!-- generated by ant-design-issue-helper. DO NOT REMOVE --> | Inactive | low | Critical |
533,816,807 | pytorch | [TensorBoard] Installed tensorboard 1.14.0, TestTensorBoardWriter.test_writer failed. | ## 🐛 Bug
PyTorch supported TensorBoard 1.14.0, but TensorBoard 1.14.0 not support tensorboard.plugins.mesh.metadata.get_components_bitmask.
PyTorch 1.3.1 + TensorBoard 1.14.0, failed TestTensorBoardWriter.test_writer.
Because tensorboard.plugins.mesh.metadata.get_components_bitmask is supported form tensorboard 1.15.0 .
If you are not particular about supporting Tensorboard 1.14.0, would you like to update ImportError message?
original code https://github.com/pytorch/pytorch/blob/v1.3.1/torch/utils/tensorboard/__init__.py#L4-L5
```diff
except ImportError:
raise ImportError('TensorBoard logging requires TensorBoard with Python summary writer installed. '
- 'This should be available in 1.14 or above.')
+ 'This should be available in 1.15 or above.')
from .writer import FileWriter, SummaryWriter # noqa F401
```
## To Reproduce
1. install tensorboard 1.14.0.
```bash
pip install tensorboard==1.14.0
```
1. Run test.
```bash
python3 test/test_tensorboard.py -- TestTensorBoardWriter.test_writer
```
## Expected behavior
```bash
root@6ed1da09866b:/workspace/pytorch# pip install tensorboard==1.14.0
Collecting tensorboard==1.14.0
Using cached https://files.pythonhosted.org/packages/91/2d/2ed263449a078cd9c8a9ba50ebd50123adf1f8cfbea1492f9084169b89d9/tensorboard-1.14.0-py3-none-any.whl
...
Successfully installed tensorboard-1.14.0
root@6ed1da09866b:/workspace/pytorch# python3 test/test_tensorboard.py -- TestTensorBoardWriter.test_writer
/opt/conda/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint8 = np.dtype([("qint8", np.int8, 1)])
/opt/conda/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint8 = np.dtype([("quint8", np.uint8, 1)])
/opt/conda/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint16 = np.dtype([("qint16", np.int16, 1)])
/opt/conda/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint16 = np.dtype([("quint16", np.uint16, 1)])
/opt/conda/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint32 = np.dtype([("qint32", np.int32, 1)])
/opt/conda/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
np_resource = np.dtype([("resource", np.ubyte, 1)])
Fail to import hypothesis in common_utils, tests are not derandomized
DEBUG>> SummaryWriter output: d5562a97-7fb8-4a42-8633-4c001578b24b
E
======================================================================
ERROR: test_writer (__main__.TestTensorBoardWriter)
----------------------------------------------------------------------
Traceback (most recent call last):
File "test/test_tensorboard.py", line 251, in test_writer
writer.add_mesh('my_mesh', vertices=v, colors=c, faces=f)
File "/workspace/pytorch/torch/utils/tensorboard/writer.py", line 948, in add_mesh
self._get_file_writer().add_summary(mesh(tag, vertices, colors, faces, config_dict), global_step, walltime)
File "/workspace/pytorch/torch/utils/tensorboard/summary.py", line 664, in mesh
components = metadata.get_components_bitmask([
AttributeError: module 'tensorboard.plugins.mesh.metadata' has no attribute 'get_components_bitmask'
----------------------------------------------------------------------
Ran 1 test in 0.641s
FAILED (errors=1)
```
## Environment
```
PyTorch version: 1.3.0a0+ee77ccb
Is debug build: No
CUDA used to build PyTorch: 10.1.243
OS: Ubuntu 16.04.6 LTS
GCC version: (Ubuntu 5.4.0-6ubuntu1~16.04.11) 5.4.0 20160609
CMake version: version 3.5.1
Python version: 3.6
Is CUDA available: Yes
CUDA runtime version: 10.1.243
GPU models and configuration:
GPU 0: GeForce RTX 2080 Ti
GPU 1: GeForce RTX 2080 Ti
Nvidia driver version: 418.67
cuDNN version: /usr/lib/x86_64-linux-gnu/libcudnn.so.7.6.3
Versions of relevant libraries:
[pip] numpy==1.17.2
[pip] torch==1.3.0a0+ee77ccb
[pip] torchvision==0.4.1a0+d94043a
[conda] blas 1.0 mkl
[conda] mkl 2019.4 243
[conda] mkl-service 2.3.0 py36he904b0f_0
[conda] mkl_fft 1.0.14 py36ha843d7b_0
[conda] mkl_random 1.1.0 py36hd6b4f25_0
[conda] torch 1.3.0a0+ee77ccb dev_0 <develop>
[conda] torchvision 0.4.1 py36_cu101 pytorch
```
## Additional context
Add `tensorboard.plugins.mesh.metadata.get_components_bitmask` TensorBoard issue.
[Support undefined number of data points in the mesh. (#2373) ? tensorflow/tensorboard@5e5badc](https://github.com/tensorflow/tensorboard/commit/5e5badc666480d98d389f6ca46e0b4febc253df8)
| oncall: visualization | low | Critical |
533,834,578 | vscode | iPadOS: cannot move panes around (activity bar, panel) | Refs: #85990
We probably did not hook up drag and drop support for moving views around. | feature-request,web,safari,ios-ipados | low | Minor |
533,861,565 | terminal | Unusual 3rd-party keyboard layout issue; Can't type digits with Shift in Neovim in PowerShell? | <!--
🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
I ACKNOWLEDGE THE FOLLOWING BEFORE PROCEEDING:
1. If I delete this entire template and go my own path, the core team may close my issue without further explanation or engagement.
2. If I list multiple bugs/concerns in this one issue, the core team may close my issue without further explanation or engagement.
3. If I write an issue that has many duplicates, the core team may close my issue without further explanation or engagement (and without necessarily spending time to find the exact duplicate ID number).
4. If I leave the title incomplete when filing the issue, the core team may close my issue without further explanation or engagement.
5. If I file something completely blank in the body, the core team may close my issue without further explanation or engagement.
All good? Then proceed!
-->
<!--
This bug tracker is monitored by Windows Terminal development team and other technical folks.
**Important: When reporting BSODs or security issues, DO NOT attach memory dumps, logs, or traces to Github issues**.
Instead, send dumps/traces to [email protected], referencing this GitHub issue.
If this is an application crash, please also provide a Feedback Hub submission link so we can find your diagnostic data on the backend. Use the category "Apps > Windows Terminal (Preview)" and choose "Share My Feedback" after submission to get the link.
Please use this form and describe your issue, concisely but precisely, with as much detail as possible.
-->
# Environment
```none
Windows build number: 10.0.19033.0
Windows Terminal version (if applicable): 0.7.3382.0 (but it's reproduceable in earlier builds too)
Neovim: 0.4.2
PowerShell Core: 7.0.0-preview.6
Universal keyboard layout: https://github.com/braindefender/universal-layout
```
# Steps to reproduce
1. Start Terminal with PowerShell Core.
2. Start Neovim.
3. Switch to insert mode (press i).
4. Type any digit by holding Shift and pressing a key with a digit on it. Like 3.
<!-- A description of how to trigger this bug. -->
# Expected behavior
<!-- A description of what you're expecting, possibly containing screenshots or reference material. -->
The digit is typed.
# Actual behavior
<!-- What's actually happening? -->
Nothing is typed.
# Additional information
I can type digits in WSL2 Ubuntu. The rest of the setup is the same.
I also can type digits in plain PowerShell Core, without Neovim.
Also, I can type digits in Neovim under PowerShell Core if I run it not in Windows Terminal, i.e. if I run the regular old school terminal.
Also, I can type digits in Vim instead of Neovim. | Area-Input,Issue-Bug,Area-TerminalControl,Product-Terminal | low | Critical |
533,873,922 | neovim | :execute "?search" with 'nostartofline' does not change cursor column | - `nvim --version`:
```
NVIM v0.5.0-dev
Build type: RelWithDebInfo
Lua 5.1
Compilation: /usr/bin/cc -g -O2 -fdebug-prefix-map=/build/neovim-vP8h0G/neovim-0.5.0+ubuntu1+git201912060004-ed42465-00e710e=. -fstack-protector-strong -Wformat -Werror=format-security -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=1 -O2 -g -Og -g -Wall -Wextra -pedantic -Wno-unused-parameter -Wstrict-prototypes -std=gnu99 -Wshadow -Wconversion -Wmissing-prototypes -Wimplicit-fallthrough -Wvla -fstack-protector-strong -fdiagnostics-color=auto -DINCLUDE_GENERATED_DECLARATIONS -D_GNU_SOURCE -DNVIM_MSGPACK_HAS_FLOAT32 -DNVIM_UNIBI_HAS_VAR_FROM -DMIN_LOG_LEVEL=3 -I/build/neovim-vP8h0G/neovim-0.5.0+ubuntu1+git201912060004-ed42465-00e710e/build/config -I/build/neovim-vP8h0G/neovim-0.5.0+ubuntu1+git201912060004-ed42465-00e710e/src -I/build/neovim-vP8h0G/neovim-0.5.0+ubuntu1+git201912060004-ed42465-00e710e/.deps/usr/include -I/usr/include -I/build/neovim-vP8h0G/neovim-0.5.0+ubuntu1+git201912060004-ed42465-00e710e/build/src/nvim/auto -I/build/neovim-vP8h0G/neovim-0.5.0+ubuntu1+git201912060004-ed42465-00e710e/build/include
Compiled by buildd@lgw01-amd64-018
Features: +acl +iconv +tui
See ":help feature-compile"
system vimrc file: "$VIM/sysinit.vim"
fall-back for $VIM: "/usr/share/nvim"
Run :checkhealth for more info
```
- `vim -u DEFAULTS` (version: ) behaves differently? Yes it does
- Operating system/version: Ubuntu 18.04
- Terminal name/version: xfce4-terminal
- `$TERM`: screen-256color (within tmux)
### Steps to reproduce using `nvim -u NORC`
```
printf "foobar\nbarfoo\nbazfoo\nbarbar\n" > test.txt; nvim -u NONE test.txt
# Inside vim :
# Place yourself at the end of the document, end of the line
G$
# Reverse search for baz. (Within execute because this is executed within say a plugin)
:execute "?baz"
```
### Actual behaviour
The cursor is placed on the right line (line 3), but the column has not changed
### Expected behaviour
The cursor should be placed on the right line, and on column 1 :
The behavior is correct for `vim`, and also when using `?baz` directly, without execute. | bug-vim,core | low | Critical |
533,892,687 | pytorch | save model definition file | ## 🚀 Feature
I noticed that torch.save() has already saved the model definition file. However, a model still needs to be defined in advance when loading. It is inconvenient to the user-defined model structure. Since the relative paths must be consistent.
## Motivation
I always keep model in one place and often loaded failed.
## Pitch
torch.load can save forward process.
## Alternatives
<!-- A clear and concise description of any alternative solutions or features you've considered, if any. -->
## Additional context
<!-- Add any other context or screenshots about the feature request here. -->
| feature,module: serialization,triaged | low | Critical |
533,894,906 | vscode | Trailing comma-aware move line up/down | Some files like JSON or pre-ES6 JavaScript files don't allow trailing commas, so moving last line in a block up or moving other line to the bottom causes these files to be corrupted and in a need of manual editing.
```json5
{
"someProperty": 1,
"otherProperty": 2,
"anotherProperty": 3 /* <-- My cursor is in this line */
}
```
Now, I press <kbd>option</kbd>+<kbd>↑</kbd> and...
## Current behavior
```json5
{
"someProperty": 1,
"anotherProperty": 3
"otherProperty": 2,
}
```
...oh no! two errors!
* Expected comma json(514) [4, 3]
* Trailing comma json(519) [4, 20]
## Expected behavior
```json5
{
"someProperty": 1,
"anotherProperty": 3,
"otherProperty": 2
}
```
Everything good! | feature-request,json,editor-autoindent | medium | Critical |
533,943,009 | gin | time_format tag not work with ShouldBindJSON? | ## Description
I have the param struct like this:
```
type Param1 struct {
Timestamp time.Time `json:"Timestamp" binding:"required" time_format:"2006-01-02 15:04:05"`
}
type Param2 struct {
Timestamp time.Time `json:"Timestamp" binding:"required" time_format:"unix"`
}
```
request json for Param1:
```
{
"Timestamp": "2001-11-11 11:11:11"
}
```
response error:
```
parsing time \"\"2001-11-11 11:11:11\"\" as \"\"2006-01-02T15:04:05Z07:00\"\": cannot parse \" 11:11:11\"\" as \"T\"]
```
request json for Param2:
```
{
"Timestamp": 1575528300
}
```
response error:
```
parsing time \"1575528300\" as \"\"2006-01-02T15:04:05Z07:00\"\": cannot parse \"1575528300\" as \"\"\"]
```
## Environment
- go version: 1.13
- gin version (or commit ref): 1.5.0
- operating system: macOS
| bug | low | Critical |
533,960,085 | TypeScript | JSDoc @type annotation ignored in chained variable declaration | I'm not sure whether this is a bug or a feature request:
**TypeScript Version:** Version 3.8.0-dev.20191206
also confirmed in 3.7.2
**Search Terms:** type ignored
**Code**
```javascript
var
/** @type {string} */
str1,
/** @type {number} */
num1;
/** @type {string} */
var str2;
/** @type {number} */
var num2;
str1 = "Hello";
num1 = "World"; // TSC does not complain => not OK
str2 = "Hello";
num2 = "World"; // TSC complains => OK
```
**Expected behavior:**
```
test.js:15:1 - error TS2322: Type '"World"' is not assignable to type 'number'.
15 num1 = "World";
test.js:18:1 - error TS2322: Type '"World"' is not assignable to type 'number'.
18 num2 = "World";
```
**Actual behavior:**
```
test.js:18:1 - error TS2322: Type '"World"' is not assignable to type 'number'.
18 num2 = "World";
```
**Playground Link:** Unable to reproduce any type related warning in JS PlayGround, please test locally
**Related Issues:** #22434, #13371 (somewhat related)
The official JSDoc documentation says nothing about mixing types within one `var` statement, but I guess that if the TS notation works, the behavior for JS might be the same.
```ts
var str1: string,
num1: number;
``` | Suggestion,Awaiting More Feedback | low | Critical |
533,965,604 | flutter | Flutter crashes with ProcessException on Windows 10 | I set all the paths correctly, accepted licenses, but `flutter.bat` outputs `The system cannot find the path specified.` twice during its execution and it finally crashes with a `ProcessException` immediately after the initializing Gradle message. A similar issue was reported @SushiRoll53 in #35391, and @mikhail-tokarev said that this is related to accepting licenses, closing the issue. However, this seems to be unrelated to the licenses. In my case, all licenses accepted (as seen in the log below), all SDKs up-to-date, and I'm still getting this error even in freshly created Flutter projects. This is a serious problem that prevents me from making any updates to my existing applications and keeping them updated. Please find the log below:
<!-- Thank you for using Flutter!
If you are looking for support, please check out our documentation
or consider asking a question on Stack Overflow:
* https://flutter.dev/
* https://api.flutter.dev/
* https://stackoverflow.com/questions/tagged/flutter?sort=frequent
If you have found a bug or if our documentation doesn't have an answer
to what you're looking for, then fill our the template below. Please read
our guide to filing a bug first: https://flutter.dev/docs/resources/bug-reports
-->
## Steps to Reproduce
<!-- Please tell us exactly how to reproduce the problem you are running into. -->
1. Create a fresh Flutter project.
2. Make any `flutter run` or `flutter build` command and get the error.
3. See the terminal output below:
```
D:\test-app\testapp>flutter build appbundle
The system cannot find the path specified.
The system cannot find the path specified.
Initializing gradle... Sending crash report to Google.
Crash report sent (report ID: e3186bb0f5870f65)
Initializing gradle... \Oops; flutter has exited unexpectedly.
Crash report written to D:\test-app\testapp\flutter_05.log;
please let us know at https://github.com/flutter/flutter/issues. -
```
## Logs
<!--
Include the full logs of the commands you are running between the lines
with the backticks below. If you are running any "flutter" commands,
please include the output of running them with "--verbose"; for example,
the output of running "flutter --verbose create foo".
-->
Flutter crash report; please file at https://github.com/flutter/flutter/issues.
## command
flutter build appbundle
## exception
ProcessException: ProcessException: Process "D:\test-app\testapp\android\gradlew.bat" exited abnormally:
------------------------------------------------------------
Gradle 4.10.2
------------------------------------------------------------
Build time: 2018-09-19 18:10:15 UTC
Revision: b4d8d5d170bb4ba516e88d7fe5647e2323d791dd
Kotlin DSL: 1.0-rc-6
Kotlin: 1.2.61
Groovy: 2.4.15
Ant: Apache Ant(TM) version 1.9.11 compiled on March 23 2018
JVM: 1.8.0_202-release (JetBrains s.r.o 25.202-b03)
OS: Windows 10 10.0 amd64
The system cannot find the path specified.
Command: D:\test-app\testapp\android\gradlew.bat -v
```
#0 runCheckedAsync (package:flutter_tools/src/base/process.dart:259:7)
<asynchronous suspension>
#1 _initializeGradle (package:flutter_tools/src/android/gradle.dart:300:9)
<asynchronous suspension>
#2 _ensureGradle (package:flutter_tools/src/android/gradle.dart:281:37)
<asynchronous suspension>
#3 buildGradleProject (package:flutter_tools/src/android/gradle.dart:484:31)
<asynchronous suspension>
#4 buildAppBundle (package:flutter_tools/src/android/app_bundle.dart:43:10)
<asynchronous suspension>
#5 BuildAppBundleCommand.runCommand (package:flutter_tools/src/commands/build_appbundle.dart:47:11)
<asynchronous suspension>
#6 FlutterCommand.verifyThenRunCommand (package:flutter_tools/src/runner/flutter_command.dart:490:18)
#7 _asyncThenWrapperHelper.<anonymous closure> (dart:async-patch/async_patch.dart:71:64)
#8 _rootRunUnary (dart:async/zone.dart:1132:38)
#9 _CustomZone.runUnary (dart:async/zone.dart:1029:19)
#10 _FutureListener.handleValue (dart:async/future_impl.dart:137:18)
#11 Future._propagateToListeners.handleValueCallback (dart:async/future_impl.dart:678:45)
#12 Future._propagateToListeners (dart:async/future_impl.dart:707:32)
#13 Future._completeWithValue (dart:async/future_impl.dart:522:5)
#14 Future._asyncComplete.<anonymous closure> (dart:async/future_impl.dart:552:7)
#15 _rootRun (dart:async/zone.dart:1124:13)
#16 _CustomZone.run (dart:async/zone.dart:1021:19)
#17 _CustomZone.runGuarded (dart:async/zone.dart:923:7)
#18 _CustomZone.bindCallbackGuarded.<anonymous closure> (dart:async/zone.dart:963:23)
#19 _microtaskLoop (dart:async/schedule_microtask.dart:41:21)
#20 _startMicrotaskLoop (dart:async/schedule_microtask.dart:50:5)
#21 _runPendingImmediateCallback (dart:isolate-patch/isolate_patch.dart:116:13)
#22 _RawReceivePortImpl._handleMessage (dart:isolate-patch/isolate_patch.dart:173:5)
```
## flutter doctor
```
[32m[✓][39m Flutter (Channel stable, v1.9.1+hotfix.6, on Windows, locale en-US)
[32m•[39m Flutter version 1.9.1+hotfix.6 at C:\flutter
[32m•[39m Framework revision 68587a0916 (3 months ago), 2019-09-13 19:46:58 -0700
[32m•[39m Engine revision b863200c37
[32m•[39m Dart version 2.5.0
[32m[✓][39m Android toolchain - develop for Android devices (Android SDK version 29.0.2)
[32m•[39m Android SDK at C:\android-sdk
[32m•[39m Android NDK location not configured (optional; useful for native profiling support)
[32m•[39m Platform android-29, build-tools 29.0.2
[32m•[39m ANDROID_HOME = C:\android-sdk
[32m•[39m Java binary at: C:\Program Files\Android\Android Studio\jre\bin\java
[32m•[39m Java version OpenJDK Runtime Environment (build 1.8.0_202-release-1483-b03)
[32m•[39m All Android licenses accepted.
[32m[✓][39m Android Studio (version 3.5)
[32m•[39m Android Studio at C:\Program Files\Android\Android Studio
[32m•[39m Flutter plugin version 29.0.2
[32m•[39m Dart plugin version 181.5616
[32m•[39m Java version OpenJDK Runtime Environment (build 1.8.0_202-release-1483-b03)
[32m[✓][39m VS Code (version 1.40.2)
[32m•[39m VS Code at C:\Users\ailabs\AppData\Local\Programs\Microsoft VS Code
[32m•[39m Flutter extension version 3.6.0
[32m[✓][39m Connected device (1 available)
[32m•[39m GM 8 • 7bf26878 • android-arm64 • Android 9 (API 28)
[32m•[39m No issues found!
``` | c: crash,tool,platform-windows,t: gradle,customer: crowd,P2,team-tool,triaged-tool | low | Critical |
533,995,729 | create-react-app | Suggestion: Add one line in CRA-originated index.html that disable Chrome's automatic translation | ### Is your proposal related to a problem?
A deployed CRA-originated app ran into crashes. The error text:
> react-dom.production.min.js:4260 DOMException: Failed to execute 'removeChild' on 'Node': The node to be removed is not a child of this node.
This is caused by Chrome's automatic translation feature. It seems it can modify a DOM node such that React's rendering process then fails. It is probably a fairly unusual error, but tricky to diagnose.
### Describe the solution you'd like
The suggestion is to add
<meta name=“google” content=“notranslate” />
which according to Googles support pages (and in practice) then disables the automatic translation feature, and in turn React's rendering does not crash.
### Describe alternatives you've considered
It is likely that React already tries at least somewhat to handle effects of browser plugins that change the DOM under the renderer process' feet, but maybe the Chrome translation is able to hide or somehow surprise any such protections. So there might be more interesting ways to solve this.
Other alternatives is to educate all users on how to disable this translation feature.
### Additional context
Only Chrome has been specifically investigated. There are likely other browsers that offer automatic translation that can cause similar issues.
| issue: proposal,needs triage | low | Critical |
534,037,826 | godot | Empty scene handling inconsistentcy | **Godot version:** v3.1.2.stable.official & v3.2.beta.custom_build.0aea0d0b3
**OS/device including version:** macOS Catalina 10.15.1
**Issue description:**
If there are two or more scenes open (whether empty or not), empty scenes are treated just like "normal" non-empty scenes, nothing special.
If the only open scene is an empty scene:
* Some commands (e.g. "New Scene") treat the empty scene as a normal scene
* Some other commands (e.g. "Open Scene") treat the empty scene as a placeholder: the newly opened scene will replace the empty scene; closing an empty scene does nothing...
I think this behavior is strange.
* If an empty scene is just a placeholder, then there should not be two of them open at the same time. Create a New Scene over an empty scene should do nothing.
* "Open Scene" over an empty scene behave differently when the only open scene is an empty scene, and when there are two scenes open.
IMHO, some clearer ways may be:
* Don't special treat empty scenes at all
* Treat empty scene as a placeholder: only one empty scene allowed, no matter how many scenes are currently open | enhancement,discussion,topic:editor | low | Minor |
534,054,197 | flutter | Rendering order inside Grid/ListView | <!-- Thank you for using Flutter!
If you are looking for support, please check out our documentation
or consider asking a question on Stack Overflow:
* https://flutter.dev/
* https://api.flutter.dev/
* https://stackoverflow.com/questions/tagged/flutter?sort=frequent
If you have found a bug or if our documentation doesn't have an answer
to what you're looking for, then fill our the template below. Please read
our guide to filing a bug first: https://flutter.dev/docs/resources/bug-reports
-->
## Use case
I'd like to be able to change the drawing order of children in a GridView. Right now they are drawn in the order they appear in the children list. If they overlap there is no way to control how they overlap.
<!--
Please tell us the problem you are running into that led to you wanting
a new feature.
Is your feature request related to a problem? Please give a clear and
concise description of what the problem is.
Describe alternative solutions you've considered. Is there a package
on pub.dev/flutter that already solves this?
-->
## Proposal
Add rendering order for child indexes, something like CSS's "z-index".
<!--
Briefly but precisely describe what you would like Flutter to be able to do.
Consider attaching images showing what you are imagining.
Does this have to be provided by Flutter directly, or can it be provided
by a package on pub.dev/flutter? If so, maybe consider implementing and
publishing such a package rather than filing a bug.
-->
## Video
| c: new feature,framework,f: scrolling,P3,team-framework,triaged-framework | medium | Critical |
534,079,388 | go | cmd/compile: -d=checkptr doesn't detect invalid pointer fields in converted pointers-to-structs | <!-- Please answer these questions before submitting your issue. Thanks! -->
### What version of Go are you using (`go version`)?
<pre>
$ go version
go version devel +0915a19a11 Fri Dec 6 05:12:15 2019 +0000 linux/amd64
</pre>
### What operating system and processor architecture are you using (`go env`)?
<details><summary><code>go env</code> Output</summary><br><pre>
$ go env
GO111MODULE="on"
GOARCH="amd64"
GOBIN=""
GOCACHE="~/.cache/go-build"
GOENV="~/.config/go/env"
GOEXE=""
GOFLAGS=" -mod="
GOHOSTARCH="amd64"
GOHOSTOS="linux"
GOINSECURE=""
GONOPROXY=""
GONOSUMDB=""
GOOS="linux"
GOPATH="~/go"
GOPRIVATE=""
GOPROXY="https://proxy.golang.org,direct"
GOROOT="~/projects/go"
GOSUMDB="sum.golang.org"
GOTMPDIR=""
GOTOOLDIR="~/projects/go/pkg/tool/linux_amd64"
GCCGO="gccgo"
AR="ar"
CC="gcc"
CXX="g++"
CGO_ENABLED="1"
CGO_CFLAGS="-g -O2"
CGO_CPPFLAGS=""
CGO_CXXFLAGS="-g -O2"
CGO_FFLAGS="-g -O2"
CGO_LDFLAGS="-g -O2"
PKG_CONFIG="pkg-config"
GOGCCFLAGS="-fPIC -m64 -pthread -fmessage-length=0 -fdebug-prefix-map=/tmp/go-build392569888=/tmp/go-build -gno-record-gcc-switches"
</pre></details>
### What did you do?
```golang
package tmp
import (
"testing"
"unsafe"
)
func TestUnsafeCast(t *testing.T) {
var x uint32 = 0
p := (*uint64)(unsafe.Pointer(&x))
t.Log(*p)
}
func TestUnsafeStructCast(t *testing.T) {
var x uint32 = 0
type A struct {
p *uint32
}
type B struct {
p *uint64
}
v := &A{ptr:&x}
p := (*B)(unsafe.Pointer(v))
t.Log(*p.p)
}
```
### What did you expect to see?
Both `TestUnsafeCast` and `TestUnsafeStructCast` fail with `go test -gcflags=-d=checkptr`, since both use invalid `unsafe.Pointer` casts.
### What did you see instead?
Only the `TestUnsafeCast` fails. Pointer fields are not verified properly. | help wanted,NeedsInvestigation,compiler/runtime | medium | Critical |
534,140,314 | pytorch | ReduceLROnPlateau detects a plateau during a steady decrease after a spike | I'm not sure if this is best described as a bug report, feature request, or even documentation clarification. I'm describing an unintuitive and undesired behaviour, but not necessarily a wrong behaviour.
The name and documentation of `ReduceLROnPlateau` suggests that it will step the LR when the loss is no longer decreasing, but that's not how it behaves. It actually steps the LR when the loss is not below the best it has ever achieved. This causes problems if the loss spikes up during training then continues decreasing from that spike.
Here's a loss graph where this happens:
Everything after step ~250 is treated as a plateau by `ReduceLROnPlateau` because it's not below the lowest point, despite the fact that it's clearly trending down and not a plateau in the conventional sense of the word. In this example, this stalls the training because the LR is repeatedly decimated and it never gets a chance to get back below the lowest point. With the scheduler disabled, it keeps decreasing rapidly and quickly gets below the previous lowest point with no further spikes.
The documentation states:
> Reduce learning rate when a metric has stopped improving. Models often benefit from reducing the learning rate by a factor of 2-10 once learning stagnates. This scheduler reads a metrics quantity and if no improvement is seen for a ‘patience’ number of epochs, the learning rate is reduced.
I don't think the behaviour and the description match. The loss is clearly improving at the point when the LR is being reduced.
I'd like the plateau detection to be more sensible about not triggering when there's a downward trend. That's tricky, because a cycle of spike, gentle decrease, spike, gentle decrease should be detected as a plateau, but a one off case shouldn't. Perhaps one way to approach this would be to give `best` a maximum age, so that it's only ever looking at the most recent `n` steps. I wonder if `patience` would be a good default value of `n`?
cc @vincentqb | module: optimizer,triaged | low | Critical |
534,148,035 | flutter | Focus can be stolen from the current ModalRoute | This involves dialogs or any modal routes that maintain their state. If an event happens and they spawn an auto-focusing FocusNode, the route will steal focus from the rest of the application, sometimes even invisibly.
Example app:
```dart
import 'dart:async';
import 'package:flutter/foundation.dart';
import 'package:flutter/material.dart';
void main() {
// If testing on desktop:
debugDefaultTargetPlatformOverride = TargetPlatform.fuchsia;
runApp(MaterialApp(
home: HomePage(),
));
}
class HomePage extends StatefulWidget {
@override
_HomePageState createState() => _HomePageState();
}
class _HomePageState extends State<HomePage> {
Timer _timer;
void _stealSoon() {
setState(() {
_timer = Timer(Duration(seconds: 5), () {
setState(() {});
});
});
}
@override
Widget build(BuildContext context) {
return Scaffold(
body: Padding(
padding: const EdgeInsets.all(24.0),
child: SizedBox(
width: double.infinity,
child: Column(
children: <Widget>[
RaisedButton(
child: Text('Open dialog'),
onPressed: () {
showDialog(
context: context,
child: AlertDialog(
title: Text('Hi!'),
actions: <Widget>[
FlatButton(
autofocus: true,
child: Text('OK'),
onPressed: () {
Navigator.of(context).pop();
},
),
],
),
);
},
),
RaisedButton(
child: Text('Steal focus soon'),
onPressed: _timer == null ? _stealSoon : null,
),
if (_timer != null && !_timer.isActive)
TextField(
autofocus: true,
),
],
),
),
),
);
}
}
```
Clicking "Steal focus soon" and then "Open dialog" will show this issue. The topmost modal route should be the only thing that can request and maintain focus.
This can happen in an actual application. e.g. a dialog shows up, and the route below goes to a username+password login page because credentials expired. If the username field has autofocus, this will steal focus from the dialog. Worse, this can happen with an opaque route if another opaque route is on top of it, assuming the previous route has maintainState set.
cc @gspencergoog | framework,f: routes,f: focus,has reproducible steps,P2,found in release: 3.0,found in release: 3.1,team-framework,triaged-framework | low | Critical |
534,181,597 | flutter | Text and icons do not render correctly in CanvasKit on older Android devices | On older Android devices, with the latest Chrome browser installed, text and icons either do not render correctly, or not at all, when built with CanvasKit support:
Consider the test web app at http://mjohnsullivan.github.io/flutter_web_test
On a Pixel 2, it renders correctly:
On a Nexus 6, the text and icons do not render:
Here's the details on the Nexus 6:
Device: Nexus 6
OS: Android 7.1.1
Chrome version: 78.0.3904.108
Flutter: Channel master, v1.12.17-pre.80, on Mac OS X 10.14.6 18G1012, locale en-US | c: regression,framework,dependency: skia,a: typography,platform-web,e: OS-version specific,e: web_canvaskit,has reproducible steps,P2,found in release: 3.3,found in release: 3.4,team-web,triaged-web | low | Major |
534,191,649 | create-react-app | no-unused-expressions warning for optional chaining | <!--
Please note that your issue will be fixed much faster if you spend about
half an hour preparing it, including the exact reproduction steps and a demo.
If you're in a hurry or don't feel confident, it's fine to report bugs with
less details, but this makes it less likely they'll get fixed soon.
In either case, please use this template and fill in as many fields below as you can.
Note that we don't provide help for webpack questions after ejecting.
You can find webpack docs at https://webpack.js.org/.
-->
### Describe the bug
I just upgrade `react-scripts` to 3.3.0. I am able to use nullish coalescing, but when I attempt to use optional chaining, I can't compile as I get this error:
```
./src/components/Select/Select.tsx
Line 141:5: Expected an assignment or function call and instead saw an expression no-unused-expressions
Search for the keywords to learn more about each error.
```
The line in question is:
```ts
this.props.onChange?.(resultingValue);
```
### Did you try recovering your dependencies?
No, I don't have time at the moment. If this might help I'll try it.
### Which terms did you search for in User Guide?
I searched for optional chaining, no-unused-expressions, eslint optional chaining.
### Environment
```
npx: installed 91 in 30.467s
Environment Info:
System:
OS: Windows 10 10.0.17134
CPU: (8) x64 Intel(R) Core(TM) i7-6700 CPU @ 3.40GHz
Binaries:
Node: 12.2.0 - C:\Program Files\nodejs\node.EXE
Yarn: Not Found
npm: 6.9.0 - C:\Program Files\nodejs\npm.CMD
Browsers:
Edge: 42.17134.1098.0
Internet Explorer: 11.0.17134.1
npmPackages:
react: ^16.12.0 => 16.12.0
react-dom: ^16.12.0 => 16.12.0
react-scripts: 3.3.0 => 3.3.0
npmGlobalPackages:
create-react-app: Not Found
```
### Steps to reproduce
1. Use optional chaining.
2. Attempt to run `npm start`.
### Expected behavior
I expected the app to compile and for the code to function like:
```
this.props.onChange !== undefined && this.props.onChange(resultingValue);
```
### Actual behavior
The app failed to compile and gave the above message.
| tag: underlying tools,issue: bug report | high | Critical |
534,192,119 | PowerToys | [PowerRename] Allow Copy instead of Rename | In PowerRename, how about add a checkbox option to "Copy file instead of renaming"
Without the option being checked, the file is renamed:
file.txt is renamed to SomethingElse.txt and file.txt no longer exists
With the option checked, the file is copied:
file.txt is copied to SomethingElse.txt and and file.txt still exists.
Every other aspect of PowerRename works in the same exact manner. | Idea-Enhancement,Help Wanted,Product-PowerRename | low | Minor |
534,207,083 | rust | [rustc_codegen_ssa] Let function calls and definitions use OperandValue or custom enum instead of Bx::Value | >> For example `Bx::Value` is assumed to be able to contain a pair
>
> I know this is offtopic but I just wanted to note that pair creation is intentionally funnelled through methods with "immediate_or_packed_pair" in their name, and they're limited to function inputs/outputs, for both calls and definitions (including e.g returns).
The solution there is to have an enum (or reuse `OperandValue`) which is used for e.g. `call`, `invoke`, `ret` and function params, but nothing else.
_Originally posted by @eddyb in https://github.com/rust-lang/rust/pull/65881#issuecomment-562678719_
This is necessary to use cg_ssa with backends which can't represent pairs in `Bx::Value`. | C-enhancement,A-codegen,T-compiler,A-cranelift | low | Minor |
534,214,648 | vscode | [scss] references for all SASS functions/at-rules | Ref: https://github.com/microsoft/vscode/issues/85981
Related: https://github.com/microsoft/vscode/issues/86187
I want to link all of them to official sass documentation.
Make this data loaded from a JSON file and add `references` to all of them, so they show up in completion. With #86187 references would show up in hover as well. | feature-request,css-less-scss | low | Minor |
534,224,499 | pytorch | FATAL_ERROR "Failed to determine the source files for the regular expression backend" | Using
```
$ git describe --tags
v1.4.0a0-734-gf2a35db
```
I have set the following variables
```
$ export CUDA_HOME=/storage/users/mnaderan/cuda-10.1.168/
$ export NCCL_ROOT=/storage/users/mnaderan/nccl_2.4.8-1/
$ export CUDNN_INCLUDE_DIR=/storage/users/mnaderan/cudnn/include/
$ export CUDNN_LIB_DIR=/storage/users/mnaderan/cudnn/lib/
$ export CC=/storage/users/mnaderan/tools/gcc-6.1.0/bin/gcc
$ export CXX=/storage/users/mnaderan/tools/gcc-6.1.0/bin/g++
$ export Numa_INCLUDE_DIR=/storage/users/mnaderan/tools/numa-dev/include/
$ export Numa_LIBRARIES=/storage/users/mnaderan/tools/numa-dev/lib64/
```
However, the setup.py command fails with this error:
```
CMake Error at third_party/benchmark/CMakeLists.txt:231 (message):
Failed to determine the source files for the regular expression backend
```
This is a problem with cxx feature check according to that line
```
# C++ feature checks
# Determine the correct regular expression engine to use
cxx_feature_check(STD_REGEX)
cxx_feature_check(GNU_POSIX_REGEX)
cxx_feature_check(POSIX_REGEX)
if(NOT HAVE_STD_REGEX AND NOT HAVE_GNU_POSIX_REGEX AND NOT HAVE_POSIX_REGEX)
message(FATAL_ERROR "Failed to determine the source files for the regular expression backend")
endif()
```
I can not further narrow the issue since there is no verbosity for that error. How can I generate more info?
The complete build output is shown below
```
$ python setup.py install --user
Building wheel torch-1.4.0a0+f2a35db
-- Building version 1.4.0a0+f2a35db
cmake -DBUILD_PYTHON=True -DBUILD_TEST=True -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=/storage/users/mnaderan/pt/pytorch/torch -DCMAKE_PREFIX_PATH=/usr/lib/python2.7/site-packages -DCUDNN_INCLUDE_DIR=/storage/users/mnaderan/cudnn/include/ -DCUDNN_LIBRARY=/storage/users/mnaderan/cudnn/lib/ -DNUMPY_INCLUDE_DIR=/storage/users/mnaderan/.local/lib/python2.7/site-packages/numpy/core/include -DPYTHON_EXECUTABLE=/bin/python -DPYTHON_INCLUDE_DIR=/usr/include/python2.7 -DPYTHON_LIBRARY=/usr/lib64/libpython2.7.so.1.0 -DTORCH_BUILD_VERSION=1.4.0a0+f2a35db -DUSE_CUDA=True -DUSE_NUMPY=True /storage/users/mnaderan/pt/pytorch
-- The CXX compiler identification is GNU 6.1.0
-- The C compiler identification is GNU 6.1.0
-- Check for working CXX compiler: /storage/users/mnaderan/tools/gcc-6.1.0/bin/g++
-- Check for working CXX compiler: /storage/users/mnaderan/tools/gcc-6.1.0/bin/g++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Check for working C compiler: /storage/users/mnaderan/tools/gcc-6.1.0/bin/gcc
-- Check for working C compiler: /storage/users/mnaderan/tools/gcc-6.1.0/bin/gcc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Not forcing any particular BLAS to be found
-- Performing Test COMPILER_WORKS
-- Performing Test COMPILER_WORKS - Success
-- Performing Test SUPPORT_GLIBCXX_USE_C99
-- Performing Test SUPPORT_GLIBCXX_USE_C99 - Success
-- Performing Test CAFFE2_EXCEPTION_PTR_SUPPORTED
-- Performing Test CAFFE2_EXCEPTION_PTR_SUPPORTED - Success
-- std::exception_ptr is supported.
-- Performing Test CAFFE2_IS_NUMA_AVAILABLE
-- Performing Test CAFFE2_IS_NUMA_AVAILABLE - Failed
-- NUMA is not available
-- Performing Test CAFFE2_NEED_TO_TURN_OFF_DEPRECATION_WARNING
-- Performing Test CAFFE2_NEED_TO_TURN_OFF_DEPRECATION_WARNING - Failed
-- Turning off deprecation warning due to glog.
-- Performing Test CAFFE2_COMPILER_SUPPORTS_AVX2_EXTENSIONS
-- Performing Test CAFFE2_COMPILER_SUPPORTS_AVX2_EXTENSIONS - Success
-- Current compiler supports avx2 extension. Will build perfkernels.
-- Performing Test CAFFE2_COMPILER_SUPPORTS_AVX512_EXTENSIONS
-- Performing Test CAFFE2_COMPILER_SUPPORTS_AVX512_EXTENSIONS - Success
-- Current compiler supports avx512f extension. Will build fbgemm.
-- Performing Test COMPILER_SUPPORTS_HIDDEN_VISIBILITY
-- Performing Test COMPILER_SUPPORTS_HIDDEN_VISIBILITY - Success
-- Performing Test COMPILER_SUPPORTS_HIDDEN_INLINE_VISIBILITY
-- Performing Test COMPILER_SUPPORTS_HIDDEN_INLINE_VISIBILITY - Success
-- Performing Test COMPILER_SUPPORTS_RDYNAMIC
-- Performing Test COMPILER_SUPPORTS_RDYNAMIC - Success
-- Building using own protobuf under third_party per request.
-- Use custom protobuf build.
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE
-- Caffe2 protobuf include directory: $<BUILD_INTERFACE:/storage/users/mnaderan/pt/pytorch/third_party/protobuf/src>$<INSTALL_INTERFACE:include>
-- Trying to find preferred BLAS backend of choice: MKL
-- MKL_THREADING = OMP
-- Looking for sys/types.h
-- Looking for sys/types.h - found
-- Looking for stdint.h
-- Looking for stdint.h - found
-- Looking for stddef.h
-- Looking for stddef.h - found
-- Check size of void*
-- Check size of void* - done
-- MKL_THREADING = OMP
CMake Warning at cmake/Dependencies.cmake:141 (message):
MKL could not be found. Defaulting to Eigen
Call Stack (most recent call first):
CMakeLists.txt:378 (include)
CMake Warning at cmake/Dependencies.cmake:160 (message):
Preferred BLAS (MKL) cannot be found, now searching for a general BLAS
library
Call Stack (most recent call first):
CMakeLists.txt:378 (include)
-- MKL_THREADING = OMP
-- Checking for [mkl_intel_lp64 - mkl_gnu_thread - mkl_core - gomp - pthread - m - dl]
-- Library mkl_intel_lp64: not found
-- Checking for [mkl_intel_lp64 - mkl_intel_thread - mkl_core - gomp - pthread - m - dl]
-- Library mkl_intel_lp64: not found
-- Checking for [mkl_intel - mkl_gnu_thread - mkl_core - gomp - pthread - m - dl]
-- Library mkl_intel: not found
-- Checking for [mkl_intel - mkl_intel_thread - mkl_core - gomp - pthread - m - dl]
-- Library mkl_intel: not found
-- Checking for [mkl_gf_lp64 - mkl_gnu_thread - mkl_core - gomp - pthread - m - dl]
-- Library mkl_gf_lp64: not found
-- Checking for [mkl_gf_lp64 - mkl_intel_thread - mkl_core - gomp - pthread - m - dl]
-- Library mkl_gf_lp64: not found
-- Checking for [mkl_gf - mkl_gnu_thread - mkl_core - gomp - pthread - m - dl]
-- Library mkl_gf: not found
-- Checking for [mkl_gf - mkl_intel_thread - mkl_core - gomp - pthread - m - dl]
-- Library mkl_gf: not found
-- Checking for [mkl_intel_lp64 - mkl_gnu_thread - mkl_core - iomp5 - pthread - m - dl]
-- Library mkl_intel_lp64: not found
-- Checking for [mkl_intel_lp64 - mkl_intel_thread - mkl_core - iomp5 - pthread - m - dl]
-- Library mkl_intel_lp64: not found
-- Checking for [mkl_intel - mkl_gnu_thread - mkl_core - iomp5 - pthread - m - dl]
-- Library mkl_intel: not found
-- Checking for [mkl_intel - mkl_intel_thread - mkl_core - iomp5 - pthread - m - dl]
-- Library mkl_intel: not found
-- Checking for [mkl_gf_lp64 - mkl_gnu_thread - mkl_core - iomp5 - pthread - m - dl]
-- Library mkl_gf_lp64: not found
-- Checking for [mkl_gf_lp64 - mkl_intel_thread - mkl_core - iomp5 - pthread - m - dl]
-- Library mkl_gf_lp64: not found
-- Checking for [mkl_gf - mkl_gnu_thread - mkl_core - iomp5 - pthread - m - dl]
-- Library mkl_gf: not found
-- Checking for [mkl_gf - mkl_intel_thread - mkl_core - iomp5 - pthread - m - dl]
-- Library mkl_gf: not found
-- Checking for [mkl_intel_lp64 - mkl_gnu_thread - mkl_core - pthread - m - dl]
-- Library mkl_intel_lp64: not found
-- Checking for [mkl_intel_lp64 - mkl_intel_thread - mkl_core - pthread - m - dl]
-- Library mkl_intel_lp64: not found
-- Checking for [mkl_intel - mkl_gnu_thread - mkl_core - pthread - m - dl]
-- Library mkl_intel: not found
-- Checking for [mkl_intel - mkl_intel_thread - mkl_core - pthread - m - dl]
-- Library mkl_intel: not found
-- Checking for [mkl_gf_lp64 - mkl_gnu_thread - mkl_core - pthread - m - dl]
-- Library mkl_gf_lp64: not found
-- Checking for [mkl_gf_lp64 - mkl_intel_thread - mkl_core - pthread - m - dl]
-- Library mkl_gf_lp64: not found
-- Checking for [mkl_gf - mkl_gnu_thread - mkl_core - pthread - m - dl]
-- Library mkl_gf: not found
-- Checking for [mkl_gf - mkl_intel_thread - mkl_core - pthread - m - dl]
-- Library mkl_gf: not found
-- Checking for [mkl_intel_lp64 - mkl_sequential - mkl_core - m - dl]
-- Library mkl_intel_lp64: not found
-- Checking for [mkl_intel - mkl_sequential - mkl_core - m - dl]
-- Library mkl_intel: not found
-- Checking for [mkl_gf_lp64 - mkl_sequential - mkl_core - m - dl]
-- Library mkl_gf_lp64: not found
-- Checking for [mkl_gf - mkl_sequential - mkl_core - m - dl]
-- Library mkl_gf: not found
-- Checking for [mkl_intel_lp64 - mkl_core - gomp - pthread - m - dl]
-- Library mkl_intel_lp64: not found
-- Checking for [mkl_intel - mkl_core - gomp - pthread - m - dl]
-- Library mkl_intel: not found
-- Checking for [mkl_gf_lp64 - mkl_core - gomp - pthread - m - dl]
-- Library mkl_gf_lp64: not found
-- Checking for [mkl_gf - mkl_core - gomp - pthread - m - dl]
-- Library mkl_gf: not found
-- Checking for [mkl_intel_lp64 - mkl_core - iomp5 - pthread - m - dl]
-- Library mkl_intel_lp64: not found
-- Checking for [mkl_intel - mkl_core - iomp5 - pthread - m - dl]
-- Library mkl_intel: not found
-- Checking for [mkl_gf_lp64 - mkl_core - iomp5 - pthread - m - dl]
-- Library mkl_gf_lp64: not found
-- Checking for [mkl_gf - mkl_core - iomp5 - pthread - m - dl]
-- Library mkl_gf: not found
-- Checking for [mkl_intel_lp64 - mkl_core - pthread - m - dl]
-- Library mkl_intel_lp64: not found
-- Checking for [mkl_intel - mkl_core - pthread - m - dl]
-- Library mkl_intel: not found
-- Checking for [mkl_gf_lp64 - mkl_core - pthread - m - dl]
-- Library mkl_gf_lp64: not found
-- Checking for [mkl_gf - mkl_core - pthread - m - dl]
-- Library mkl_gf: not found
-- Checking for [mkl - guide - pthread - m]
-- Library mkl: not found
-- MKL library not found
-- Checking for [Accelerate]
-- Library Accelerate: BLAS_Accelerate_LIBRARY-NOTFOUND
-- Checking for [vecLib]
-- Library vecLib: BLAS_vecLib_LIBRARY-NOTFOUND
-- Checking for [openblas]
-- Library openblas: /usr/lib64/libopenblas.so
-- Looking for sgemm_
-- Looking for sgemm_ - found
-- Performing Test BLAS_F2C_DOUBLE_WORKS
-- Performing Test BLAS_F2C_DOUBLE_WORKS - Failed
-- Performing Test BLAS_F2C_FLOAT_WORKS
-- Performing Test BLAS_F2C_FLOAT_WORKS - Success
-- Performing Test BLAS_USE_CBLAS_DOT
-- Performing Test BLAS_USE_CBLAS_DOT - Success
-- Found a library with BLAS API (open).
-- The ASM compiler identification is GNU
-- Found assembler: /storage/users/mnaderan/tools/gcc-6.1.0/bin/gcc
-- Check if compiler accepts -pthread
-- Check if compiler accepts -pthread - yes
-- Brace yourself, we are building NNPACK
-- Performing Test NNPACK_ARCH_IS_X86_32
-- Performing Test NNPACK_ARCH_IS_X86_32 - Failed
-- Found PythonInterp: /bin/python (found version "2.7.5")
-- NNPACK backend is x86-64
-- Failed to find LLVM FileCheck
-- Found Git: /bin/git (found version "1.8.3.1")
-- git Version: v1.4.0-505be96a
-- Version: 1.4.0
-- Performing Test HAVE_CXX_FLAG_STD_CXX11
-- Performing Test HAVE_CXX_FLAG_STD_CXX11 - Success
-- Performing Test HAVE_CXX_FLAG_WALL
-- Performing Test HAVE_CXX_FLAG_WALL - Success
-- Performing Test HAVE_CXX_FLAG_WEXTRA
-- Performing Test HAVE_CXX_FLAG_WEXTRA - Success
-- Performing Test HAVE_CXX_FLAG_WSHADOW
-- Performing Test HAVE_CXX_FLAG_WSHADOW - Success
-- Performing Test HAVE_CXX_FLAG_WERROR
-- Performing Test HAVE_CXX_FLAG_WERROR - Success
-- Performing Test HAVE_CXX_FLAG_PEDANTIC
-- Performing Test HAVE_CXX_FLAG_PEDANTIC - Success
-- Performing Test HAVE_CXX_FLAG_PEDANTIC_ERRORS
-- Performing Test HAVE_CXX_FLAG_PEDANTIC_ERRORS - Success
-- Performing Test HAVE_CXX_FLAG_WSHORTEN_64_TO_32
-- Performing Test HAVE_CXX_FLAG_WSHORTEN_64_TO_32 - Failed
-- Performing Test HAVE_CXX_FLAG_WFLOAT_EQUAL
-- Performing Test HAVE_CXX_FLAG_WFLOAT_EQUAL - Success
-- Performing Test HAVE_CXX_FLAG_FSTRICT_ALIASING
-- Performing Test HAVE_CXX_FLAG_FSTRICT_ALIASING - Success
-- Performing Test HAVE_CXX_FLAG_WNO_DEPRECATED_DECLARATIONS
-- Performing Test HAVE_CXX_FLAG_WNO_DEPRECATED_DECLARATIONS - Success
-- Performing Test HAVE_CXX_FLAG_WSTRICT_ALIASING
-- Performing Test HAVE_CXX_FLAG_WSTRICT_ALIASING - Success
-- Performing Test HAVE_CXX_FLAG_WD654
-- Performing Test HAVE_CXX_FLAG_WD654 - Failed
-- Performing Test HAVE_CXX_FLAG_WTHREAD_SAFETY
-- Performing Test HAVE_CXX_FLAG_WTHREAD_SAFETY - Failed
-- Performing Test HAVE_CXX_FLAG_COVERAGE
-- Performing Test HAVE_CXX_FLAG_COVERAGE - Success
-- Performing Test HAVE_STD_REGEX
-- Performing Test HAVE_STD_REGEX
-- Performing Test HAVE_STD_REGEX -- compiled but failed to run
-- Performing Test HAVE_GNU_POSIX_REGEX
-- Performing Test HAVE_GNU_POSIX_REGEX
-- Performing Test HAVE_GNU_POSIX_REGEX -- failed to compile
-- Performing Test HAVE_POSIX_REGEX
-- Performing Test HAVE_POSIX_REGEX
-- Performing Test HAVE_POSIX_REGEX -- compiled but failed to run
CMake Error at third_party/benchmark/CMakeLists.txt:231 (message):
Failed to determine the source files for the regular expression backend
-- Configuring incomplete, errors occurred!
See also "/storage/users/mnaderan/pt/pytorch/build/CMakeFiles/CMakeOutput.log".
See also "/storage/users/mnaderan/pt/pytorch/build/CMakeFiles/CMakeError.log".
Traceback (most recent call last):
File "setup.py", line 751, in <module>
build_deps()
File "setup.py", line 310, in build_deps
cmake=cmake)
File "/storage/users/mnaderan/pt/pytorch/tools/build_pytorch_libs.py", line 56, in build_caffe2
rerun_cmake)
File "/storage/users/mnaderan/pt/pytorch/tools/setup_helpers/cmake.py", line 318, in generate
self.run(args, env=my_env)
File "/storage/users/mnaderan/pt/pytorch/tools/setup_helpers/cmake.py", line 142, in run
check_call(command, cwd=self.build_dir, env=env)
File "/usr/lib64/python2.7/subprocess.py", line 542, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['cmake', '-DBUILD_PYTHON=True', '-DBUILD_TEST=True', '-DCMAKE_BUILD_TYPE=Release', '-DCMAKE_INSTALL_PREFIX=/storage/users/mnaderan/pt/pytorch/torch', '-DCMAKE_PREFIX_PATH=/usr/lib/python2.7/site-packages', '-DCUDNN_INCLUDE_DIR=/storage/users/mnaderan/cudnn/include/', '-DCUDNN_LIBRARY=/storage/users/mnaderan/cudnn/lib/', '-DNUMPY_INCLUDE_DIR=/storage/users/mnaderan/.local/lib/python2.7/site-packages/numpy/core/include', '-DPYTHON_EXECUTABLE=/bin/python', '-DPYTHON_INCLUDE_DIR=/usr/include/python2.7', '-DPYTHON_LIBRARY=/usr/lib64/libpython2.7.so.1.0', '-DTORCH_BUILD_VERSION=1.4.0a0+f2a35db', '-DUSE_CUDA=True', '-DUSE_NUMPY=True', '/storage/users/mnaderan/pt/pytorch']' returned non-zero exit status 1
```
Thanks for any idea. | module: build,triaged,module: third_party | low | Critical |
534,232,362 | pytorch | Adding max_norm constraint to an Embedding layer leads to an error | ## Max_norm constraint results in strange behavior
Here is a simplified (and meaningless) network where this problem occurs:
```python
class DDImodel(nn.Module):
def __init__(self, max_separation = 155, position_embedding_size = 50, output_classes = 5):
super(DDImodel, self).__init__()
self.max_separation = max_separation
self.position_embedding = nn.Embedding(max_separation, position_embedding_size, max_norm=3.0)
self.hidden2class = nn.Linear(position_embedding_size, output_classes)
def forward(self, inputs):
batch_size = inputs.shape[0]
positions = torch.arange(0, self.max_separation).repeat(batch_size, 1)
out = self.position_embedding(positions)
out = torch.sum(out, dim=1)
out = self.hidden2class(out)
return F.log_softmax(out, dim=1)
```
Without max_norm, everything works fine. But when I switch it on, I get an error:
> RuntimeError: leaf variable has been moved into the graph interior
Interestingly, it depends on the batch size. With batch size = 6, it works, but breaks when it is 7 and higher. Also, it always works with CUDA, but it doesn't work on a CPU.
Also, you can see some strange results on the computation graphs. With batch size of 6 (first image), it looks sensible, but when it is 7 (second image), there is this strange long sequence of CopySlice blocks. The error occurs only when there is this long sequence.
Environment:
PyTorch version: 1.3.1
Is debug build: No
CUDA used to build PyTorch: 10.1.243
OS: Ubuntu 18.04.3 LTS
GCC version: (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0
CMake version: version 3.10.2
Python version: 3.6
Is CUDA available: No
CUDA runtime version: No CUDA
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
Versions of relevant libraries:
[pip3] msgpack-numpy==0.4.4.3
[pip3] numpy==1.17.3
[pip3] torch==1.3.1
[pip3] torchtext==0.4.0
[pip3] torchviz==0.0.1
[conda] Could not collect | module: nn,triaged | low | Critical |
534,233,995 | pytorch | CUDA error: initialization error (multiprocessing) with Python 3.7 | ## 🐛 Bug
The problem is similar to the one described in #2517 and #21092. However, the solutions mentioned in those issues do not work in this case.
## To Reproduce
Steps to reproduce the behavior:
1. Anaconda3 with Python 3.7
2. Pytorch 1.1.0
3. Run the code below to generate some processes in parallel with a single GPU
4. Use the code borrowed from @floringogianu in #2517 with a little change, as below:
```
import torch
import torch.multiprocessing as mp
try: # otherwise it complains: context has already been set
mp.set_start_method("spawn")
except: pass
def sub_processes(A, B, D, i, j, size):
D[(j * size):((j + 1) * size), i] = torch.mul(B[:, i], A[j, i])
def task(A, B):
size1 = A.shape
size2 = B.shape
D = torch.zeros([size1[0] * size2[0], size1[1]]).cuda()
D.share_memory_()
for i in range(1):
processes = []
for j in range(size1[0]):
p = mp.Process(target=sub_processes, args=(A, B, D, i, j, size2[0]))
p.start()
processes.append(p)
for p in processes:
p.join()
return D
if __name__ == "__main__":
A = torch.rand(3, 3).cuda()
B = torch.rand(3, 3).cuda()
C = task(A,B)
print(C)
```
## More detail about my setting:
- OS: Ubuntu 18.4
- CUDA: 10.2
- GPU: Titan X
## Workaround:
Downgrade Python to 3.6.9 and it works. The expected output is as follow:
```
tensor([[0.5072, 0.0000, 0.0000],
[0.4780, 0.0000, 0.0000],
[0.4990, 0.0000, 0.0000],
[0.3349, 0.0000, 0.0000],
[0.3156, 0.0000, 0.0000],
[0.3295, 0.0000, 0.0000],
[0.4351, 0.0000, 0.0000],
[0.4100, 0.0000, 0.0000],
[0.4280, 0.0000, 0.0000]], device='cuda:0')
```
So I hope someone could take a look into the reason as to why it does not work with Python 3.7 multiprocessing library.
cc @ezyang @gchanan @zou3519 | needs reproduction,module: multiprocessing,triaged | low | Critical |
534,238,070 | godot | Error when opening a custom dialog in response to a quit request | Godot 3.2 beta2
Windows 10 64 bits
I have a custom confirmation popup which works fine if I open it on a key press. But when I make it open in reaction to `NOTIFICATION_WM_QUIT`, this error prints in the console:
```
ERROR: Parent node is busy setting up children, move_child() failed. Consider using call_deferred("move_child") instead (or "popup" if this is from a popup).
At: scene/main/node.cpp:332
```
Test project:
[PopupBug.zip](https://github.com/godotengine/godot/files/3933825/PopupBug.zip)
Run, close window. This makes the dialog show up, error in console. You can press `Quit` to actually close the app. | bug,topic:core | low | Critical |
534,242,679 | pytorch | Don't ship protoc in wheels | Apparently there is a torch/bin/protoc in our whl release.
cc @ezyang | module: binaries,triaged | low | Minor |
534,244,996 | rust | Incorrect hints for `Fn` type on stable and incorrect suggestions on nightly | Code
```rust
fn call_fn<F: Fn() -> ()>(f: &F) {
f()
}
fn call_any<F: std::any::Any>(f: &F) {
call_fn(f)
}
fn main() {
call_any(&|| {});
}
```
currently produces on stable
```
error[E0277]: expected a `std::ops::Fn<()>` closure, found `F`
--> src/main.rs:6:13
|
1 | fn call_fn<F: Fn() -> ()>(f: &F) {
| ------- ---------- required by this bound in `call_fn`
...
6 | call_fn(f)
| ^ expected an `Fn<()>` closure, found `F`
|
= help: the trait `std::ops::Fn<()>` is not implemented for `F`
= note: wrap the `F` in a closure with no arguments: `|| { /* code */ }
= help: consider adding a `where F: std::ops::Fn<()>` bound
```
- `note: wrap the F in a closure with no arguments: || { /* code */ }` - ???
- `where F: std::ops::Fn<()>` - incorrect `Fn` syntax
On nightly situation is more interesting:
```
error[E0277]: expected a `std::ops::Fn<()>` closure, found `F`
--> src/main.rs:6:13
|
1 | fn call_fn<F: Fn() -> ()>(f: &F) {
| ------- ---------- required by this bound in `call_fn`
...
5 | fn call_any<F: std::any::Any>(f: &F) {
| -- help: consider further restricting this bound: `F: std::ops::Fn<()> +`
6 | call_fn(f)
| ^ expected an `Fn<()>` closure, found `F`
|
= help: the trait `std::ops::Fn<()>` is not implemented for `F`
= note: wrap the `F` in a closure with no arguments: `|| { /* code */ }
```
- ` -- help: consider further restricting this bound: F: std::ops::Fn<()> +` has missing type and incorrect `Fn` syntax (`Fn()` is correct)
([Playground](https://play.rust-lang.org/?version=stable&mode=debug&edition=2018&gist=1b1dcac0ec2206eeff63889ffd17d503)) | A-diagnostics,A-closures,T-compiler,C-bug,D-invalid-suggestion | low | Critical |
534,251,911 | flutter | Feature Request: Textfield allow override of key events before Textfield is changed. | Hello. It would be nice to be able to get some callbacks with the chance to change behavior before it is applied to the Textfield.
Here is an example.
A search box where the user presses the up/down arrow keys to select suggestions. The default behavior of the Textfield is to move the cursor to the beginning or end of the text field. Getting a callback with the key event before the state of the Textfield is updated and a chance to change what happens would be great.
The other option would be allow rawkeyboardlistener to swallow or forward events to its children based on a handler. Right now if I wrap a TextField inside a Rawkeyboardlistener they both receive the key events without discrimination.
I’m not sure what approach would be better. But either way it seems a sdk change is needed to make this happen.
| a: text input,c: new feature,framework,P3,team-framework,triaged-framework | low | Minor |
534,271,089 | go | net/http, x/net/http2: TCP connection is not closed after http client request times out after missing RST (TCP) or GOAWAY(HTTP/2) frame | ### What version of Go are you using (`go version`)?
<pre>
$ go version
go version go1.13.4 linux/amd64
</pre>
### Does this issue reproduce with the latest release?
I believe so.
### What operating system and processor architecture are you using (`go env`)?
<details><summary><code>go env</code> Output</summary><br><pre>
$ go env
GO111MODULE=""
GOARCH="amd64"
GOBIN=""
GOCACHE="/data/home/jminter/.cache/go-build"
GOENV="/data/home/jminter/.config/go/env"
GOEXE=""
GOFLAGS=""
GOHOSTARCH="amd64"
GOHOSTOS="linux"
GONOPROXY=""
GONOSUMDB=""
GOOS="linux"
GOPATH="/data/home/jminter/go"
GOPRIVATE=""
GOPROXY="direct"
GOROOT="/usr/lib/golang"
GOSUMDB="off"
GOTMPDIR=""
GOTOOLDIR="/usr/lib/golang/pkg/tool/linux_amd64"
GCCGO="gccgo"
AR="ar"
CC="gcc"
CXX="g++"
CGO_ENABLED="1"
GOMOD=""
CGO_CFLAGS="-g -O2"
CGO_CPPFLAGS=""
CGO_CXXFLAGS="-g -O2"
CGO_FFLAGS="-g -O2"
CGO_LDFLAGS="-g -O2"
PKG_CONFIG="pkg-config"
GOGCCFLAGS="-fPIC -m64 -pthread -fmessage-length=0 -fdebug-prefix-map=/tmp/go-build951550205=/tmp/go-build -gno-record-gcc-switches"
</pre></details>
### What did you do?
With a simple client like this, and using it to make a dumb GET request to some server:
c := &http.Client{
Timeout: 30 * time.Second,
}
If the server doesn't respond within 30 seconds, the client gets a timeout error as expected. But, Go does not close the underlying TCP connection to the server. If the client keeps retrying, unfortunately its subsequent requests get queued on the same TCP connection.
If the server doesn't close the connection, or perhaps if there's an underlying TCP connection (e.g. connectivity) break which means the client and server aren't actually communicating, the client doesn't redial to establish a new connection, at least perhaps until the local kernel socket send buffer is full, or there's a TCP keepalive timeout, or something (conjecture).
You can see this when it happens because netstat continues to show the connection, and you can see its local Recv-Q figure incrementing on each retry.
Setting DisableKeepAlives works around this problem, but it's a pretty big hammer.
IMO it would make more sense for Go to abandon and close the underlying TCP connection if an HTTP timeout has occured on it.
### What did you expect to see?
I'd like to see Go net/http close underlying TCP connections if the HTTP request they are running times out.
### What did you see instead?
It keeps the underlying TCP connection open. | NeedsInvestigation | medium | Critical |
534,271,865 | TypeScript | Add a visibility mechanism similar to `friend` or `InternalsVisibleTo` |
## Suggestion
A way to allow certain other modules to access properties or methods of a class.
## Use Cases
Sometimes, we want other modules (other code) to have access to certain properties or methods, but otherwise the end user not to have access to the properties/methods. This is labeled as "package protected" or similar terms in other languages.
## Examples
This is how it could work, with some sort of new non-runtime syntax:
```js
import SomeClass from './SomeClass'
export class Foo {
visible in SomeClass
private foo: number = 123
}
```
or
```js
import SomeClass from './SomeClass'
export class Foo {
private foo visible in SomeClass: number = 123
}
```
or maybe even as a special decorator, globally-declared, virtual decorator
```js
import SomeClass from './SomeClass'
export class Foo {
@visibleIn(SomeClass)
private foo: number = 123
}
```
`SomeClass` would then be able to access the private property/method:
```js
import {Foo} from './Foo'
export class SomeClass {
doSomethingWithFoo(o: Foo) {
console.log(o.foo) // OK, no type error.
}
}
```
The code inside of `SomeClass` could be allowed to access the private properties. This is great at design time when the author controls the source for both sides, but still wants to hide private parts from the end user on the outside of the APIs.
This allows for patterns like object managers being able to control certain aspects of the objects they manage (for example a renderer in a game engine that manages how objects in a scene render, and the renderable properties that the engine manager accesses, which are private to the end user) are composed from reactions to changes in the objects' public APIs by end users).
At the moment, the only way to allow objects to access properties of other objects is to make them public, but this does not allow for us to hide (at least in the type system) certain implementation details of a cross-class or cross-module whole system from the end user.
## Checklist
My suggestion meets these guidelines:
* [x] This wouldn't be a breaking change in existing TypeScript/JavaScript code
* [x] This wouldn't change the runtime behavior of existing JavaScript code
* [x] This could be implemented without emitting different JS based on the types of the expressions
* [x] This isn't a runtime feature (e.g. library functionality, non-ECMAScript syntax with JavaScript output, etc.)
* [ ] This feature would agree with the rest of [TypeScript's Design Goals](https://github.com/Microsoft/TypeScript/wiki/TypeScript-Design-Goals).
- I'm not sure about this. This feature wouldn't change semantics of JavaScript, or add any JavaScript features, but would only add additional type capabilities to limit use of APIs in certain ways. In plain JS, before I ported a JavaScript project to TypeScript, I had an object manager that accessed properties prefixed with `__` to denote that the properties were not to be used by the end user. The engine accessed the properties for internal implementation of the system.
| Suggestion,Awaiting More Feedback | medium | Critical |
534,273,050 | rust | Make it easier to run build in CI Docker image | The `isCiBranch` function now requires either the Azure Pipelines or Github Actions check to pass:
https://github.com/rust-lang/rust/blob/9630dbbc3caca60f2482e6eae8904aa6bda54f93/src/ci/shared.sh#L60-L67
This means that a plain `./src/ci/docker/run.sh $image_name` will no longer work. Setting `TF_BUILD=True` appears to work, as it makes the script think it's running on Azure pipelines. However, discovering this required digging around in several shell scripts, and I'm not totally sure that it does the right thing.
Additionally, the `./src/ci/docker/run.sh` is not very discoverable - I don't think it's mentioned outside of the CI scripts and the README in `./src/ci/docker`.
It would be nice if there was a simple way to do a local build of the compiler in a given Docker image (perhaps a top-level script adjacent to `x.py`). This would simplify debugging PRs where the PR builder tests pass, but the full merge-commit test suite fails. | C-enhancement,T-bootstrap,T-infra | low | Critical |
534,286,788 | PowerToys | [Fluent UX] Shortcut Guide v2 | PowerToys should do its best to look like a system level integration. Shortcut guide is a great example of where we can be but needs a little bit more polish.
- #792 - Screen preview in shortcut view only shows current display
- #179 - shortcut list update
- #235 - configurable list
- #853 - adjustments in fade in effect and interaction model
# Current screen
# Update: Possible Proposed adjustments
Here is one idea we've been working with the shell team with. It would allow for contextual hooks for applications, popular items can be included in, teaching moments for 'popular' and
**In context of full screen:**
| Product-Shortcut Guide,Area-Localization,Status-In progress,Area-User Interface,Cost-Large,External Dependency-WinUI 3,UI refresh | high | Critical |
534,304,135 | go | cmd/go, net, runtime: TestScript flake with "Failed to find getsockname procedure in ws2_32.dll: The specified procedure could not be found." | Just seen at https://storage.googleapis.com/go-build-log/40de7642/windows-amd64-2016_533e366a.log, there is a panic `Failed to find getsockname procedure in ws2_32.dll: The specified procedure could not be found.`
Perhaps this is just a flake but failing to find getsockname in the Windows Socket DLL looks serious. Kindly paging some related/interested parties @bcmills @alexbrainman @bradfitz @randall77 | NeedsInvestigation,compiler/runtime | low | Critical |
534,307,782 | rust | Supertrait method visible in generic but not non-generic code | Suppose you `use` a trait but not its supertrait, then try to call a method from the supertrait using dot syntax. This works if the receiver is a generic parameter bounded on the child trait, but not if it's a concrete type that implements the child trait.
Example [(playground link)](https://play.rust-lang.org/?gist=ee809f4bdeec5c6cf7b1751ee1642193&version=stable&mode=debug&edition=2015):
```rust
mod private {
pub trait Super {
fn foo(self);
}
pub trait Child: Super {}
impl Super for u32 {
fn foo(self) {}
}
impl Child for u32 {}
}
use private::Child;
fn huh<T: Child>(x: T) {
x.foo(); // OK
}
fn hah(x: u32) {
huh(x); // OK
x.foo(); // error: no method named `foo` found
// for type `u32` in the current scope
}
```
This might be intended behavior, but in a quick search I didn't find any documentation of it, and it seems very strange. | A-trait-system,A-associated-items,T-lang,C-discussion,T-types | low | Critical |
534,311,210 | godot | Assigned type (Class) doesn't match the function argument's type (Class) | Godot 3.2 beta2
I am getting a strange error with typed GDScript, where it finds incompatibility between argument type and parameter type... while they are actually the same.
```
SCRIPT ERROR: GDScript::reload: Parse Error: At "update_episode_data_from_file()" call, argument 1. Assigned type (Project) doesn't match the function argument's type (Project).
At: res://main.gd:294
```
I have a `script_data.gd` file with a `Project` class inside.
I also have a `script_parser.gd` file which const-imports the previous one, in order to define the following function:
```gdscript
const ScriptData = preload("./script_data.gd")
...
static func update_episode_data_from_file(project: ScriptData.Project, path: String) -> Array:
...
```
Then I const-imported both scripts inside `main.gd`, in which I have the following:
```gdscript
const ScriptData = preload("./script_data.gd")
const ScriptParser = preload("./script_parser.gd")
...
var _project : ScriptData.Project = null
...
func _open_project(fpath: String):
...
var dir := fpath.get_base_dir()
for ep_file_rpath in data.episode_files:
var path := dir.plus_file(ep_file_rpath) as String
ScriptParser.update_episode_data_from_file(_project, path)
```
This code works fine, but recently I've been getting this typing error in the editor, and it obscures any error happening elsewhere after this line...
So far the only thing that fixes it is to close the editor and reopen it. Then a few hours later, it reappeared again.
I tried to repro in a simpler project, but could not find a pattern yet... | bug,topic:gdscript,confirmed | medium | Critical |
534,319,905 | TypeScript | Symbols in const assertions should be considered unique | **TypeScript Version:** 3.8.0-dev.20191205
**Search Terms:**
- const assertion unique symbol
- "unique symbol" "as const"
**Code**
```ts
const prop = Symbol()
const dictionary = {
prop: Symbol()
}
const sealedDictionary = {
prop: Symbol()
} as const
class MyClass {
// Works because "prop" is a unique symbol
[prop]: 1
// Does not work because "dictionary.prop" is not a unique symbol
[dictionary.prop]: 2
// Does not work because "sealedDictionary.prop" is not a unique symbol,
// even though it probably should be?
[sealedDictionary.prop]: 3
}
```
**Expected behavior:**
`MyClass` should probably accept `sealedDictionary.prop` as a computed property name since it should be considered a unique symbol.
**Actual behavior:**
TypeScript complains with TS1166:
> A computed property name in a class property declaration must refer to an expression whose type is a literal type or a 'unique symbol' type.
**Playground Link:** ~[Provided](https://www.typescriptlang.org/play/?target=99&ts=3.8.0-dev.20191205#code/MYewdgzgLgBADgJxHGBeGBlAngWwEYgA2AFAJQCwAUKJLACYCWwUD4AhglmjAN5UzwkcAFyZcBEhUoBfKjWgwIAUzaEldACJMW7Ttz6UBiZKOz4iZKtJhsIMeVCpzCtuwFksAYRcQ7BgQD0ATAA6iAIANZ2eErAbACuyjAARMZwyTAMdmww8WAMAI7xSoriRPwwANppALqiAIxOhjBBMBogSnZgILAA7uERMDFxiSXJjMysYBxYAHRpGVkw3bA5eYXFpeaEFZUTOtOc80J1MABMTYHB7Z3LPTD9kUOxCUnJyqrqWpO6cwuZXXua3yRRKEDKhAANBVWkoAG5KMAwKAACxA8QA5ijMrBjHg2HhCFwIGj4oQ6M8APy7D5qTTaKYzY7IU4AZisQA)~ [Updated](https://www.typescriptlang.org/play?ts=3.8.2#code/MYewdgzgLgBADgJxHGBeGBlAngWwEYgA2AFAJQCwAUKJLACYCWwUD4AhglmjAN5UzwkcAFyZcBEhUoBfKjWgwIAUzaEldACJMW7Ttz6UBiZKOz4iZKtJhsIMeVCpzCtuwFksAYRcQ7BgQD0ATAA6iAIANZ2eErAbACuyjAARMZwyTAMdmww8WAMAI7xSoriRPwwANppALqiAIxOhjBBMBogSnZgILAA7uERMDFxiSXJjMysYBxYAHRpGVkw3bA5eYXFpeaEFZUTOtOc80J1MABMTYHB7Z3LPTD9kUOxCUnJyqrqWpO6cwuZXXua3yRRKEDKhAANBVWkoAG5KMAwKAACxA8QA5ijMrBjHg2HhCFwIGj4oQ6M8APy7D5qTTaKYzY7IU4AZisQA) (old link seems to be broken in new playground)
**Related Issues:** —
| Suggestion,Awaiting More Feedback | low | Critical |
534,332,769 | godot | Classes are inconsistently highlighted | Godot 3.2 beta 2
Given the following code:
```gdscript
class A:
var a
static func test():
var ep : A = null
ep = A.new() as A
```
Highlighting is inconsistent:
Here on more involved code:
| bug,topic:gdscript,topic:editor,confirmed | low | Major |
534,350,428 | pytorch | How to set not to build libtorch_cpu.so and libmkl_*.so dependencies? | ``` linux-vdso.so.1 (0x00007fffa4bfc000)
libtorch_cpu.so => /home/xxxxx/workfiles/work/pytorch/torch/lib/./libtorch_cpu.so (0x00007f63d4f6c000)
librt.so.1 => /lib64/librt.so.1 (0x00007f63d4d52000)
libgcc_s.so.1 => /lib64/libgcc_s.so.1 (0x00007f63d4b3c000)
libdl.so.2 => /lib64/libdl.so.2 (0x00007f63d4938000)
libmkl_intel_lp64.so => /lib/libmkl_intel_lp64.so (0x00007f63d3e06000)
libmkl_gnu_thread.so => /lib/libmkl_gnu_thread.so (0x00007f63d25cd000)
libmkl_core.so => /lib/libmkl_core.so (0x00007f63ce494000)
libpthread.so.0 => /lib64/libpthread.so.0 (0x00007f63ce275000)
libm.so.6 => /lib64/libm.so.6 (0x00007f63cdf73000)
libc10.so => /home/xxxxx/workfiles/work/pytorch/torch/lib/./libc10.so (0x00007f63cdd31000)
libstdc++.so.6 => /lib64/libstdc++.so.6 (0x00007f63cd9ae000)
libgomp.so.1 => /lib64/libgomp.so.1 (0x00007f63cd788000)
libc.so.6 => /lib64/libc.so.6 (0x00007f63cd3dc000)
/lib64/ld-linux-x86-64.so.2 (0x000055795895f000)
```
| module: build,triaged,module: mkl | low | Minor |
534,357,757 | vue | When v-show render instruction is passed a reference object, the instruction will not work | ### Version
2.6.10
### Reproduction link
[https://codepen.io/SamirGuo/pen/vYBezXz](https://codepen.io/SamirGuo/pen/vYBezXz)
### Steps to reproduce
```html
<div id="app">
<template>
<div>
<button @click="visible = !visible">显示/隐藏</button>
<my-component :visible="visible" style="border-bottom: solid 1px #ccc" />
</div>
</template>
</div>
```
```js
const directive = {
name: 'show',
value: true
};
Vue.component('MyComponent', {
props: {
mystyle: '',
visible: true
},
render(h) {
directive.value = this.visible;
return h('div', {
// do not work!
directives: [directive],
// work well!
// directives: [{
// name: 'show',
// value: this.visible
// }],
domProps: {
innerText: 'sssssss'
},
style: this.mystyle
});
}
});
var Main = {
data() {
return {
visible: true
};
}
};
var Ctor = Vue.extend(Main)
new Ctor().$mount('#app')
```
### What is expected?
v-show to be applied
### What is actually happening?
v-show isn't applied
---
When the instruction update logic assigns oldValue, oldVnode.data.directives ["show"] and vnode.data.directives ["show"] are actually the same object.
So dir.oldValue = oldDir.value is actually equivalent to dir.oldValue = dir.value; in the later update event, dir.oldValue === dir.value
<!-- generated by vue-issues. DO NOT REMOVE --> | discussion,has workaround | low | Major |
534,363,820 | PowerToys | Help make it easier for users to choose apps that start automatically | I noticed this tweet https://twitter.com/MicrosoftHelps/status/1203086731815407617 and thought that sounds okay for technical people but too hard for a lot of users. They also provide a link to https://support.microsoft.com/en-us/help/4026268/windows-10-change-startup-apps which seems slightly different. Maybe this is something the Windows 10 team will improve but it also seems like a a good thing for a Power Toy. | Idea-New PowerToy | low | Minor |
534,382,296 | godot | Calling a getter on a null object in a getter creates loop | **Godot version:**
v3.2-beta2_win32
Still works properly in v3.1.2.stable
seems to happen in all versions higher than 3.1.2.stable. Also the nightly builds and my self built godot project seem to have the same bug
**OS/device including version:**
Windows 10
**Issue description:**
There seems to be a problem/Change with the way how getters are being called:
In the reproduction-project, if i have 2 objects(Object1, Object2) with declared getters.
Object1 hold Object2 as variable.
In the get_call for the object2, the object1 calls the getter-method of the object2.
If i now run a script where i create the Object1 but dont set the Object2 within Object1, the crashing will happen.
Now the game just starts looping endlessly (depending on the version it will either loop endlessy and kind of lock down the editor and game or crash the game and quit debugging and return to the editor)
I did this to have something of a "converter-method" in my object, so that i wouldd get back a different data-structure when i accessed the variable from outside.
But as this would correctly tell me: "Nonexistent function 'get_object2' in base 'Nil'" in 3.1.2 i assume this is currently a bug.
**Steps to reproduce:**
Open reproduction project and press start
**Minimal reproduction project:**
(Edit: replaced with version with uncompiled scripts)
(Edit2: replaced the zip with one that contains the project.godot for better ease of opening)
[reproduction project get loop bug 2.zip](https://github.com/godotengine/godot/files/9277039/reproduction.project.get.loop.bug.2.zip)
Sidenote:
@willnationsdev This also has implications for the custom-resource branch (https://github.com/godotengine/godot-proposals/issues/18), where the same error would happen for Custom-Resources, but slightly different: even if the resources are properly set (and not null). The debugger would crash, show wrong info, or not show up at all, but the game would continue running. So i dont know in how far the bug of this issue is the culprit for that behaviour. I couldnt build | bug,topic:gdscript,confirmed | low | Critical |
534,397,121 | node | net: pending socket re-use | It looks like `net.Socket` instances are intended to be re-used however while reading through the code I spot some potential problems:
- `_destroy()`: doesn't wait for connect to finish
- `connect`: doesn't first destroy if already connected, before connecting again
- `connect`: doesn't wait for pending destroy to finish, before calling e.g. `_undestroy()`
Not sure if these are actual problems that need to be fixed?
@mcollina | net | low | Minor |
534,419,956 | pytorch | Spurious negative output in convolution of positive tensors | The gist of the issue is that at one point in my network in my project I feed positive float32 inputs `X` and `W` into a 1d (or 2d) convolution, but unexpectedly obtain explicitly negative outputs.
## Details
The X input is nonnegative about 1e-2 with many explicit 0 (after an elementwise square, or abs), and the W input is positive but with quite large range of [1e–8, 1e+8] (after an exp). I take `F.conv2d(X, W, None)` and **expect nonnegative** outputs (finite or infinite), but actually get **significantly negative** values. The downstream computations rely on the result of this operation to be at least nonnegative.
I tried `torch.backends.cudnn.deterministic = True`, or fp64, or clamping the `W` value and each time the problem seemed to had gone away. In my project I ended up clamping the result of the convolution to nonnegative range, and it sufficed.
Having observed this oddity got me worried, since in my understanding of floating point arithmetic the loss of precision and round off could yield imprecise results, but not **flip the sign** of the overall result. I read in the docs, that non-deterministic cudnn algorithms might induce this behaviour and yield inconsistent results, but I did not expect that atomic additions (according to cudnn reference) could spuriously flip the sign bit of a 32bit float.
I understand that Github issues are not a Q&A support forum, but I think that this inconsistency, might of interest, and would very much appreciate any discussion of guarantees of single precision operations.
## Minimally reproducing example
I came up with a minimal example of this odd behaviour. My observations are as follows (based on this snippet, and on the real data within the network in my project):
* negative outputs of various magnitude occur **only if** deterministic is *False*, and data is *fp32*, and the input `X` is a moderately large-dim tensor.
* `X` and `W` of dims `...x12x12` and `...x4x4`, respectively, give a lot of very small fp32 numbers, some of which are negative
* `X` and `W` of dims `...x11x11` and `...x3x3`, respectively, produce very rare negative tiny fp32, but mostly expected zeros
These suggest that some optimizations within cudnn might be responsible.
The following code constructs explicitly nonnegative `X` and `W`: `X` all zeros except for a single **1** at the *centre*, `W` is a 4x4 kernel with constant **1e-3**. I attach the pickled [result.gz](https://github.com/pytorch/pytorch/files/3935238/result.gz)
```python
import gzip
import torch
import torch.nn.functional as F
# change these
torch.backends.cudnn.deterministic = False
devtype = dict(device=torch.device("cuda:0"), dtype=torch.float32)
# x -- zeros with a spike
x = torch.zeros(1024, 1, 12, 12).to(**devtype)
x[..., 5, 5] = 1.
# small nonnegative values
w = (torch.ones(1, 1, 4, 4) * 1e-3).to(**devtype)
assert x.ge(0).all() and w.ge(0).all()
# conv1d/2d
result = F.conv2d(x, w, None)
with gzip.open("result.gz", "wb") as fout:
torch.save(dict(x=x, w=w, result=result), fout)
assert result.ge(0).all()
```
To unpack the attached `.gz` pickled result.
```python
import gzip
import torch
with gzip.open("result.gz", "rb") as fin:
pack = torch.load(fin)
pack["result]
```
Here is a link to a [colab notebook](https://colab.research.google.com/drive/136n8oKDitchxEkXMhV7AJeYfFUayiS4X), where it can be seen, that the output is spurious even for `X` is `25 x 1 x 12 x 12`. Especially striking is that `24 x 1 x 12 x 12` appears to be correct.
## Specs and versions
Below are very abridged specs:
`Intel(R) Xeon(R) CPU E5-2698 v4 @ 2.20GHz, 256GB RAM`, `4x GeForce GTX 1080 Ti` with `Driver Version: 390.77` on `Linux Mint 18.3 Sylvia (GNU/Linux 4.10.0-38-generic x86_64)` and conda `python 3.7.4` packages:
* `pytorch 1.1.0 py3.7_cuda9.0.176_cudnn7.5.1_0 pytorch`
* `cudatoolkit 9.0 h13b8566_0`
cc @jlin27 @mruberry | module: docs,module: convolution,triaged | low | Major |
534,423,123 | godot | Editing AudioStreamSample does not update project files | **Godot version:**
3.2.master @8eb183aebb9c79ff92d6f566af7ad2f91696ce08
**OS/device including version:** Arch Linux
**Issue description:** I have an AudioStreamPlayer that has a stream containing a .wav sound attached. When I edit the properties of said stream (looping properties in my case) those changes won't be saved to disk nor have effect when running a scene. They work in the editor however.
**Steps to reproduce:** Just create an AudioStreamPlayer, drop a .wav into the stream property and set some valid looping range of the stream. Set the AudioStreamPlayer's autoplay proeprty to true and run the scene. The sound will loop, but the range is not applied, so the sample as a whole is looped.
When instead of running the scene you tick the playing property of the AudioStreamPlayer in the editor, the sample loops just the part given in the stream's properties like expected.
There are two workarounds I know of: either make the stream unique or attach a script to the scene/AudioStreamPlayer/what ever node you like and set the loop properties manually, for example in the _ready function.
**Minimal reproduction project:**
[BugAudioStreamPlayer.zip](https://github.com/godotengine/godot/files/3935289/BugAudioStreamPlayer.zip)
Since the looping range is not saved, you have to set some before running the scene/setting "playing". I used values 10000 for Loop Begin and 15000 for Loop End. | bug,confirmed,topic:audio,topic:import | low | Critical |
534,433,929 | node | repl / eval: CommonJS globals leak into ESM modules | * **Version**: v14.0.0-pre / cf5ce2c9e1
* **Platform**: `Linux lt2.cfware.com 5.3.11-200.fc30.x86_64 #1 SMP Tue Nov 12 19:25:25 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux`
* **Subsystem**: repl, CLI `--eval` and `--print` without `--input-type=module`
Create `script.mjs`:
```js
console.log('script.mjs', typeof require);
```
Running `node ./script.mjs` or `node --input-type=module --eval "import('./script.mjs')"` both produce the correct output `script.mjs undefined`.
Now run `import('./script.mjs')` in repl, this produces output `script.mjs function`. Same for `node --eval "import('./script.mjs')"`.
Using `--print` in place of `--eval` does not change `typeof require`.
CC @nodejs/modules-active-members | repl,cli,experimental,esm | medium | Major |
534,454,273 | go | cmd/compile: redundant typecheck errors when using map[T]func(Undefined) | <!-- Please answer these questions before submitting your issue. Thanks! -->
### What version of Go are you using (`go version`)?
<pre>
$ go version
go version devel +da4d58587e Sat Dec 7 15:57:30 2019 +0000 linux/amd64
</pre>
### Does this issue reproduce with the latest release?
Yes.
### What did you do?
https://play.golang.org/p/PYwYgOfQo5S
### What did you expect to see?
Three simple errors, just saying that `Undefined` is undefined on each of the three lines.
### What did you see instead?
A redundant error on the `map[...]func(...)` line:
```
$ go build f.go
# command-line-arguments
./f.go:6:14: undefined: Undefined
./f.go:7:16: undefined: Undefined
./f.go:8:4: cannot use map[int]func(<T>) literal (type map[int]func(<T>)) as type map[int]int in assignment
./f.go:8:19: undefined: Undefined
```
It distracted me in some real code, particularly because the shorter and more helpful error is below it. We probably want all three cases to error in a consistent way. | NeedsInvestigation,compiler/runtime | low | Critical |
534,463,086 | TypeScript | @types's definition doesn't match its own type | https://github.com/DefinitelyTyped/DefinitelyTyped/pull/40731 broke the following global declaration.
cc @rbuckton
**TypeScript Version:** 3.7.x-dev.20191207
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:**
**Code**
https://github.com/falsandtru/spica/blob/v0.0.289/global.test.d.ts and @types/[email protected]
```ts
import assert from 'power-assert';
type Assert = typeof assert;
declare global {
const assert: Assert;
}
```
**Expected behavior:**
pass
**Actual behavior:**
`TypeScript error: global.test.d.ts(6,9): Error TS2403: Subsequent variable declarations must have the same type. Variable 'assert' must be of type 'typeof assert', but here has type 'typeof assert'.`
**Playground Link:** <!-- A link to a TypeScript Playground "Share" link which demonstrates this behavior -->
**Related Issues:** #32808
| Bug | low | Critical |
534,472,998 | node | Investigate flaky test parallel/test-module-loading-globalpaths | <!--
Thank you for reporting a possible bug in Node.js.
Please fill in as much of the template below as you can.
Version: output of `node -v`
Platform: output of `uname -a` (UNIX), or version and 32 or 64-bit (Windows)
Subsystem: if known, please specify the affected core module name
If possible, please provide code that demonstrates the problem, keeping it as
simple and free of external dependencies as you can.
-->
* **Version**: master
* **Platform**: Windows
* **Subsystem**: test
<!-- Please provide more details below this comment. -->
```
internal/modules/cjs/loader.js:1026
throw err;
^
Error: Cannot find module 'foo'
Require stack:
- C:\workspace\node-test-binary-windows-js-suites\node\test\parallel\test-module-loading-globalpaths.js
at Function.Module._resolveFilename (internal/modules/cjs/loader.js:1023:17)
at Function.Module._load (internal/modules/cjs/loader.js:910:27)
at Module.require (internal/modules/cjs/loader.js:1085:19)
at require (internal/modules/cjs/helpers.js:72:18)
at Object.<anonymous> (C:\workspace\node-test-binary-windows-js-suites\node\test\parallel\test-module-loading-globalpaths.js:15:15)
at Module._compile (internal/modules/cjs/loader.js:1190:30)
at Object.Module._extensions..js (internal/modules/cjs/loader.js:1229:10)
at Module.load (internal/modules/cjs/loader.js:1045:32)
at Function.Module._load (internal/modules/cjs/loader.js:949:14)
at Function.executeUserEntryPoint [as runMain] (internal/modules/run_main.js:71:12) {
code: 'MODULE_NOT_FOUND',
requireStack: [
'C:\\workspace\\node-test-binary-windows-js-suites\\node\\test\\parallel\\test-module-loading-globalpaths.js'
]
}
C:\workspace\node-test-binary-windows-js-suites\node\test\.tmp.355\install\node.exe - Access is denied.
Can't clean tmpdir: C:\workspace\node-test-binary-windows-js-suites\node\test\.tmp.355
Files blocking: [ 'install' ]
C:\workspace\node-test-binary-windows-js-suites\node\test\common\tmpdir.js:136
throw e;
^
Error: ENOTEMPTY: directory not empty, rmdir 'C:\workspace\node-test-binary-windows-js-suites\node\test\.tmp.355'
at Object.rmdirSync (fs.js:783:3)
at rmdirSync (C:\workspace\node-test-binary-windows-js-suites\node\test\common\tmpdir.js:86:10)
at rimrafSync (C:\workspace\node-test-binary-windows-js-suites\node\test\common\tmpdir.js:41:7)
at process.onexit (C:\workspace\node-test-binary-windows-js-suites\node\test\common\tmpdir.js:121:5)
at process.emit (events.js:314:22) {
errno: -4051,
syscall: 'rmdir',
code: 'ENOTEMPTY',
path: 'C:\\workspace\\node-test-binary-windows-js-suites\\node\\test\\.tmp.355'
}
```
On: https://ci.nodejs.org/computer/test-rackspace-win2008r2_vs2017-x64-4/
Ref: https://ci.nodejs.org/job/node-test-binary-windows-js-suites/50/RUN_SUBSET=1,nodes=win2008r2-COMPILED_BY-vs2017/testReport/junit/(root)/test/parallel_test_module_loading_globalpaths/
| flaky-test | low | Critical |
534,481,757 | flutter | Support for interative Drag-Drop Reorderable GridView | Hope GridView can support interactive drag drop reorder.
| c: new feature,framework,f: scrolling,would be a good package,c: proposal,P3,team-framework,triaged-framework | medium | Major |
534,503,842 | godot | FileDialog access scope can be broken with relative path | **Godot version:** v3.1.2.stable.offical & v3.2.beta.custom_build.8eb183aeb
**OS/device including version:** macOS Catalina 10.15.1 / MacBook Pro (Retina, 13-inch, Early 2015)
**Issue description:**
FileDialog has three access scopes: Resource, UserData, and FileSystem.
It's expected to only affect / select files within the scope. But relative paths are not checked properly. So we can actually
* Create folders that's out of scope
* Select files / folders that's out of scope
**Steps to reproduce:**
As the editor uses FileDialog:
1. Open any project, and create a non-empty scene
2. Click menu: Scene -> Save Scene As
3. The "Save Scene As" dialog appears
* The root has a root of "res://", and clicking "Up" button has no effect, everything is fine
4. Click "Create Folder" button, Enter the name "../x", and click OK
* A folder named "x" is created outside your project root
* Fortunately, the main file list view is not switched to that directory
5. In the file name edit box, change the name to "../Scene.tscn", and click Save
* The scene will be saved outside your project root
* FileSystem panel adds a "Scene.tscn" entry under the "res://" node. Clicking on that entry produces error.
```
Switch Scene Tab
scene/resources/resource_format_text.cpp:1292 - Method/Function Failed.
Switch Scene Tab
Cannot open file 'res://Scene.tscn'.
Failed loading resource: res://Scene.tscn.
Cannot navigate to 'res://Scene.tscn' as it has not been found in the file system!
```
(The scene got added as a "res://" child thing may be a separate issue.) | bug,confirmed,topic:gui | low | Critical |
534,504,607 | flutter | video_player on 4K Amazon Fire TV stick shows small video on green background | # Steps to Reproduce
Create a video player using the video_player plugin and play it on a 4K Amazon Fire TV stick.
No matter what widget you use to contain the video, the final video will play at a smaller size than the containing widget and the rest of the widget will be filled with green.
The `flt_video_player` plugin does not share this behavior. | e: device-specific,platform-android,engine,p: video_player,package,dependency: android,P2,team-android,triaged-android | low | Major |
534,506,863 | godot | GraphEdit set_selected does not trigger node_selected signal | **Godot version:**
3.2 beta 2 mono
**OS/device including version:**
ArcoLinux
**Issue description:**
Using set_selected on GraphEdit or setting `selected = true` on a GraphNode will not trigger the node_selected signal; only by clicking on the graph node will the signal fire. Ideally the signal will fire when the GraphNode is selected, whatever method is used, rather than only when the mouse clicks on it.
[GraphEditSignal.zip](https://github.com/godotengine/godot/files/3935978/GraphEditSignal.zip)
| discussion,topic:core | low | Minor |
534,507,807 | terminal | Option to not clear selection on copy | <!--
🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
I ACKNOWLEDGE THE FOLLOWING BEFORE PROCEEDING:
1. If I delete this entire template and go my own path, the core team may close my issue without further explanation or engagement.
2. If I list multiple bugs/concerns in this one issue, the core team may close my issue without further explanation or engagement.
3. If I write an issue that has many duplicates, the core team may close my issue without further explanation or engagement (and without necessarily spending time to find the exact duplicate ID number).
4. If I leave the title incomplete when filing the issue, the core team may close my issue without further explanation or engagement.
5. If I file something completely blank in the body, the core team may close my issue without further explanation or engagement.
All good? Then proceed!
-->
# Description of the new feature/enhancement
Currently when you copy text it clears your selection. While this is current behavior for some terminals, this is different than standard Windows behavior. Most Windows programs simply leave the text selected and require some other action to deselect the text. This is particularly bad if you have remapped Copy to Ctrl-C, then accidentally hitting it twice can result in killing your running process accidentally. Additionally, I frequently will copy text, then need to come back to it later to recopy, but due to this issue I have to re-select the text frequently.
I just want copy/paste to work the same here that it does in every other program.
<!--
A clear and concise description of what the problem is that the new feature would solve.
Describe why and how a user would use this new functionality (if applicable).
-->
<!--
A clear and concise description of what you want to happen.
-->
| Help Wanted,Area-TerminalControl,Product-Terminal,Issue-Task,Priority-2,good first issue | low | Critical |
534,525,246 | flutter | Bringing Fluent Design to Flutter for additional Mobile design options and Desktop nativity | We already have Material and Cupertino for Flutter, having redmond too can be a value addition while we get ready for Desktop.
I would love to start writing up the code but I just wanted to make sure if someone is already doing it. Even better, if someone on the Flutter team is doing it.
Is it even good or do we have the need to have Fluent design? For the most part, it's a wrap around Material Design.
I'm open to further discussions on this. | c: new feature,framework,platform-windows,c: proposal,a: desktop,P3,team-windows,triaged-windows | high | Critical |
534,525,403 | go | cmd/go: document the meaning of solitary "command-line-arguments" | <!-- Please answer these questions before submitting your issue. Thanks! -->
### What version of Go are you using (`go version`)?
<pre>
$ go version
</pre>
### Does this issue reproduce with the latest release?
### What operating system and processor architecture are you using (`go env`)?
<details><summary><code>go env</code> Output</summary><br><pre>
$ go env
GO111MODULE="on"
GOARCH="amd64"
GOBIN="/home/user/bin"
GOCACHE="/home/user/.cache/go-build"
GOENV="/home/user/.config/go/env"
GOEXE=""
GOFLAGS=""
GOHOSTARCH="amd64"
GOHOSTOS="linux"
GONOPROXY=""
GONOSUMDB=""
GOOS="linux"
GOPATH="/home/user"
GOPRIVATE=""
GOPROXY="https://proxy.golang.org,direct"
GOROOT="/home/user/go"
GOSUMDB="sum.golang.org"
GOTMPDIR=""
GOTOOLDIR="/home/user/go/pkg/tool/linux_amd64"
GCCGO="gccgo"
AR="ar"
CC="gcc"
CXX="g++"
CGO_ENABLED="1"
GOMOD="/dev/null"
CGO_CFLAGS="-g -O2"
CGO_CPPFLAGS=""
CGO_CXXFLAGS="-g -O2"
CGO_FFLAGS="-g -O2"
CGO_LDFLAGS="-g -O2"
PKG_CONFIG="pkg-config"
GOGCCFLAGS="-fPIC -m64 -pthread -fmessage-length=0 -fdebug-prefix-map=/tmp/go-build773661967=/tmp/go-build -gno-record-gcc-switches"
</pre></details>
### What did you do?
```
~ $ go list -m
command-line-arguments
```
and look for "command-line-arguments" in documentation and the source tree.
### What did you expect to see?
An explanation of the significance of the output.
### What did you see instead?
Crickets (though also some bread crumbs in the source and commit messages).
It would be nice if the meaning were properly documented. | Documentation,help wanted,NeedsFix | low | Critical |
534,528,086 | neovim | LSP: hook into all traffic | requested in https://github.com/neovim/nvim-lsp/issues/45 :
> implement a catch-all callback for any server notifications
Something like `lsp.on_response(handled)` which is `nil` by default, otherwise all server => client responses pass through it, and `handled=true` if there is a handler for it already.
| enhancement,lsp | low | Major |
534,531,954 | TypeScript | Intellisense breaks when typing a function | **TypeScript Version:** 3.8.0-dev.20191207
Module.ts file:
/**
* My awesome function description
*/
export let myFunc = () => {
// do something...
return;
}
In index.ts:
import { myFunc } from "./module.ts";
myFunc();
Hovering over "myFunc" gives proper Intellisense, including the description given in the function declaration ("My awesome function description"):
[link to screenshot: Intellisense showing function description][1]
However, if, in module.ts, I declare a type in the module and apply it to the function declaration like:
declare global {
/** MyFunc Type Def */
type myFuncType = () => void;
}
/**
* My awesome function description
*/
export let myFunc:myFuncType = () => {
// do something...
return;
}
then now Intellisense does not show the function description anymore in index.ts:
[link to screenshot: Intellisense not showing function description][2]
In the defining file, module.ts, Intellisense works in both cases.
[1]: https://i.stack.imgur.com/G5AE0.png
[2]: https://i.stack.imgur.com/IuXCI.png | Bug,Domain: JSDoc,Domain: Quick Info | low | Minor |
534,571,789 | react-native | Using `toUpperCase` or `textTransform: 'uppercase'` breaks on an Android controlled TextInput | React Native version: 0.61.4 and lower
Trying to force capitalization of characters inside a TextInput is broken on Android.
- `autoCapitalize="characters"` doesn't seem to do anything
- Using `toUpperCase` in the onChangeText hook causes duplication of letters
- Using `textTransform: 'uppercase'` in styles block causes the same duplication of letters
- Using `textTransform: 'uppercase'` in a non-controlled TextInput does nothing
## Steps To Reproduce
1. Create a controlled TextInput and either use onTextChanged to modify the text to uppercase or use text transform in the styles block
2. Type multiple lowercase characters into the text box
Describe what you expected to happen:
Characters should be capitalized
What actually happens:
Characters are capitalized and duplicated
Snack, code example, screenshot, or link to a repository:
https://snack.expo.io/@nmi09/rn-android-capitalize-input-bug
| Component: TextInput,Platform: Android,Bug,Never gets stale | high | Critical |
534,572,225 | vue | v-once on template not working inside a v-for in two cases | ### Version
2.6.10
### Reproduction link
[https://codepen.io/xiangyuecn/pen/eYmmPNP](https://codepen.io/xiangyuecn/pen/eYmmPNP)
moved from https://github.com/vuejs/vue/issues/10892
### Steps to reproduce
The repro adds entries to an array, causing a re render that. The template intentionally displays the elapsed time to see which v-once work and which do not
### What is expected?
All v-once to never render again
### What is actually happening?
Test 3 first template with a `v-once` re renders all the time. **Workaround** is to append an empty `span`: `<template v-once>[{{ getTime() }}] {{ obj.msg }}<span/></template>`
Test 5: the second template with `v-once` still rerenders. **Workaround** is to use a different tag like a span or use the `v-once` on the parent.
<!-- generated by vue-issues. DO NOT REMOVE --> | bug,has workaround | medium | Minor |
534,580,056 | rust | `#![windows_subsystem = "windows"]` hides cargo test output | **Problem**
Calling `cargo test` on a project that has the `#![windows_subsystem = "windows"]` global attribute does not show the expected human-readable output.
Instead of showing
```
> cargo test
Compiling no_output v0.1.0 (C:\Programming\no_output)
Finished dev [unoptimized + debuginfo] target(s) in 0.71s
Running target\debug\deps\no_output-c8b57bda8502ba55.exe
running 1 test
test tests::test_panic ... FAILED
failures:
---- tests::test_panic stdout ----
thread 'tests::test_panic' panicked at 'This tape has just self-destructed!', src\main.rs:9:9
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace.
failures:
tests::test_panic
test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out
error: test failed, to rerun pass '--bin no_output'
```
it shows
```
> cargo test
Compiling no_output v0.1.0 (C:\Programming\no_output)
Finished dev [unoptimized + debuginfo] target(s) in 1.40s
Running target\debug\deps\no_output-c8b57bda8502ba55.exe
error: test failed, to rerun pass '--bin no_output'
```
which not quite as informative as the former message (it does not tell which tests failed).
**Steps**
This requires a Windows computer.
1. Create a new project with `cargo new`.
2. Copy-paste the following into your `main.rs` file:
```
#![windows_subsystem = "windows"]
fn main() {
println!("Hello, world!");
}
#[cfg(test)]
mod tests {
#[test]
fn test_panic() {
panic!("This tape has just self-destructed!");
}
}
```
3. call `cargo test` on your project.
**Possible Solution(s)**
My guess is that this is due to executables targeting the Windows subsystem having no console and standard output initialized by default on Windows. Therefore, 2 possible solutions come to my mind:
- forcing `cargo test`-compiled executables to target the console subsystem when compiled for Windows;
- or having cargo explicitly initialize the console in the test framework automatically included in test executables: AFAIK this can be achieved by calling the [`AllocConsole()` function](https://docs.microsoft.com/en-us/windows/console/allocconsole) provided by the Windows API.
**Notes**
Output of `cargo version`: this happens on pretty much any cargo version.
Tested on `cargo 1.39.0 (1c6ec66d5 2019-09-30)` and `cargo 1.41.0-nightly (626f0f40e 2019-12-03)`.
Apologies if this has already been reported. I was unable to find a similar issue in the repository 😅 | O-windows,T-libs-api,A-libtest,C-bug | low | Critical |
534,582,917 | rust | [cg_ssa] TerminatorCodegenHelper::do_call assumes that destination doesnt have more than one predecessor | https://github.com/rust-lang/rust/blob/4abb0ad2731e9ac6fd5d64d4cf15b7c82e4b5a81/src/librustc_codegen_ssa/mir/block.rs#L127-L131
It appends the ret value storage code directly to the destination bb. This means that if two calls have the same destination, the destination bb will contain the ret value storage code for both calls. | A-codegen,T-compiler,C-bug,A-cranelift | low | Minor |
534,585,182 | godot | Reparenting node in Area.body_entered causes crash | **Godot version:**
```
3.1.2.stable.custom_build
```
**OS/device including version:**
```
Linux 5.4.2-arch1-1 x86_64 GNU/Linux
```
**Issue description:**
If you reparent a `Node` while handling the `Area.body_entered` signal, godot crashes.
My use case was an object that can "grab" another object like a magnet.
**Steps to reproduce:**
1. Create an `Area` with this script:
```
extends Area
func _ready():
connect("body_entered", self, "on_body_entered")
func on_body_entered(body: Node):
body.get_parent().remove_child(body)
add_child(body)
```
2. Move a `KinematicBody` into the area
```
Running: /usr/bin/godot --path /tmp/example --remote-debug 127.0.0.1:6007 --allow_focus_steal_pid 71021 --debug-collisions --position 448,240 res://Spatial.tscn
Godot Engine v3.1.2.stable.custom_build - https://godotengine.org
OpenGL ES 3.0 Renderer: GeForce GTX 970/PCIe/SSE2
ERROR: build: Condition ' O == __null ' is true. Continuing..:
At: core/math/quick_hull.cpp:402.
ERROR: build: Condition ' O == __null ' is true. Continuing..:
At: core/math/quick_hull.cpp:402.
ERROR: build: Condition ' O == __null ' is true. Continuing..:
At: core/math/quick_hull.cpp:402.
ERROR: build: Condition ' O == __null ' is true. Continuing..:
At: core/math/quick_hull.cpp:402.
ERROR: get_tree: Condition ' !data.tree ' is true. returned: __null
At: scene/main/node.h:263.
handle_crash: Program crashed with signal 11
```
I can try to repro with debug bits if desired.
**Minimal reproduction project:**
[example.zip](https://github.com/godotengine/godot/files/3937028/example.zip)
| bug,topic:core,confirmed,topic:physics,crash | low | Critical |
534,585,746 | rust | format!() ignores width when displaying enums | Similar to the issue reported in #55584 (and #55749), the [width](https://doc.rust-lang.org/std/fmt/#width) parameter is ignored when displaying enums, e.g. this fails:
```
assert_eq!(&format!("{:10}", Enum::FOO), "foo ");
```
[**Playground Example**](https://play.rust-lang.org/?version=stable&mode=debug&edition=2018&gist=490658bb2407ea934e7e0ff1ebf53742)
Notice this is with `Display`, not `Debug`. There's some question in the linked issues what the desired behavior for `Debug` is, but I think it's a bug that `Display` formatting would not respect width (and precision).
I'd be willing to take a crack at fixing this if I could get some code pointers. | C-enhancement,A-docs | low | Critical |
534,588,833 | godot | Light2D mask is ignored on shader with render_mode enabled (like blend_add) | **Godot version:**
v3.2.beta.custom_build.8eb183aeb
**OS/device including version:**
Windows 10
**Issue description:**
I do not know if this is a bug or correct behavior, but for me it was unexpected and undesired so I have created this ticket.
When shader use **render_mode blend_add** (or some other) then Light2D mask is not working for it.
This is what I want (render_mode is not set, light mask is a circle)
But when I enable render mode to (for example) blend_add then mask is ignored
**Minimal reproduction project:**
[Light2DShaderBug.zip](https://github.com/godotengine/godot/files/3937048/Light2DShaderBug.zip)
| bug,topic:rendering,confirmed,topic:2d | low | Critical |
534,600,478 | vscode | Support local paths in hover MarkdownString | Currently paths to images / links used in hover docs must be absolute, otherwise they will not work. Consider a file main.ts:
```typescript
/**
* [link](main.ts)
*/
function main() {
}
```
If you hover over main and click the link, it will attempt to open /main.ts file, which does not exist.
I think it would be sensible if the behavior was exactly the same as with markdown preview.
Example of file main.md:
```markdown
[link](main.ts)
```
Here, clicking the link would properly open the main.ts file, assuming that it's in the same directory.
Here is a project with these two files:
[project.zip](https://github.com/microsoft/vscode/files/3937151/project.zip)
| feature-request,upstream,api,markdown | high | Critical |
534,626,408 | pytorch | Wrong initialization with kaiming_uniform_ | ## 🐛 Bug
Standard deviation of the uniform distribution is (b-a)/(12)^0.5 where a and b are the bounds of the distribution. In the torch.nn.init.kaiming_uniform_ :
```
fan = _calculate_correct_fan(tensor, mode)
gain = calculate_gain(nonlinearity, a)
std = gain / math.sqrt(fan)
bound = math.sqrt(3.0) * std
```
Scaling elements by constant will scale variance by its square. This is the case with kaiming_normal_. But in uniform distribution, code is directly affecting the bounds and standard deviation is directly affected by the bounds. It should be like:
```
fan = _calculate_correct_fan(tensor, mode)
gain = calculate_gain(nonlinearity, a)
std = ( (gain)**2 ) / math.sqrt(fan)
bound = math.sqrt(3.0) * std
```
( In uniform distribution variance will be scaled by 1/4 after Relu so you can also double check using that information)
## Environment
PyTorch version: 1.3.0
CUDA used to build PyTorch: 10.1
Python version: 3.7
Versions of relevant libraries:
[pip3] numpy==1.17.3
[pip3] torch==1.3.0
[pip3] torchvision==0.4.1
| module: nn,triaged | low | Critical |
534,649,846 | electron | Refresh in-app-purchase App Store receipt for Mac | ### Preflight Checklist
<!-- Please ensure you've completed the following steps by replacing [ ] with [x]-->
* [x] I have read the [Contributing Guidelines](https://github.com/electron/electron/blob/master/CONTRIBUTING.md) for this project.
* [x] I agree to follow the [Code of Conduct](https://github.com/electron/electron/blob/master/CODE_OF_CONDUCT.md) that this project adheres to.
* [x] I have searched the issue tracker for a feature request that matches the one I want to file, without success.
### Problem Description
<!-- Is your feature request related to a problem? Please add a clear and concise description of what the problem is. -->
Some users reported that they have purchased in app products but it's not reflected in the app. It happens in my app and several others as in this tickets #14945. The in-app receipt is not updated after user purchased the product.
### Proposed Solution
It would be helpful to have a way to refresh the in-app receipt using `SKReceiptRefreshRequest`.
### Alternatives Considered
<!-- A clear and concise description of any alternative solutions or features you've considered. -->
If user purchases the product again, it works. I also checked with Apple and they said that "if the user deletes the app, and re-installs it from the App Store, the new appStoreReceipt should include the recent transaction". However these solutions are quite bad from user perspective. Having a way to refresh the receipt would make it a lot more reliable to implement in-app purchase.
### Additional Information
<!-- Add any other context about the problem here. -->
| enhancement :sparkles: | low | Major |
534,656,537 | godot | Dialog text is not copyable | Godot: 3.1.1
The text in certain dialogs can not be selected and copied.
For example on an error dialog popup, the text is not selectable or copyable, which is the standard expected behaviour.
Another example would be the about popup dialog where the version number is not copyable.
| enhancement,discussion,topic:gui | low | Critical |
534,703,784 | pytorch | false CHECK FAILED at ../aten/src/ATen/core/function_schema_inl.h | ## ❓ Questions and Help
I've converted custom pytorch model. it raises error when i'm trying to deploy on android.
with no backtrace.
i wanna know what causing this error.
here is the full error:
at android.app.ActivityThread$H.handleMessage(ActivityThread.java:1589)
at android.os.Handler.dispatchMessage(Handler.java:106)
at android.os.Looper.loop(Looper.java:164)
at android.app.ActivityThread.main(ActivityThread.java:6494)
at java.lang.reflect.Method.invoke(Native Method)
at com.android.internal.os.RuntimeInit$MethodAndArgsCaller.run(RuntimeInit.java:440)
at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:807)
Caused by: com.facebook.jni.CppException: false CHECK FAILED at ../aten/src/ATen/core/function_schema_inl.h (checkArg at ../aten/src/ATen/core/function_schema_inl.h:194)
(no backtrace available)
at org.pytorch.Module$NativePeer.runMethod(Native Method)
at org.pytorch.Module$NativePeer.access$100(Module.java:60)
at org.pytorch.Module.runMethod(Module.java:46)
at org.pytorch.helloworld.MainActivity.onCreate(MainActivity.java:67)
| module: internals,triaged | low | Critical |
534,712,421 | pytorch | error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型 |
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/order_preserving_flat_hash_map.h(1501): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/order_preserving_flat_hash_map.h(1559): note: 参见对正在编译的类 模板 实例化“ska_ordered::order_preserving_flat_hash_map<K,V,H,E,A>”的引用
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/order_preserving_flat_hash_map.h(1505): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/order_preserving_flat_hash_map.h(1513): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/order_preserving_flat_hash_map.h(1595): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/order_preserving_flat_hash_map.h(1632): note: 参见对正在编译的类 模板 实例化“ska_ordered::flat_hash_set<T,H,E,A>”的引用
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/order_preserving_flat_hash_map.h(1600): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/order_preserving_flat_hash_map.h(1604): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/order_preserving_flat_hash_map.h(1608): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/order_preserving_flat_hash_map.h(1612): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1383): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1441): note: 参见对 正在编译的类 模板 实例化“ska::flat_hash_map<K,V,H,E,A>”的引用
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1387): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1395): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1477): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1514): note: 参见对 正在编译的类 模板 实例化“ska::flat_hash_set<T,H,E,A>”的引用
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1482): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1486): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1490): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型
D:/STL/software/Anaconda/Anaconda33/envs/DB/lib/site-packages/torch/include\c10/util/flat_hash_map.h(1494): error C3203: “templated_iterator”: 未专用化的 类 模板 不能用作 模板 变量，该变量属于 模板 参数“_Ty1”，应为 real 类型 | needs reproduction,module: internals,triaged | low | Critical |
534,723,354 | godot | VisibilityEnabler2D does not work on instanced scenes | **Godot version:** 2.1.6 (happens on 3.1 as well)
**OS/device including version:** Windows 10, 64-bit
**Issue description:** `VisibilityEnabler2D` does not work on instanced scenes
**Steps to reproduce:**
- Run example
- Move away from the other 2 characters (WASD), to trigger `VisibilityEnabler2D`
- Go to `Debugger` -> `Live Scene Tree`.
- Find `player1` -> `character_container`, and click `AnimationPlayer` or `Particles2D`. They are still active.
**Minimal reproduction project:**
[visibilityenabler2dbug.zip](https://github.com/godotengine/godot/files/3938172/visibilityenabler2dbug.zip)
| discussion,topic:core | low | Critical |
534,779,962 | rust | Bootstrap should be able to learn target information from Rustc target database | There is lots of behavior in bootstrap that just does string matching on the target name to partially alter its behavior. A lot of the time, this information could've been determined by looking at a particular target option (from `rustc --print=target-spec-json`). | C-enhancement,T-bootstrap | low | Minor |
534,788,145 | kubernetes | Record 401 Unauthorized in the request counter metrics | <!-- Please only use this template for submitting enhancement requests -->
/sig api-machinery
/sig instrumentation
/kind feature
@kubernetes/sig-api-machinery-misc @logicalhan
**What would you like to be added**:
currently authn is working as a filter handler which is executed before the individual implementation of the resource APIs. however, the `apiserver_request_total` metric is recorded after the execution of those filter handlers. that makes us never recoding any of the `401 Unauthorized` bits.
consider promote such commonly-used metrics as `apiserver_request_total` to be a dedicated filter handler and make it works before the authn filter so we can observe the occurrence of those 401 requests. the pros and cons of recording `401` i can think of is:
__pros__:
- reminds cluster admin of certification expiration
- reminds cluster admin of potential improper external signer setups
__cons__:
- 401 requests can be sent by spammy "unfriendly" client, recording them can be overwhelming the normal requests.
- recording latencies (and other dimension of the requests) can be useless and also costing unnecessary performance loss at the apiserver.
- these authenticated requests can be considered "out-of-scope" of the apiserver in some sense
| priority/backlog,sig/api-machinery,kind/feature,sig/instrumentation,lifecycle/frozen | low | Major |
534,834,235 | create-react-app | react-scripts start should support CI=true, watchAll=false etc. | ### Is your proposal related to a problem?
<!--
Provide a clear and concise description of what the problem is.
For example, "I'm always frustrated when..."
-->
I'm running `react-scripts start` in a CI environment in order to run cypress end-to-end tests. As the project has grown this has started hitting the `inotify.max_user_instances` kernel limit which cannot be configured from within a container, this is reported as:
```
events.js:183
throw er; // Unhandled 'error' event
^
Error: watch /path/to/somewhere ENOSPC
at FSWatcher.start (fs.js:1382:19)
at Object.fs.watch (fs.js:1408:11)
at createFsWatchInstance (/path/to/somewhere
```
### Describe the solution you'd like
<!--
Provide a clear and concise description of what you want to happen.
-->
`react-scripts start` should support flags already used with `react-scripts test` that turn off webpack's file watching, these include env var `CI` and commandline flags `--watchAll=false`
### Describe alternatives you've considered
<!--
Let us know about other solutions you've tried or researched.
-->
I could use my production express server or something else to serve these files, but these aren't dependencies of the app whereas it's already using webpack-dev-server with configurable watch options.
| issue: proposal,needs triage | medium | Critical |
534,838,995 | react | DevTools: Provide full file path for React Native component stacks | ## Overview
In React Native, we're working on a new RedBox experience for errors and warnings called LogBox. In LogBox, we separate out component stack traces and show them similar to call stacks.
We'd like to be able to tap on these components and open them (like we can with call stacks).
## Solutions
In React we have the full file path context, but when we build the component stack trace, we strip the full path so that it displays only the file name.
There are two options to achieve this:
- For React Native, don't strip the full path [here](https://github.com/facebook/react/blob/b438699d3620bff236282b049204e1221b3689e9/packages/react-devtools-shared/src/backend/describeComponentFrame.js#L25). This will result in longer component stack messages with every frame having the full path.
- Keep the message the same, and instead add structured component stack frame info including the full file path. | Component: Developer Tools,Partner | low | Critical |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.