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 |
|---|---|---|---|---|---|---|
591,028,254 |
vscode
|
Add option `editor.lineDecorationsWidth` as official
|
<!-- ⚠️⚠️ Do Not Delete This! feature_request_template ⚠️⚠️ -->
<!-- Please read our Rules of Conduct: https://opensource.microsoft.com/codeofconduct/ -->
<!-- Please search existing issues to avoid creating duplicates. -->
<!-- Describe the feature you'd like. -->
Is there any chance that the setting of `editor.lineDecorationsWidth` will be official?
I currently have the "Unknown configuration setting" notification 😬
A lot of people using this option and it would be good to add it as official
|
feature-request,editor-core
|
high
|
Critical
|
591,041,244 |
go
|
x/mobile: key events not fired for predictive keyboards
|
On a Samsing Galaxy S10, running Android 10 (this was not an issue with the keyboard they shipped in Android 9) key events are no longer fired when the (default) predictive text is turned on.
Some other apps, such as Termux (https://github.com/termux/termux-app/issues/686) report that words appear after space/enter etc - however that is not true for gomobile. Each time a word should display a single "unknown key code" event is fired and no characters.
### What version of Go are you using (`go version`)?
1.12, 1.13, 1.14
### Does this issue reproduce with the latest release?
Yes
### What operating system and processor architecture are you using (`go env`)?
I think the environment on the Android phone is most relevant, but I don't know what to provide.
Code works fine on older keyboards, or if you turn off the predictive text
### What did you do?
* run a gomobile app on Galaxy S10 with Android 10, cause a keyboard to appear (I used fyne_demo app).
* type "hello " (ending in a space) and see that only a space is added to the entry.
* type "g " and you will see that "g" is shown followed by a space.
### What did you expect to see?
I want to see characters as they are entered. There are various docs about how to disable prediction (such as https://developer.android.com/reference/android/text/InputType#TYPE_NULL) but I cannot figure out how to set those options in a default gomobile app, as the root view does not seem to be configurable in GoNativeActivity.java.
### What did you see instead?
(as above)
|
NeedsInvestigation,mobile
|
low
|
Minor
|
591,067,121 |
go
|
spec: extend `Type assertions` section
|
Today I stubled upon [this](https://polyfloyd.net/post/golang-subtle-pitfalls/#2-type-assertions-of-value-types-vs-pointer-types) issue and I was hoping to find, in the reference doc, more examples of type assertions on pointer types but, unfortunately, there seems to be none.
Would it make sense to extend the current [Type assertions](https://golang.org/ref/spec#Type_assertions) section with this?
(BTW I'm a Go newbie so apologies in advance if I missed something in the existing doc!)
|
Documentation,NeedsFix
|
low
|
Minor
|
591,084,836 |
godot
|
Exported variables don't update in real-time
|
**Godot version:**
3.2 mono
**OS/device including version:**
ubuntu 19.10
**Issue description:**
If you export variady read the bles using tool, and change their value, it is only updated when you click another node and the click back to the changed node.
**Steps to reproduce:**
* Create 2 nodes: A and B
* At B attach a script (can not be built-in)
* code this:
```
tool
extends Node
export var a:int = 0
export var update:bool = false
func process(delta):
if update:
a = int(rand_range(0,100))
update = false
```
* Save scene close it and open again to take effects
* Value is changed, you cant see
* Click Node A
* Click Back to Node B
* Should see the change
**Minimal reproduction project:**
N/A
|
enhancement,topic:editor,documentation
|
low
|
Major
|
591,180,300 |
pytorch
|
UserRRef should store error if it sees any and prevent subsequent usage
|
@jjlilley brought this up in #35331. Currently, if `UserRRef` encountered an error internally or when running user-facing APIs, it would report the error but nothing prevents the application from using the `UserRRef` again. However, as some error could have messed up RRef internal states, especially for the ones that have impact on reference counting, we should store those error in the UserRRef, delete the UserRRef and prevent subsequent usage on the RRef.
cc @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @gqchen @aazzolini @rohan-varma @xush6528 @jjlilley @osalpekar
|
feature,triaged,module: rpc
|
low
|
Critical
|
591,200,979 |
pytorch
|
how to do 3d data augmentation in parallel on the gpu?
|
I have a lot of 3d data and need to do various data augmentation. I want to do data augmentation in parallel on the gpu, but it seems that pytorch does not allow gpu operation in the dataloader. Is there any good way?
cc @ngimel @SsnL
|
module: dataloader,module: cuda,triaged
|
low
|
Minor
|
591,201,932 |
TypeScript
|
Add a stable module target for top-level await
|
**TypeScript Version:**
3.9-beta
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:**
top level await es2020 module
**tsconfig.json**
"target":"es2020"
"module":"es2020"
**Code**
```ts
await 1;
```
**Expected behavior:**
OK if "module":"esnext" in tsconfig.json
OK if "module":"es2020" in tsconfig.json
**Actual behavior:**
OK if "module":"esnext" in tsconfig.json
KO if "module":"es2020" in tsconfig.json <----- :'(
`Top-level 'await' expressions are only allowed when the 'module' option is set to 'esnext' or 'system', and the 'target' option is set to 'es2017' or higher. ts(1378)`
**Playground Link:**
"module":"es2020" is not available on Typescript playground.
"module":"es2020" is not available in documentation.
**Related Issues:**
https://github.com/microsoft/TypeScript/pull/35813 - Add top-level await for esnext and system modules
Basically, just add "es2020" to the list.
Easy to fix, bro.
|
Needs Investigation
|
low
|
Minor
|
591,228,451 |
godot
|
Drag Drop on TextureButton does not pass on to parent container
|
**Godot version:**
3.2.1
**OS/device including version:**
Win 10
**Issue description:**
Drag and drop events are not passed from a TextureBbutton to a parent container
**Minimal reproduction project:**
[DragDropExample.zip](https://github.com/godotengine/godot/files/4410145/DragDropExample.zip)
In the example i provided there are three gui's that i have made with slightly different scripts and or nodes. The basic idea is that you can drag an image from the left to the gui on the right and then it will replace the icon in the center.
The top one is the one i would expect to be working, but it does not. In this situation the can_drop is attached to the container of the gui. It would therefore expect it to be droppable on the entire surface, but it is not droppable on the texturebuttons inside the container.
The middle example is my current compromise. In this situation the can_drop is attached to the central textureButton. So it is atleast possible to drag and drop on a texturebutton.
The third example has again the can_drop attached to the container, but now the central icon has been replaced with a texturerect, This case it is possible to drop on the rect and the surrounding areas, but still not on the little buttons.
This seems to me inconsistent and for me its also a big inconvenience as i would like to keep all of the nodes as TextureButtons.
|
bug,topic:gui
|
low
|
Minor
|
591,240,897 |
opencv
|
OpenCV 4.2.0 installation for Python 3 on Ubuntu 16.04.04 / ARM
|
I am trying to install OpenCV for further usage with both Python 2.7 and Python 3.5 on Ubuntu 16.04.4 running on OrangePI. It compiles succesfully, but installation directory of `cv2-xxx.so` is incorrect in case of Python 3.5.
Configuration:
- OpenCV => 4.2.0
- Operating System / Platform => Ubuntu 16.04.4 / ARM
- Compiler => GCC 5.4.0
Steps:
- (while in root source directory) `mkdir build && cd build`
- `cmake -DCMAKE_BUILD_TYPE=Release -DBUILD_NEW_PYTHON_SUPPORT=ON -DINSTALL_PYTHON_EXAMPLES=ON -DBUILD_EXAMPLES=ON -DINSTALL_C_EXAMPLES=ON ..`
The output of last command for Python 3 looks like:
```
-- Python 3:
-- Interpreter: /usr/bin/python3 (ver 3.5.2)
-- Libraries: /usr/lib/arm-linux-gnueabihf/libpython3.5m.so (ver 3.5.2)
-- numpy: /usr/lib/python3/dist-packages/numpy/core/include (ver 1.11.0)
-- install path: lib/python3.5/dist-packages/cv2/python-3.5
```
The install path is incorrect as by default Python 3.5 tries to search for SO library in a `/usr/lib/python3/dist-packages` directory:
```
$ python3.5 -c "from distutils.sysconfig import *; print(get_python_lib())"
/usr/lib/python3/dist-packages
```
Thus, `import cv2` can't load SO library after `make install`.
The full output of `pythom3 -m site` looks like that:
```
sys.path = [
'/home/orangepi/python3/opencv-4.2.0/build',
'/usr/lib/python35.zip',
'/usr/lib/python3.5',
'/usr/lib/python3.5/plat-arm-linux-gnueabihf',
'/usr/lib/python3.5/lib-dynload',
'/usr/local/lib/python3.5/dist-packages',
'/usr/lib/python3/dist-packages',
]
USER_BASE: '/home/orangepi/.local' (doesn't exist)
USER_SITE: '/home/orangepi/.local/lib/python3.5/site-packages' (doesn't exist)
ENABLE_USER_SITE: True
```
|
question (invalid tracker)
|
low
|
Minor
|
591,241,345 |
rust
|
No error code for "union patterns should have exactly one field"
|
It would be nice to have a specific error code for `union patterns should have exactly one field` [error](https://github.com/rust-lang/rust-gha/blob/master/src/librustc_typeck/check/pat.rs#L948). Currently, there is [no such code](https://doc.rust-lang.org/error-index.html).
|
C-enhancement,A-diagnostics,T-compiler
|
low
|
Critical
|
591,292,532 |
opencv
|
videoio: MSMF: Doesn't save convertFormat property, no way to read raw video
|
##### System information (version)
- OpenCV => 4.2, 4.3-pre (master)
- Operating System / Platform => Windows 10 64 Bit
- Compiler => Visual Studio 2019
##### Detailed description
I am trying to read a video from an uncompressed grayscale 8-bit AVI file without changing the color format. In OpenCV 3.4 this was easily possible. In OpenCV 4.0 this was possible but I had to set CAP_PROP_CONVERT_RGB to false and reshape the frames because they would be read as 1 x size instead of width x height. In OpenCV 4.2, it looks like CAP_PROP_CONVERT_RGB isn't stored at all.
I think this was broken in 6c8f50d1f3f99
##### Steps to reproduce
```.cpp
cv::VideoCapture cap("testfile.avi", cv::CAP_MSMF);
std::cout << cap.get(cv::CAP_PROP_CONVERT_RGB) << std::endl; // Prints 1
cap.set(cv::CAP_PROP_CONVERT_RGB, false);
std::cout << cap.get(cv::CAP_PROP_CONVERT_RGB) << std::endl; // Prints 1
cv::Mat frame;
cap >> frame;
// prints 760x760 ch: 3 for a 760x760 grayscale image
std::cout << "raw: " << frame.cols << "x" << frame.rows << " ch: " << frame.channels();
```
##### Issue submission checklist
- [ ] I report the issue, it's not a question
<!--
OpenCV team works with answers.opencv.org, Stack Overflow and other communities
to discuss problems. Tickets with question without real issue statement will be
closed.
-->
- [ ] I checked the problem with documentation, FAQ, open issues,
answers.opencv.org, Stack Overflow, etc and have not found solution
<!--
Places to check:
* OpenCV documentation: https://docs.opencv.org
* FAQ page: https://github.com/opencv/opencv/wiki/FAQ
* OpenCV forum: https://answers.opencv.org
* OpenCV issue tracker: https://github.com/opencv/opencv/issues?q=is%3Aissue
* Stack Overflow branch: https://stackoverflow.com/questions/tagged/opencv
-->
- [ ] I updated to latest OpenCV version and the issue is still there
<!--
master branch for OpenCV 4.x and 3.4 branch for OpenCV 3.x releases.
OpenCV team supports only latest release for each branch.
The ticket is closed, if the problem is not reproduced with modern version.
-->
- [ ] There is reproducer code and related data files: videos, images, onnx, etc
<!--
The best reproducer -- test case for OpenCV that we can add to the library.
Recommendations for media files and binary files:
* Try to reproduce the issue with images and videos in opencv_extra repository
to reduce attachment size
* Use PNG for images, if you report some CV related bug, but not image reader
issue
* Attach the image as archite to the ticket, if you report some reader issue.
Image hosting services compress images and it breaks the repro code.
* Provide ONNX file for some public model or ONNX file with with random weights,
if you report ONNX parsing or handling issue. Architecture details diagram
from netron tool can be very useful too. See https://lutzroeder.github.io/netron/
-->
|
priority: low,category: videoio(camera),platform: win32
|
low
|
Critical
|
591,296,972 |
pytorch
|
Clarify tensor storage communication behavior in RPC
|
For now, RPC could either send tensor data or tensor **storage** data to the destination depending on the sparsity of the tensor. Say, if the tensor is just a view of the 1st row of another tensor that contains 100 rows, RPC would clone the tensor and just send the data in the 1st row. If, however, if the tensor is a view of the first 60 rows, RPC would send the entire storage. This is an internal optimization to balance memory overhead vs comm overhead. As @lw pointed out in [this comment](https://github.com/pytorch/pytorch/pull/35483#discussion_r400196335), this will have user-visible differences and it's important that we have a clear design on this behavior and clarify that in our API docs.
cc @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @gqchen @aazzolini @rohan-varma @xush6528 @jjlilley @osalpekar
|
module: docs,feature,triaged,module: rpc
|
low
|
Minor
|
591,321,890 |
angular
|
Angular binding DOM property instead of web component property
|
### Description
Currently when we use web components in angular 8, with module schema set to CUSTOM_ELEMENTS_SCHEMA. Angular sets and update DOM properties, instead of web component properties when property name is similar to a attribute which has a corresponding DOM property. For example:
If the web component has exposed a property with name readonly, when we bind this property in our template like:
`<my-custom-element [readonly]="readonly"></my-custom-element>`
Instead of updating and binding with readonly prop of the web component. Angular binds it with readOnly property of the dom. Also this readOnly property is then also available on the web component reference. But my web component readonly property is always undefined and is never set.
## 🔬 Minimal Reproduction
As displayed in the below stackblitz, even on changing the readonly property of the web component the input is never made readonly, and the web component setter is never called. Also you can observe that readonly property of web component is always undefined wheres as if you display readOnly property of DOM is getting updated
https://stackblitz.com/edit/angular-rlntip
|
type: bug/fix,freq1: low,area: core,state: confirmed,core: binding & interpolation,cross-cutting: custom elements,P3
|
low
|
Minor
|
591,329,327 |
pytorch
|
BatchNormFuncOptions object cant be printed in C++
|
Repro:
```
auto options = F::BatchNormFuncOptions().weight(w).bias(b).momentum(m).eps(e).training(t);
std::cout << options << std::endl;
```
Expected result:
Print value for each property.
Actual Result:
```
‘torch::nn::functional::BatchNormFuncOptions’ is not derived from ‘const torch::ExpandingArray<D, T>’
std::cout << options << std::endl;
^~~~~~~
```
cc @yf225
|
module: cpp,module: nn,triaged
|
low
|
Minor
|
591,349,194 |
godot
|
OS.alert() called multiple times when used in a FileDialog file selection signal.
|
**Godot version:**
3.2.1 Official
**OS/device including version:**
Windows 10 64-bit
Not sure yet if this is exclusive to Windows, perhaps others can confirm or deny on Linux/Mac.
**Issue description:**
I've found when double clicking a FileDialog file to open it, if there is an OS.alert() in that signal code, it will spam the alert().
Screenshots are an attempt to show the next window that appeared after closing the first. They stagger upwards.
**Steps to reproduce:**
Run the demo project and double click the icon.png file.
**Minimal reproduction project:**
[File Path Signal Spam OS.Alert.zip](https://github.com/godotengine/godot/files/4411054/File.Path.Signal.Spam.OS.Alert.zip)
|
bug,topic:porting
|
low
|
Major
|
591,349,341 |
vscode
|
Search for ignored sync settings or extensions
|
Testing https://github.com/microsoft/vscode/issues/93801
Changing extensions and settings that have been ignored to being synced again was smooth, but could we making searching for which extensions and settings easier? Like being able to search for "@ sync: ignored" in extensions view or in settings search
|
feature-request,settings-editor
|
low
|
Major
|
591,352,427 |
TypeScript
|
Partial Semantic Editing Operations
|
# Background
We've heard from many TypeScript and JavaScript developers that large codebases suffer from massive startup delays in the editor. This is often due to program construction time. Before any semantic operation can be triggered, TypeScript has to effectively do the following:
1. Find out which project a file belongs to
1. Load that project
1. loop: Read each of the given files
1. Parse the file
1. Look through the imports/exports and resolve each import path
1. If new files were found, add them to the program's file list and go to `loop`.
All this is unfortunately synchronous today, and even if it wasn't, this might still end up being a lot of work for larger projects. For users with many files and non-trivial resolution schemes, we've witnessed program construction times that have taken *minutes*.
This can be frustrating for lots of users whose first tasks might just include navigation and some basic editing.
## Basic Code Navigation
Let's take the following code and assume a user requests to go-to-definition on the `import` statement:
```ts
import * as foo from "./foo";
```
If TypeScript is still loading the program, it can't answer where `./foo` resolves to.
But it's a relative path, and for most users this feels like something so obvious and cheap to compute. Most users could guess which file it points to be right. If there was a lightweight way to say "it should probably take you here" until a program is fully loaded up, it might get 99% of the job done for 99% of people. If there was a way to guess 2-3 locations and give the editor a list of locations to look up, that might just solve the issue.
## Basic Completions
Let's take the following and assume the user requests completions in each of the comments:
```ts
import { SomeClass } from "some-module";
function foo() {
function bar() {
b/*inBar*/
}
function baz() {
SomeClass./*afterDot*/
}
new /*afterNew*/
}
class CurrentClass {
/*inClass*/
}
```
Completions at `/*inBar*/` and `/*afterNew*/` technically need a full program to tell us what globals are available, but a lot of the time users might be able to get by with locals until the program is fully loaded - after all, this was the basic experience in lots of editors before they had TypeScript support. We could still build up the list of local symbols and list those as the highest-priority completions, followed by a list of local identifiers.
When it comes to `/*afterDot*/` something like what's proposed in https://github.com/Microsoft/TypeScript/issues/5334 might be good enough - just provide a list of local identifiers.
A completion at `/*inClass*/` could limit itself to contextual keywords (e.g. `public`, `private`, `protected`, `readonly`, `abstract`, whatever) and that would still be useful.
# Proposal
Some editors spin up two TS Server instances: one which looks for full semantic information, and one which only answers file-local syntactic questions. So one idea we've had is: could we give the syntactic server some basic heuristics to provide answers to editor questions that appear obvious, or where incomplete answers are good enough?
The proposal is to enable syntax-only instances of TypeScript's language service to answer a set of semantic-like questions that includes
* Go-to-Definition on
* relative paths
* locally declared functions and variables
* Completions based on
* Locally scoped identifiers
* Keywords based on different syntactic contexts
## Go to Definition
### On Imports to Relative Paths
When it comes to relative paths, there's at least the following resolution scheme:
* If it has an extension
* remove it, try to resolve by appending `.ts`, `.tsx`, `.d.ts`, `.js`, `.jsx`
* try loading the file by that path (e.g. `.json` files)
* Try appending `.ts`, `.tsx`, `.d.ts`, `.js`, `.jsx` to the current path
TypeScript can potentially provide an ordered list of lookup locations back to the editor, or just guess at one. The editor can try all of these, wait, and return the first "preferred" definition, but we'd need to sync with editor teams to see how feasible this would be.
Alternatively, the syntax server can perform the work, but there are some complications. See the open questions below.
### On Imported Bindings
In
```ts
import { foo } from "./foo";
/**/foo();
```
It might be okay for go-to-definition to be the same as `go-to-definition` for `./foo` itself if it's prohibitively expensive to load/parse/bind a file on the fly (though we're going to do that once it's opened in the editor anyway).
### On Local Bindings
TypeScript can potentially bind each source file on demand, and walk up the AST to resolve the requested symbol, and jump to the definition. This means that go-to-definition definitely wouldn't work on globals across files, and it might be able to jump to the correct file of certain imports, but it would at least definitely work within the current file.
## Completions
Completions could work by providing the list of contextual keywords based on the current AST (which we already have at least in part), along with a list of either file-local identifier completions or whatever's in the current chain of file-local symbol tables.
## Editor Usage
Editors could use this new functionality by changing certain commands to be multi-step commands: see if the semantic language service is in the middle of loading a project first, use the syntactic version of the command, and then try to fire semantic a syntax command if that fails (e.g. for go-to-definition, none of the operations worked).
Alternatively, it can have both operations race against each other. Try both concurrently, see which completes first, and then either cancel the operation the semantic server if it loses or throw away the results of the syntax server if it loses.
# Prior Art
## Current Architecture
The syntax server existing in the first place gives some legitimacy to this approach - a server for answering questions that doesn't know the entire state of the world.
## Roslyn
Roslyn has some of the same concepts but with a different approach. Roslyn makes a lot of their architecture asynchronous and lazy, and if results aren't currently available, they'll provide an incomplete experience - for example, completion lists aren't always available for the entire program. The approach is drastically different from what's being proposed here, but it's fundamentally the same idea: work with what you've got until you've got it all.
# Open Questions
## Can we return locations that don't exist in navigation?
We mention above the idea of *guessing* file locations - but those files might not exist on disk! Some editors might not expect an invalid location to be returned. We'll have to chat with editor teams to understand whether this is actually a problem or not. Without an opt-in flag, it might be.
## Does go-to-definition need to return a ranked list of files to try?
Today, go-to-definition just returns a list of locations. In this version of go-to-definition, we might want to provide a list of potential lookup locations for the editor - but they're not all equally weighted. You don't want to provide the `.js` file in the same list of definitions as the `.ts` file. It seems like this version of go-to-definition needs to be slightly altered from the existing version.
## Can the server just resolve the files itself instead returning a ranked list of files?
Technically it can but TSServer is synchronous. Maybe for a max of 5-10 file existence checks this wouldn't be too bad, but we should figure out if it's better for editors to decide.
## Do we need new TSServer commands?
Maybe! Technically, we can have existing TS Server commands respond differently between the syntax server and the semantic server, but to be honest, I think we'd be better off not doing this. Some of this new functionality might need opt-in flags, so I think it would be better to make this entire experience opt-in via a separate language server command.
## Is TypeScript providing its own list of completions actually better than most editors' naive implementation of completions?
It's not entirely clear whether TypeScript's "look at every identifier in the file" approach is actually going to be better than what an editor like VS Code does in a plaintext file. We should understand what they do for plaintext and see if there's actually anything we can provide that's better.
## Should editors ever fall back to sytnax operations once the semantic server is done?
Operations like *go-to-definition* can often trigger new project loads. This will block the semantic server again. It feels like using the syntax server is the right call, but this is another place where "not everything works perfectly" might be confusing.
## Would this experience too different/incomplete?
It might be very jarring if a user got drastically different results from a normal editing session. Hopefully visual indicators from the editor can give a hint that results will be slightly incomplete, but a big goal here is to capture as much of the current experience in a lightweight way.
|
Suggestion,In Discussion,Domain: Completion Lists,Domain: Symbol Navigation
|
medium
|
Major
|
591,374,253 |
flutter
|
AsyncSnapshot should support null as data
|
Dart's `Stream`s can contain `null` as data and listeners get still notified.
With Flutter's `StreamBuilder` though, users have no way of finding out which of the following states an `AsyncSnapshot` with `null` as `data` represents:
- A `Stream` that contains `null` as data.
- A `Stream` without an event.
## Use cases
In some cases, you might want to distinguish between both states, representing "there is no data" vs "the data is still loading".
## Proposal
Maybe `AsyncSnapshot` could support nullable data by having an extra field indicating whether the data comes from the `Stream`?
|
framework,c: proposal,P2,team-framework,triaged-framework
|
low
|
Minor
|
591,380,256 |
TypeScript
|
VSCode ignores JSDoc's @property when specify @typedef type in 1.43.2 and 1.44.0-insider
|
<!-- ⚠️⚠️ Do Not Delete This! bug_report_template ⚠️⚠️ -->
<!-- Please read our Rules of Conduct: https://opensource.microsoft.com/codeofconduct/ -->
<!-- Please search existing issues to avoid creating duplicates. -->
<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: https://code.visualstudio.com/insiders/ -->
Seems like VSCode ignores `@property` when specify type inside `@typedef` in 1.43.2 and 1.44.0-insider. It cause type hint to show not so correct type information and also not detected inside autocomplete/intellisense.
<!-- Use Help > Report Issue to prefill these. -->
- VSCode Version: 1.43.2 and 1.44.0-insider
- OS Version: Ubuntu 18.04 x86_64
Steps to Reproduce:
1. Create JSDoc type definition file, for example `types.js`.
2. Import the type into `form.js`.
3. Ensure the type hint for the specified JSDoc is working properly (no `any`) and the type is detected available inside autocomplete/intellisense.
Temporary Solution:
- Remove type from `@typedef`, such us from `@typedef {Object<string, any>} Person` into `@typedef Person`.
- Or, remove `<string, any>` from Object, such us `@typedef {Object<string, any>} Person` into `@typedef {Object} Person`.
Reproduce Repo: https://github.com/jeffryang24/sandbox/tree/master/javascript/vscode-jsdoc-type-hint-issue
<!-- Launch with `code --disable-extensions` to check. -->
Does this issue occur when all extensions are disabled?: Yes/~~No~~
|
Needs Investigation
|
low
|
Critical
|
591,401,031 |
nvm
|
Ability to configure what rc/profile file nvm installs into
|
<!-- Thank you for being interested in nvm! Please help us by filling out the following form if you‘re having trouble. If you have a feature request, or some other question, please feel free to clear out the form. Thanks! -->
#### Operating system and version:
Debian/unstable.
#### How did you install `nvm`?
<!-- (e.g. install script in readme, Homebrew) -->
Ran the installer.
#### What steps did you perform?
Installed with the standard install script. Tried to use ansible.
#### What happened?
#2174. Ansible started a non-interactive login shell & did not start nvm.
#### What did you expect to happen?
I would have liked some way to, on install, control how & where nvm installed itself into my shell and/or shells. A supplemental environment variable for the installer that let me specify that I prefer `.zshenv`/`.bash_profile` over the significantly less used/less useful `.zshrc`/`.bashrc` files would have given me an install preference to install nvm in a way that would be useful in more situations, available to more systems, such as Ansible.
This change by nvm would go a long way towards helping users such as those in #2174 make more active use of nvm wherever they need it.
#### Is there anything in any of your profile files that modifies the `PATH`?
Oh heck yes tons of things.
|
feature requests,installing nvm
|
low
|
Minor
|
591,404,717 |
PowerToys
|
[Run][WindowWalker Plugin] Search parent window names
|
# Summary of the new feature/enhancement
<!--
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).
-->
Trying it out for the first time, I wanted to switch to 'RoyalTS'. The issue is that depending on what tab I have selected, the window title is not actually 'RoyalTS'. In this case it was 'Single Box' so I wasn't able to find it.
# Proposed technical implementation details (optional)
When searching, search parent window properties as well:
<!--
A clear and concise description of what you want to happen.
-->
Able to select windows by their parent window name or other metadata details.
|
Idea-Enhancement,Product-PowerToys Run,Run-Plugin
|
low
|
Minor
|
591,414,091 |
pytorch
|
Unexpected behaviour for affine_grid and grid_sample with 3D inputs
|
## 🐛 Bug
When performing affine transformations on 3D tensors I’ve noticed some unexpected behaviour which does not occur for 2D tensors.
Specifically, suppose we perform two 3D rotations with two rotation matrices `A` and `B`. Given a 3D input tensor `x`, we can first rotate it with `B` to get the rotated tensor `x_B`. We can then further rotate `x_B` by `A` to get the rotation corresponding to `B` followed by `A`, i.e. to the rotation matrix `AB`. We can denote this resulting tensor by `x_AB`.
Alternatively, we could have directly taken the matrix product `C = AB` and transformed the tensor with `C`, to obtain `x_C`. Now `x_AB` and `x_C` should be approximately equal up to sampling errors. However in practice this is not the case. As an example:
Clearly, the cube rotated by `B` followed by `A` is not the same as the one rotated by `C=AB`. Note that this does not happen in 2D. If we perform similar transformations to a 2D input, we obtain:
As can be seen, the transformations by `AB` and `C` are roughly the same (`AB` has more sampling noise since it corresponds to two affine transformations and two grid_samples), so the error does not occur in 2D.
Any idea why the difference in 3D might be occurring? Thank you!
## To Reproduce
```python
import torch
def get_3d_cube(grid_size=32, cube_size=16):
"""Returns a 3d cube which is useful for debugging and visualizing 3d
rotations.
Args:
grid_size (int): Size of sides of 3d grid.
cube_size (int): Size of sides of cube within grid.
"""
cube = torch.zeros(1, 1, grid_size, grid_size, grid_size)
hs = grid_size // 2 # Half grid size
hcs = cube_size // 2 # Half cube size
cube[:, :, -hcs + hs:hs + hcs, hs + hcs, hs + hcs] = 1
cube[:, :, -hcs + hs:hs + hcs, hs + hcs, hs - hcs] = 1
cube[:, :, -hcs + hs:hs + hcs, hs - hcs, hs + hcs] = 1
cube[:, :, -hcs + hs:hs + hcs, hs - hcs, hs - hcs] = 1
cube[:, :, hs + hcs, -hcs + hs:hs + hcs, hs + hcs] = 1
cube[:, :, hs - hcs, -hcs + hs:hs + hcs, hs + hcs] = 1
cube[:, :, hs + hcs, -hcs + hs:hs + hcs, hs - hcs] = 1
cube[:, :, hs - hcs, -hcs + hs:hs + hcs, hs - hcs] = 1
cube[:, :, hs + hcs, hs + hcs, -hcs + hs:hs + hcs] = 1
cube[:, :, hs - hcs, hs + hcs, -hcs + hs:hs + hcs] = 1
cube[:, :, hs + hcs, hs - hcs, -hcs + hs:hs + hcs] = 1
cube[:, :, hs - hcs, hs - hcs, -hcs + hs:hs + hcs] = 1
return cube
# Create a cube to use for debugging
cube = get_cube_3d(128, 64)
# 3D rotation matrices with shape (1, 3, 3)
rotation_A = torch.tensor([[[ 0.7198, -0.6428, -0.2620],
[ 0.5266, 0.7516, -0.3973],
[ 0.4523, 0.1480, 0.8795]]])
rotation_B = torch.tensor([[[ 0.6428, 0.0000, 0.7660],
[ 0.0000, 1.0000, 0.0000],
[-0.7660, 0.0000, 0.6428]]])
# C = AB, so C is a rotation by B, followed by a rotation by A
rotation_C = rotation_A @ rotation_B
# Convert to affine matrices of shape (1, 3, 4), where the 4th column
# corresponds to a zero translation
translations = torch.zeros(1, 3, 1)
affine_A = torch.cat([rotation_A, translations], dim=2)
affine_B = torch.cat([rotation_B, translations], dim=2)
affine_C = torch.cat([rotation_C, translations], dim=2)
# Get grid for each affine transform
grid_A = torch.nn.functional.affine_grid(affine_A, size=cube.shape)
grid_B = torch.nn.functional.affine_grid(affine_B, size=cube.shape)
grid_C = torch.nn.functional.affine_grid(affine_C, size=cube.shape)
# Rotate cube by B
cube_B = torch.nn.functional.grid_sample(cube, grid_B, mode="bilinear")
# Rotate resulting cube by A, to obtain AB transformation
cube_AB = torch.nn.functional.grid_sample(cube_B, grid_A, mode="bilinear")
# Rotate original cube by C
cube_C = torch.nn.functional.grid_sample(cube, grid_C, mode="bilinear")
print((cube_AB - cube_C).abs().sum()) # This is large
```
## Expected behavior
Rotating a tensor by `B` followed by `A` should be approximately the same as rotating by `AB`.
## Environment
- PyTorch Version (e.g., 1.0): 1.4.0
- OS (e.g., Linux): macOS Catalina
- How you installed PyTorch (`conda`, `pip`, source): pip
- Build command you used (if compiling from source): N/A
- Python version: 3.6.8
- CUDA/cuDNN version: N/A
- GPU models and configuration: N/A
- Any other relevant information: N/A
cc @albanD @mruberry
|
module: numerical-stability,module: nn,triaged
|
low
|
Critical
|
591,416,511 |
kubernetes
|
Propagate context deadline in TokenReview REST handler
|
The TokenReview REST handler constructs a new request and passes it through the authentication handler. Any deadline associated with the original context is not passed down.
We should propagate the original context deadline/cancellation, either by wrapping the parent context, or by extracting/propagating the deadline, or by waiting on the original context's Done() channel and propagating a cancellation
/sig auth
/cc @enj @mikedanese
|
kind/bug,priority/important-soon,sig/auth,lifecycle/frozen
|
low
|
Minor
|
591,421,844 |
TypeScript
|
Function expressions in a property assignment of a prototype object should be methods, not nested classes
|
```js
/** @class */
module.exports.C = function() {
this.x = 1
}
module.exports.C.prototype = {
m1() {
this.a = 1
},
m2: function() {
this.b = 2
}
}
var c = new module.exports.C()
c.a
c.b
```
**Expected behavior:**
`c.a` and `c.b` both work.
**Actual behavior:**
Only `c.a` works; `c.b` says that 'd' does not exist on type 'C'. There's also an error on `this.b = 2` when noImplicitThis: true.
bindThisPropertyAssignment needs to understand that function expressions might be part of a property assignment in an object literal. The code to handle this will probably look like this:
```ts
// For `{ x: function() }` should modify the object literal's members, not behave like a fresh class (as long as it doesn't have @class on it)
if (isPropertyAssignment(thisContainer.parent)) {
// wow I hope those parent pointers are set!
const containingClass = thisContainer.parent.parent;
const symbolTable = containingClass.symbol.members!;
if (hasDynamicName(node)) {
bindDynamicallyNamedThisPropertyAssignment(node, containingClass.symbol);
}
else {
declareSymbol(symbolTable, containingClass.symbol, node, SymbolFlags.Property | SymbolFlags.Assignment, SymbolFlags.None, /*isReplaceableByMethod*/ true);
}
break;
}
```
I expect this to break a lot of tests, but it didn't. However, I haven't tried it on the user tests yet.
|
Bug
|
low
|
Critical
|
591,422,729 |
flutter
|
[web] add chrome android support for web installers repo
|
[web_drivers](https://github.com/flutter/web_installers/tree/master/packages/web_drivers) libraries should be able to, more info on: https://chromedriver.chromium.org/getting-started/getting-started---android
- [ ] create an emulator
- [ ] run the emulator
- [ ] run the bridge (adb)
- [ ] find the version of chrome on the emulator and install the driver.
|
a: tests,framework,platform-web,c: proposal,P2,team-web,triaged-web
|
low
|
Minor
|
591,440,380 |
flutter
|
Handle the UITouchPhaseRegion enum values.
|
The following enum values were added to `UITouchPhase` in iOS 13.4:
* `UITouchPhaseRegionEntered`
* `UITouchPhaseRegionExited`
* `UITouchPhaseRegionMoved`
These [don't seem to be documented](https://developer.apple.com/documentation/uikit/uitouchphase/uitouchphaseregionentered?language=objc) currently. This issue tracks the handling of these new enums. Xcode and/or the iOS SDK will have to be updated before these enums can be used in the engine as well.
|
platform-ios,engine,t: xcode,P2,team-ios,triaged-ios
|
low
|
Minor
|
591,442,658 |
flutter
|
Handle the UITouchTypeIndirectPointer enum value.
|
The following enum value was added to the `UITouchType` enum in iOS 13.4:
* `UITouchTypeIndirectPointer`
This [doesn't seem to be documented](https://developer.apple.com/documentation/uikit/uitouchtype/uitouchtypeindirectpointer?language=objc) currently. This issue tracks the handling of this new enum value. Xcode and/or the iOS SDK will have to be updated before this enum can be used in the engine as well.
|
platform-ios,engine,t: xcode,P2,team-ios,triaged-ios
|
low
|
Minor
|
591,449,019 |
godot
|
Light2D-tracking code (also occluders, likely others) depends on undefined behavior
|
**Godot version:** master`
**OS/device including version:** n/a
**Issue description:**
While trying to track down some strange behavior involving the interaction of a moving light and a stationary one, I discovered that the light-tracking code unwittingly relies on undefined behavior -- i.e. it performs operations that are specifically forbidden by well-formed C++ code.
1. When a light is added to the canvas, it is first allocated using `canvas_light_create` (https://github.com/godotengine/godot/blob/master/servers/rendering/rendering_server_canvas.cpp#L1007) which ultimately calls `new`.
2. Lights are tracked in the `lights` member of a `struct Canvas` (https://github.com/godotengine/godot/blob/master/servers/rendering/rendering_server_canvas.h). This is a `Set<RasterizerCanvas::Light *>` -- a set of "pointer to RasterizierCanvas::Light".
3. The `Set` class (https://github.com/godotengine/godot/blob/master/core/set.h#L40) is based on a binary tree.
4. The default comparison operator for `Set` is `Comparator<T>` (defined here: https://github.com/godotengine/godot/blob/master/core/typedefs.h#L218 )
```
template <class T>
struct Comparator {
_ALWAYS_INLINE_ bool operator()(const T &p_a, const T &p_b) const { return (p_a < p_b); }
};
```
5. Since `T` here is `RasterizerCanvas::Light *`, and there are no template specializations for `Comparator`, this boils down to direct pointer comparison.
However, in C and C++, the results of comparing two pointers using `<` or `>` is *undefined* unless both pointers come from the same allocation (i.e. array declaration, `malloc`, or `new`). There are some special exceptions, but none of them apply here.
The most significant consequence of this is that, if the compiler's optimizer becomes aggressive enough, it could make some assumptions that cause the application to break in highly surprising ways:
- That the only objects ever passed into the Set were allocated independently. In that case, the only defined comparison that will ever occur is a light with itself; therefore, the comparison `less(a, b)` can simply always be false, which would cause some very interesting effects.
- That the first comparison check, if any, will always use '<' to compare two unrelated (i.e. separately allocated) pointers. Since that's undefined behavior, the compiler can assume that no comparisons will ever be made (the resulting program is likely to crash immediately on startup, or it will never add more than one light, or maybe it won't ever add any lights to the canvas at all.)
I'll be honest, there's not a lot of good, simple ways to fix this.
- The ideal method would likely be to have the Set be of _references_, not pointers; the mistake would have produced a compile-time error.
However, you can't modify a reference, so the element allocation would have to be adjusted to call a non-default constructor. Plus, of course, all the code that uses `Set` would have to be modified accordingly.
- Switching to a hash table would work, although, unless you're willing to take a dependency on the standard library, I don't see a portable way to produce a hashcode from a pointer.
The least-invasive-prone way I can think to fix this would be to do:
```
template<class T>
struct RidComparator {
_ALWAYS_INLINE_ bool operator()(const T &p_a, const T &p_b) const {
return (p_a->primary_rid() < p_b->primary_rid());
}
};
template <class T, class C = RidComparator<T>, class A = DefaultAllocator>
class RidSet : public Set<T, C, A>
{
};
```
then and stick a `primary_rid` method on `Light` and `LightOccluderInstance` and whatnot, returning the RID that uniquely identifies the object in question. However, while I'm certain that doing this would _work_, I don't know if there's a better way.
**Steps to reproduce:** n/a
**Minimal reproduction project:** n/a
|
bug,topic:core,topic:rendering,topic:2d
|
low
|
Critical
|
591,490,122 |
vscode
|
Confusion about how EnvironmentVariableCollection handles multiple extensions changing the same variable
|
Refs #93601
**Repro**
1. In Extension 1, create a persisted `EnvironmentVariableCollection` that calls `.replace('TEST', 'replace')`
1. In extension 2, create a non persisted `EnvironmentVariableCollection` that calls `.append('TEST', 'append')`
1. Install both extensions
1. Activate extension 1.
1. Then restart VS Code
1. Now activate extension 2
1. Create a new terminal and check the value of `TEST`
**Expected**
I expected `TEST` to be `replaceappend`
This is because I expect the initial value of `TEST` to be `replace` from the persisted collection in extension 1, then when extension 2 is activated it appends on its value.
**Actual**
The value of `TEST` is `replace`
|
bug,terminal-process
|
low
|
Minor
|
591,530,337 |
pytorch
|
[jit] Structural typing shouldn't work
|
This example runs but should throw an error since `SomethingElse` and `Point` should be different types
```python
import torch
@torch.jit.script
class Point(object):
def __init__(self, x: int, y: int):
self.x = x
self.y = y
def some_method(self):
return self.x + self.y
class SomethingElse(object):
def __init__(self):
self.x = 1
self.y = 1
@torch.jit.script
def fn(x: Point):
print(x)
return x
# Not passing in a Point
fn(SomethingElse())
```
cc @suo
|
oncall: jit,triaged
|
low
|
Critical
|
591,591,650 |
pytorch
|
Python 3 build PyTorch got core dump
|
## ❓ Questions and Help
### Please note that this issue tracker is not a help form and this issue will be closed.
We have a set of [listed resources available on the website](https://pytorch.org/resources). Our primary means of support is our discussion forum:
- [Discussion Forum](https://discuss.pytorch.org/)
|
needs reproduction,module: crash,triaged
|
low
|
Minor
|
591,630,428 |
flutter
|
Patch ICU to not violate -W-implicit-int-float-conversion
|
Specifically, https://fuchsia.googlesource.com/third_party/icu/+/refs/heads/master/source/i18n/plurrule.cpp#1664
/cc @chinmaygarde
|
engine,P2,team-engine,triaged-engine
|
low
|
Minor
|
591,631,330 |
flutter
|
Patch wuffs to not violate -Wnon-c-typedef-for-linkage"
|
Multiple places in //third_party/wuffs/release/c/wuffs-v0.2.c
/cc @chinmaygarde
|
engine,P2,team-engine,triaged-engine
|
low
|
Minor
|
591,658,649 |
pytorch
|
Backward function causes device error in C++ when changing module's device repeatly.
|
# 🐛 Bug
<!-- A clear and concise description of what the bug is. -->
## To Reproduce
Steps to reproduce the behavior:
1. Compile & Run code below.
```C++
#include <torch/torch.h>
// Define a new Module.
struct Net : torch::nn::Module {
Net() {
// Construct and register two Linear submodules.
fc1 = register_module("fc1", torch::nn::Linear(784, 64));
fc2 = register_module("fc2", torch::nn::Linear(64, 32));
fc3 = register_module("fc3", torch::nn::Linear(32, 10));
}
// Implement the Net's algorithm.
torch::Tensor forward(torch::Tensor x) {
// Use one of many tensor manipulation functions.
x = torch::relu(fc1->forward(x.reshape({x.size(0), 784})));
x = torch::dropout(x, /*p=*/0.5, /*train=*/is_training());
x = torch::relu(fc2->forward(x));
x = torch::log_softmax(fc3->forward(x), /*dim=*/1);
return x;
}
// Use one of many "standard library" modules.
torch::nn::Linear fc1{nullptr}, fc2{nullptr}, fc3{nullptr};
};
int main() {
// Create a new Net.
auto net = std::make_shared<Net>();
// Create a multi-threaded data loader for the MNIST dataset.
auto data_loader = torch::data::make_data_loader(
torch::data::datasets::MNIST("./data").map(
torch::data::transforms::Stack<>()),
/*batch_size=*/64);
// Instantiate an SGD optimization algorithm to update our Net's parameters.
torch::optim::SGD optimizer(net->parameters(), /*lr=*/0.01);
for (size_t epoch = 1; epoch <= 10; ++epoch) {
size_t batch_index = 0;
// Iterate the data loader to yield batches from the dataset.
torch::Device device = epoch % 2 == 1 ? torch::kCUDA : torch::kCPU;
net->to(device);
std::cout << "Device of network is " << device << std::endl;
for (auto& batch : *data_loader) {
// Reset gradients.
optimizer.zero_grad();
// Execute the model on the input data.
torch::Tensor prediction = net->forward(batch.data.to(device));
// Compute a loss value to judge the prediction of our model.
torch::Tensor loss = torch::nll_loss(prediction, batch.target.to(device));
// Compute gradients of the loss w.r.t. the parameters of our model.
if(batch_index == 0) std::cout << "Device of loss is " << loss.device() << "\n";
loss.backward();
if(batch_index == 0) std::cout << "loss.backward() success" << "\n";
// Update the parameters based on the calculated gradients.
optimizer.step();
// Output the loss and checkpoint every 100 batches.
if (++batch_index % 100 == 0) {
std::cout << "Epoch: " << epoch << " | Batch: " << batch_index
<< " | Loss: " << loss.item<float>() << std::endl;
// Serialize your model periodically as a checkpoint.
torch::save(net, "net.pt");
}
}
}
}
```
and output is
```
Device of network is cuda
Device of loss is cuda:0
loss.backward() success
Epoch: 1 | Batch: 100 | Loss: 2.2953
Epoch: 1 | Batch: 200 | Loss: 2.252
Epoch: 1 | Batch: 300 | Loss: 2.22181
Epoch: 1 | Batch: 400 | Loss: 2.13327
Epoch: 1 | Batch: 500 | Loss: 2.04376
Epoch: 1 | Batch: 600 | Loss: 1.80017
Epoch: 1 | Batch: 700 | Loss: 1.65806
Epoch: 1 | Batch: 800 | Loss: 1.47708
Epoch: 1 | Batch: 900 | Loss: 1.40919
Device of network is cpu
Device of loss is cpu
terminate called after throwing an instance of 'torch::utils::FutureError'
what(): expected device cuda:0 but got device cpu (compute_types at /pytorch/aten/src/ATen/native/TensorIterator.cpp:246)
frame #0: c10::Error::Error(c10::SourceLocation, std::string const&) + 0x46 (0x7f0850510576 in /home/sonic/Downloaded_Library/libtorch/lib/libc10.so)
frame #1: at::TensorIterator::compute_types() + 0x17d4 (0x7f08426fe214 in /home/sonic/Downloaded_Library/libtorch/lib/libtorch_cpu.so)
frame #2: at::TensorIterator::build() + 0x44 (0x7f0842700104 in /home/sonic/Downloaded_Library/libtorch/lib/libtorch_cpu.so)
frame #3: at::TensorIterator::binary_op(at::Tensor&, at::Tensor const&, at::Tensor const&, bool) + 0x146 (0x7f08427007b6 in /home/sonic/Downloaded_Library/libtorch/lib/libtorch_cpu.so)
frame #4: at::native::add_out(at::Tensor&, at::Tensor const&, at::Tensor const&, c10::Scalar) + 0x3c (0x7f084242155c in /home/sonic/Downloaded_Library/libtorch/lib/libtorch_cpu.so)
frame #5: <unknown function> + 0x101717d (0x7f0805af717d in /home/sonic/Downloaded_Library/libtorch/lib/libtorch_cuda.so)
frame #6: <unknown function> + 0x2b517ea (0x7f084443e7ea in /home/sonic/Downloaded_Library/libtorch/lib/libtorch_cpu.so)
frame #7: <unknown function> + 0x2cf9c5c (0x7f08445e6c5c in /home/sonic/Downloaded_Library/libtorch/lib/libtorch_cpu.so)
frame #8: torch::autograd::AccumulateGrad::apply(std::vector<at::Tensor, std::allocator<at::Tensor> >&&) + 0xdc (0x7f08445e81bc in /home/sonic/Downloaded_Library/libtorch/lib/libtorch_cpu.so)
frame #9: <unknown function> + 0x2cf5745 (0x7f08445e2745 in /home/sonic/Downloaded_Library/libtorch/lib/libtorch_cpu.so)
frame #10: torch::autograd::Engine::evaluate_function(std::shared_ptr<torch::autograd::GraphTask>&, torch::autograd::Node*, torch::autograd::InputBuffer&) + 0x16f3 (0x7f08445dfa43 in /home/sonic/Downloaded_Library/libtorch/lib/libtorch_cpu.so)
frame #11: torch::autograd::Engine::thread_main(std::shared_ptr<torch::autograd::GraphTask> const&, bool) + 0x3d2 (0x7f08445e0822 in /home/sonic/Downloaded_Library/libtorch/lib/libtorch_cpu.so)
frame #12: torch::autograd::Engine::thread_init(int) + 0x39 (0x7f08445d8e99 in /home/sonic/Downloaded_Library/libtorch/lib/libtorch_cpu.so)
frame #13: <unknown function> + 0xc70f (0x7f085074970f in /home/sonic/Downloaded_Library/libtorch/lib/libtorch.so)
frame #14: <unknown function> + 0x76db (0x7f08416d56db in /lib/x86_64-linux-gnu/libpthread.so.0)
frame #15: clone + 0x3f (0x7f0803dec88f in /lib/x86_64-linux-gnu/libc.so.6)
Aborted (core dumped)
```
## Expected behavior
Run w/o any segmentation faults.
## Environment
PyTorch version: 1.2.0
Is debug build: No
CUDA used to build PyTorch: 10.0.130
OS: Ubuntu 18.04.4 LTS
GCC version: (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0
CMake version: version 3.10.2
Python version: 3.6
Is CUDA available: Yes
CUDA runtime version: 10.0.130
GPU models and configuration: GPU 0: GeForce RTX 2070 SUPER
Nvidia driver version: 430.50
cuDNN version: /usr/local/cuda-10.0/targets/x86_64-linux/lib/libcudnn.so.7.5.0
Versions of relevant libraries:
[pip3] numpy==1.17.1
[pip3] torch==1.2.0
[pip3] torchvision==0.4.0
[conda] Could not collect
cc @ezyang @SsnL @albanD @zou3519 @gqchen @yf225
|
module: cpp,module: autograd,triaged
|
low
|
Critical
|
591,660,456 |
pytorch
|
Dropout on sparse tensors
|
## ❓ Questions and Help
I‘d like to know if there is a dropout layer that i can use it on sparse tensor.
I will be appreciate it if you can show me the name? Just like the sparse_dropout in tensorflow!
We have a set of [listed resources available on the website](https://pytorch.org/resources). Our primary means of support is our discussion forum:
- [Discussion Forum](https://discuss.pytorch.org/)
cc @vincentqb
|
module: sparse,module: nn,triaged,enhancement
|
low
|
Minor
|
591,663,131 |
TypeScript
|
Suggestion: Reference expression type during cast
|
## Search Terms
cast reference expression type
## Suggestion
I would like to be able to reference original type of the expression during cast.
For example:
```typescript
const o = {
x: { a: 1 } as typeof & { b?: number }
}
// type of o.x is now { a: number; b?: number }
```
I am not attached to `typeof` syntax here, can be anything that makes sense.
## Use Cases
1. Defining optional properties without actually initializing them.
Current approach would require me to do something like:
```typescript
const o = {
x: { a: 1 } as { a: number; b?: number }
}
```
which is unweildy and repeats all the properties.
It is possible to do through a function, but it has an overhead and is complicated due to partial generics:
```typescript
function extend<T>(T data) {
return <TExtra>() => data as T & TExtra;
}
const o = {
x: extend({ a: 1 })<{ b?: number }>()
}
```
2. Adding indexers to object literals:
```typescript
function map1(key: string) {
return { a: 1, b: 2 }[key] || 3; // does not work due to https://github.com/microsoft/TypeScript/issues/14951
}
function map2(key: string) {
return ({ a: 1, b: 2 } as { [key: string]: typeof[keyof typeof]|undefined })[key] || 3;
}
```
## Examples
See use cases.
## 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.)
* [x] This feature would agree with the rest of [TypeScript's Design Goals](https://github.com/Microsoft/TypeScript/wiki/TypeScript-Design-Goals).
|
Suggestion,Awaiting More Feedback
|
low
|
Major
|
591,753,841 |
PowerToys
|
Feature request: stretch and resize windows
|
Three different feature:
1. Stretch windows:
Allows the user to double click (example) a window edge, that will stretch (resize) the window up to the next window adjacent to the edge clicked.
2. ~Magnet windows:~ already covered by https://github.com/microsoft/PowerToys/issues/8
~Windows will "snap" (move/jump) into place when reaching a certain distance from another window or a screen edge.~
3. Window resize:
Allows the user to alt+click and drag(example) two overlapping window edges (of two different adjacent windows) and resize them both at the same time while keeping the edges overlapping.
Included gifs from several other window management programs as examples.
All of those will allow further customization after setting windows in designated zones.
|
Idea-Enhancement,Product-Window Manager
|
low
|
Major
|
591,769,208 |
godot
|
Some InputEvents aren't received by custom viewports
|
**Godot version:**
Godot 3.2.1
**OS/device including version:**
Windows 10
**Issue description:**
Custom viewports doesn't receive some input events (like keypresses).
**Steps to reproduce:**
Add a Viewport, a TextureRect with a ViewportTexture attached to it, add a node inside of the viewport with a script using `_input` and a print statement.
**Minimal reproduction project:**
[viewport_test.zip](https://github.com/godotengine/godot/files/4414136/viewport_test.zip)
--------
Maybe there's something to do on my end to pass the input events, but i wasn't able to find anything useful in the documentation or on Google.
|
bug,documentation,topic:input
|
low
|
Major
|
591,781,699 |
rust
|
Add option to set global `no_run` for doc test
|
As this [issue](https://github.com/GuillaumeGomez/doc-comment/issues/20).
I want to guarantee all examples in external files can pass compiling, but I would not like to run them.
The main reason not to add no_run for each block is many markdown renders cannot correctly recognize a "rust,no_run" block.
So, could we add a option for `#[doc]` like `#[doc(include = "../README.md", no_run)]`?
|
T-rustdoc,C-feature-request,A-doctests
|
low
|
Minor
|
591,783,476 |
rust
|
Associated types expansion issue
|
Adding an bound on an associated type (`where T::BarOutput: TraitFoo`) causes the compiler to erroneously complain about the original bound:
`the trait bound ``T: TraitBar<u64>`` is not satisfied`
The compiler is able to expand the associated type `BazOutput` enough to generate that error, but not enough to realize it would satisfy that bound.
```rust
pub trait TraitFoo {}
pub trait TraitBar<T> {
type BarOutput;
}
pub trait TraitBaz {
type BazOutput;
}
pub trait TraitFum<T> {}
pub struct StructZot;
impl TraitBaz for StructZot {
type BazOutput = u64;
}
macro_rules! test {
(1) => {
impl<T: TraitBar<<Self as TraitBaz>::BazOutput>> TraitFum<T> for StructZot {}
};
(2) => {
impl<T: TraitBar<<StructZot as TraitBaz>::BazOutput>> TraitFum<T> for StructZot {}
};
(3) => {
impl<T: TraitBar<u64>> TraitFum<T> for StructZot {}
};
(4) => {
impl<T: TraitBar<<Self as TraitBaz>::BazOutput>> TraitFum<T> for StructZot where
T::BarOutput: TraitFoo
{
}
};
(5) => {
impl<T: TraitBar<<StructZot as TraitBaz>::BazOutput>> TraitFum<T> for StructZot where
T::BarOutput: TraitFoo
{
}
};
(6) => {
impl<T: TraitBar<u64>> TraitFum<T> for StructZot where T::BarOutput: TraitFoo {}
};
}
//test!{1} // Works
//test!{2} // Works
//test!{3} // Works
//test!{4} // Fails
test!{5} // Fails
//test!{6} // Works
```
([Playground](https://play.rust-lang.org/?version=nightly&mode=debug&edition=2018&gist=ff6ba725e845b52ec5106df7a546698b))
I expected to see all 6 versions compile cleanly.
Instead, this happened:
```
error[E0277]: the trait bound `T: TraitBar<u64>` is not satisfied
--> src/lib.rs:35:9
|
35 | / impl<T: TraitBar<<StructZot as TraitBaz>::BazOutput>> TraitFum<T> for StructZot where
36 | | T::BarOutput: TraitFoo
37 | | {
38 | | }
| |_________^ the trait `TraitBar<u64>` is not implemented for `T`
...
49 | test!{5} // Fails
| -------- in this macro invocation
|
= note: this error originates in a macro (in Nightly builds, run with -Z macro-backtrace for more info)
help: consider further restricting type parameter `T`
|
36 | T::BarOutput: TraitFoo, T: TraitBar<u64>
| ^^^^^^^^^^^^^^^^^^
```
|
A-trait-system,A-associated-items,T-compiler,C-bug,A-lazy-normalization,T-types,fixed-by-next-solver
|
low
|
Critical
|
591,794,246 |
pytorch
|
Add KFAC optimizer
|
## 🚀 Feature
Add the [Kronecker-factored Approximate Curvature (KFAC)](https://arxiv.org/abs/1503.05671) optimizer and/or [E-KFAC](https://arxiv.org/abs/1806.03884).
## Motivation
These are both approximate second order methods with promising results in terms of convergence rate. KFAC uses an block-diagonal approximation to the Fisher information matrix to approximate natural gradients. This allows for larger learning rates and thus improved convergence rate. Having a fully features built-in optimizer allows researchers and deep learning engineers to quickly iterate and test.
## Pitch
A KFAC implementation that works out of the box for all supported layers should be added to the built-in optimizers.
## Additional context
[An unmaintained Pytorch implementation](https://github.com/alecwangcq/KFAC-Pytorch)
[The standard TF implementation](https://github.com/tensorflow/kfac)
[KFAC for CNN](https://arxiv.org/abs/1602.01407)
[KFAC for RNN](https://openreview.net/forum?id=HyMTkQZAb)
cc @vincentqb
|
module: optimizer,triaged,needs research
|
low
|
Minor
|
591,806,185 |
create-react-app
|
Expose env vars both with and without the REACT_APP_ prefix
|
### 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..."
-->
A third party library I'm using leverages a certain env variable, say `XYZ`. Currently there's no way for me to supply that.
### Describe the solution you'd like
<!--
Provide a clear and concise description of what you want to happen.
-->
I realize it's a good safety measure for CRA to only pick up env vars with the prefix, and I don't object to that. However CRA could populate `process.env` with both keys i.e. both `REACT_APP_XYZ` and `XYZ`. That wouldn't violate the safety in any way, but still allows for code to expect any named env var.
|
issue: proposal,needs triage
|
low
|
Minor
|
591,825,325 |
pytorch
|
Hook on input tensor not called when using autograd.grad()
|
## 🐛 Bug
A backward hook registered on the input Tensor of the output expression is called when gradients are calculated by running `autograd.backward(output)` but not when running `autograd.grad(output, input)`.
In both cases the gradients are calculated and are the same except that `grad()` returns a tuple.
The documentation of `register_hook()` says
> The hook will be called every time a gradient with respect to the Tensor is computed.
And for `autograd.grad()`
> Computes and returns the sum of gradients of outputs w.r.t. the inputs.
So is it a bug or did I misunderstand the documentation or functionalities?
## To Reproduce
Steps to reproduce the behavior:
```python
import torch
# backward()
def hook_backward(grad):
print("backward hook called using backward()")
input = torch.Tensor([2])
input.requires_grad = True
input.register_hook(hook_backward)
output = 5 * input
torch.autograd.backward(output)
print(f"backward() output: {input.grad}")
# grad()
def hook_grad(grad):
print("backward hook called using grad()")
input = torch.Tensor([2])
input.requires_grad = True
input.register_hook(hook_grad)
output = 5 * input
grad = torch.autograd.grad(output, input)
print(f"grad() output: {grad}")
```
Output:
```
backward hook called using backward()
backward() output: tensor([5.])
grad() output: (tensor([5.]),)
```
## Expected behavior
Both backward hooks are called:
```
backward hook called using backward()
backward() output: tensor([5.])
backward hook called using grad()
grad() output: (tensor([5.]),)
```
## Environment
PyTorch version: 1.4.0
Is debug build: No
CUDA used to build PyTorch: None
OS: Microsoft Windows 10 Enterprise
GCC version: Could not collect
CMake version: Could not collect
Python version: 3.7
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] numpy==1.15.2
[conda] Could not collect
## Additional context
I wanted to post it in the forum but could not register or use github for login.
cc @ezyang @albanD @zou3519 @gqchen @pearu @nikitaved @soulitzer @Lezcano @Varal7 @SsnL
|
module: autograd,triaged,actionable
|
low
|
Critical
|
591,831,766 |
PowerToys
|
[Run] Vim (j,k) feature or other keys combinations to navigate items
|
# Summary of the new feature/enhancement
<!--
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).
-->
Many Vim users would love to have `j` and `k` keys to navigate listed items after focusing it with Tab key.
|
Idea-Enhancement,Help Wanted,Product-PowerToys Run,Area-User Interface,Area-Accessibility
|
low
|
Major
|
591,863,406 |
vscode
|
Investigate setting Electron's proxy from user setting
|
<!-- ⚠️⚠️ Do Not Delete This! bug_report_template ⚠️⚠️ -->
<!-- Please read our Rules of Conduct: https://opensource.microsoft.com/codeofconduct/ -->
<!-- Please search existing issues to avoid creating duplicates. -->
<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: https://code.visualstudio.com/insiders/ -->
<!-- Use Help > Report Issue to prefill these. -->
- VSCode Version: 1.43.2
- OS Version: Ubuntu 18.04
Steps to Reproduce:
1. Set proxy from OS, and in settings.json, but still can't browse or install extensions. Looking at log I get ERR_CONNECTION_TIMED_OUT
|
bug,proxy
|
low
|
Critical
|
591,866,589 |
flutter
|
[web] Uncaught (in promise) TypeError: Failed to fetch ( Actual Error message not shown because service worker exception is not handled)
|
In the flutter web app , if error occurs when the service worker makes a fetch call ( because the resource was not found in cache) , if the fetch call gives an error , its not handled.
It doesn't give the same error message as would have been given by the browser if we didn't use service worker.
It just gives this : `Uncaught (in promise) TypeError: Failed to fetch` instead of telling the exact error.
This fails to give error messages that are related to accessing 'insecure content' , 'mixed content' , 'host <name> not found' messages like how the browser would normally give.
|
tool,platform-web,a: error message,P2,team-web,triaged-web
|
low
|
Critical
|
591,910,193 |
flutter
|
ReorderableTableView
|
## Use case
I'm trying to create a set of columns, like a table, where i can drag and reorder items both within columns and from one column to another. I'm using `Draggable` and `DragTarget` but i find it difficult to keep the data in sync with the views and draggables. Maybe someone else have this problem and a generic solution is in place.
## Proposal
I propose a new `ReorderableTableView` widget, like `ReorderableListView` only you can also move items from one column to another ( where items are also reordered ).
You would have a callback like this:
```
void ReorderCallback (
int oldColumn,
int oldRow,
int newColumn,
int newRow
)
```
or better it would default to a behaviour where a new item would "fit" in between other rows, but you could subscribe to a callback whenever a item was moved.
The action to perform before dragging could be optional, that is, instead of long press, we could do a normal press and drag.
It would be nice if you could also select spesific rows ( or cells ) as non-draggable.
|
c: new feature,framework,f: scrolling,would be a good package,c: proposal,P3,team-framework,triaged-framework
|
low
|
Minor
|
591,912,965 |
PowerToys
|
Markdown Preview: Show images
|
Iit would be nice if the images referred to in Markdown files could be shown. This may be a security issue though.
> Note: I am the author of [MarkdownPreview](https://github.com/Atrejoe/MarkdownPreview), a Windows Explorer preview pane extension looking forward to drop support, already announcing people to switch to Powertoys
|
Idea-Enhancement,Product-File Explorer
|
medium
|
Critical
|
591,943,011 |
PowerToys
|
Syntax highlighting for Markdown preview
|
Markdown preview does not do any syntax highlighting. Please enable syntax highlighting.
This will prevent me from opening vscode for every README that I come across.
|
Idea-Enhancement,Product-File Explorer
|
low
|
Major
|
591,952,138 |
TypeScript
|
Support non-constructor JSX factory functions
|
**TypeScript Version:** 3.3.3 to 3.8.3
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:**
jsx factory function constructor component ElementClass
**Expected behavior:**
I'm trying to pass a factory function instead of a class component as a JSX element constructor function. It should be possible according to the Handbook, as described [here](https://www.typescriptlang.org/docs/handbook/jsx.html). See the example under the `Type checking -> class component` section (I can't deep-link with hash/anchor because the server isn't properly configured :( ). I should get no errors with the code provided below.
**Actual behavior:**
I get the following error:
```
JSX element type '{ render: () => { custom: string; }; }' is not a constructor function for JSX elements.
Property 'custom' is missing in type '{ render: () => { custom: string; }; }' but required in type 'Element'.(2605)
```
From what I can tell, this used to work, but the behavior broke somewhere between v3.1.6 and v3.3.3 (as tried in the Playground). I'm guessing there was some change that altered detection between React function components and class components and that change introduced this bug. I'm not even sure if React runtime even allows factory functions, but my use case is a custom library and I'm not using React at all. This error is the same regardless of using React or not.
Also, with the exact example from the Handbook, because the JSX.Element is not specified explicitly, there is no error, even when there should be one. See [Playground Link for Handbook example](https://www.typescriptlang.org/play/index.html?jsx=1&ts=3.3.3#code/CYUwxgNghgTiAEA7KBbEBnADlMCBSAygBrwDeAUPPAJaIAuIMAZjggKIQhr0DC066MpSrw4iUDABc8KIgCeAbmEBfcqvKQoA+AFk5PAPYpMBxCHpCqYiQAoAlGVWqmAV0Rg61U7rkAxHHQGMH5uHl6I9paiIHQuMIhk0eKM0pEAvAB8jvBO5OQAPHqGxqbmdPAA9BkKlRXwBgDWBXr+HkEh7p7eVTUVdY15mtoAcgZ0AIIAalAQ1MDFJmYWpM6hXQmjE9OzwK2Bwb5r4ZEUVjFxCStK6vmbUzNzC6UWPbXwjDBBBXfbc3vth064Uq1TeHy+QA)
<!-- Did you find other bugs that looked similar? -->
**Related Issues:**
https://github.com/microsoft/TypeScript/issues/5740
https://github.com/microsoft/TypeScript/issues/5478
**Code**
```ts
declare namespace JSX {
interface Element {
custom: string;
}
interface ElementClass {
render: () => Element;
}
}
class MyComponent {
render() {
return {
custom: "2"
};
}
}
function MyFactoryFunction() {
return { render: () => { return { custom: "2" }; } }
}
function MyFunctionComponent() {
return { custom: "2" }
}
<MyComponent />; // ok
<MyFactoryFunction />; // ok
<MyFunctionComponent />; // ok
class NotAValidComponent {}
function NotAValidFactoryFunction() {
return {};
}
<NotAValidComponent />; // error
<NotAValidFactoryFunction />; // error
```
<details><summary><b>Output</b></summary>
```ts
"use strict";
class MyComponent {
render() {
return {
custom: "2"
};
}
}
function MyFactoryFunction() {
return { render: () => { return { custom: "2" }; } };
}
function MyFunctionComponent() {
return { custom: "2" };
}
<MyComponent />; // ok
<MyFactoryFunction />; // ok
<MyFunctionComponent />; // ok
class NotAValidComponent {
}
function NotAValidFactoryFunction() {
return {};
}
<NotAValidComponent />; // error
<NotAValidFactoryFunction />; // error
```
</details>
<details><summary><b>Compiler Options</b></summary>
```json
{
"compilerOptions": {
"noImplicitAny": true,
"strictNullChecks": true,
"strictFunctionTypes": true,
"strictPropertyInitialization": true,
"strictBindCallApply": true,
"noImplicitThis": true,
"noImplicitReturns": true,
"useDefineForClassFields": false,
"alwaysStrict": true,
"allowUnreachableCode": false,
"allowUnusedLabels": false,
"downlevelIteration": false,
"noEmitHelpers": false,
"noLib": false,
"noStrictGenericChecks": false,
"noUnusedLocals": false,
"noUnusedParameters": false,
"esModuleInterop": true,
"preserveConstEnums": false,
"removeComments": false,
"skipLibCheck": false,
"checkJs": false,
"allowJs": false,
"declaration": true,
"experimentalDecorators": false,
"emitDecoratorMetadata": false,
"target": "ES2017",
"module": "ESNext"
}
}
```
</details>
**Playground Link:** [Provided](https://www.typescriptlang.org/play/index.html?jsx=1#code/CYUwxgNghgTiAEA7KBbEBnADlMCBSAygBrwDeAUPPAJaIAuIMAZjggKIQhr1mVXxgArujoB7FAC54ImLQDmAbj4BfPrQbNW8DlxD0AwtHTpe-eHESgYUgBQBKeAF4AfNs7c6SqqtXlIUY3gAWQBPfXFMUUQ9OlNzPSt7OKo4OkEYRGT+IRFxKQAiACZ8vm8veB9yJkFEMDpqKOCQgDEcMRgWmrqGxCSKFJA0jLJ4y0ZbBxcR1PTM0gFhMUl4IvyKhQqK8lVq2vrG0Oau-cRwlEjo+j6+GeH5nKWC4q3fAB5Qs4uY+AB6Zw2fj94KIANbkd4tNqiDpHPY9X7-X5A0Hgw7HHqfKLfP4A5FgvxGEwAOVEdAAggA1KAQajATGXWKkHboxok8lUmnAVp1aGdOFRa4DIZzZRKN5synU2n07GIwHwRgwaHgiUc2nc9p87qNHFIhUwJUwIA)
|
Suggestion,Awaiting More Feedback
|
low
|
Critical
|
592,008,620 |
rust
|
Cargotest enhancements
|
The Infra team recently made some decisions around the future of cargotest - and there is more to decided and implemented. This issue tracks this work.
Reference: https://paper.dropbox.com/doc/The-future-of-cargotest--AxQ6EFIbNcslFkGPv4Kbr8yhAg-KuqIf3UWJnfrfrukx8o4G
|
A-testsuite,P-low,T-bootstrap,T-infra,C-tracking-issue
|
low
|
Minor
|
592,009,951 |
TypeScript
|
Making isServer() and isClient() checks work in TypeScript
|
## Search Terms
React, Isomorphic, Client/Server, Frontend, Backend, process, server, window
<!-- List of keywords you searched for before creating this issue. Write them down here so that others can find this suggestion more easily -->
## Suggestion
Add a typing mechanism for narrowing down global variable declarations.
```ts
function isServer(): globalThis is typeof globalThis & NodeJS.Global;
```
```ts
function isClient(): globalThis is typeof globalThis & Window;
```
## Use Cases
Isomorphic JavaScript code that runs in browser and server often checks for environment to decide to do various things. Today we have to add both `dom.d.ts` and `@types/node` to bring in all global variables from both declarations such as `window` and `process` in.
In runtime however it's not safe to assume `window` exists so we have to check for it:
## Examples
```ts
if (isClient()) {
window.location.search = '?foo'
}
```
```ts
if (isServer()) {
console.log(process.env.FOO);
}
```
## 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.)
* [x] This feature would agree with the rest of [TypeScript's Design Goals](https://github.com/Microsoft/TypeScript/wiki/TypeScript-Design-Goals).
|
Suggestion,Awaiting More Feedback
|
low
|
Minor
|
592,029,837 |
create-react-app
|
Support incremental transition from flow to typescript
|
### Is your proposal related to a problem?
If you have a codebase written in Flow but want to transition to TypeScript, it would be useful for Create React App to support doing this incrementally without additional config.
At the moment, when following the ['Adding Flow'](https://create-react-app.dev/docs/adding-flow) instructions followed by the ['Adding TypeScript' instructions](https://create-react-app.dev/docs/adding-typescript) instructions, you get an error such as:
```
'import type' declarations can only be used in TypeScript files. TS8006
1 | // @flow
2 | import React from 'react';
> 3 | import type { Node } from 'react';
```
Example steps to recreate in the first 3 commits here:
https://github.com/penx/cra-flow-ts/commits/master
### Describe the solution you'd like
If you follow the ['Adding TypeScript' instructions](https://create-react-app.dev/docs/adding-typescript) and add TypeScript to a project that already uses [Flow](https://create-react-app.dev/docs/adding-flow), it should compile both Flow and TypeScript files correctly.
### Describe alternatives you've considered
This is possible by [manually editing .tsconfig](https://www.typescriptlang.org/docs/handbook/tsconfig-json.html) after it is generated, as per @tebuevd's comment below. However, Create React App should detect the presence of flow (either flow-bin being in dependencies or presence of a .flowconfig file) and setup .tsconfig to have `"include": ["src/**/*.ts", "src/**/*.tsx"]` and `"allowJs": false` when it is generated.
|
issue: proposal,needs triage
|
low
|
Critical
|
592,050,722 |
flutter
|
[gen_l10n] Checking for unused messages
|
Currently, there is no way to tell if there is a message that isn't being used in the Flutter application. It would be nice to be able to use the tool to check what messages aren't being used in the application so that they can be removed, edited, properly used, etc.
/cc @HansMuller
|
c: new feature,framework,a: internationalization,c: proposal,P3,team-framework,triaged-framework
|
low
|
Minor
|
592,067,506 |
godot
|
Godot project manager not responding
|
**Godot version:**
Godot_v3.2.1-stable_win64 but any other vesion of godot seems to be not working too (I tried with 3.2 and 3.1, win32 and win 64).
**OS/device including version:**
Windows 10 Version 10.0.18363
CPU Intel(R) Core(TM) i5-7200U CPU @ 2.50GHz, 2712 MHz, 2 core(s), 4 logical processors(s)
GPU chipset Intel(R) HD Graphics 620 (driver version 26.20.100.7325)
**Issue description:**
1. If if run the executable, the project manager will open and freeze. The window won't close when I click the [x] button.
2. I can open a project.godot : the windows explorer will ask me for the path to Godot, then freeze about 5 minutes before opening the project. If I try to open the project manager from the engine editor, the project manager window will open and freeze untill I kill the process.
I used to think that the issue came from my GPU, but Godot seemed to work perfectly fine a few days ago and I didn't change my hardware. Also, my drivers seem to be up to date.
**Steps to reproduce:**
1. Run the executable.
2. Open an old project. Once the engine editor has opened, go to project and select open project manager. (I apologize if the names aren't correct, I'm using the french version and cannot check the english names, due to the issue).
**Minimal reproduction project:**
No minimal reproduction project might be required for this issue, because it seems to happen only on my laptop when I launch the executable.
I'd like to thank you and apologize if I'm not doing this correctly. This is my first opened issue and I am willing to make things better for the engine.
|
topic:editor,confirmed
|
medium
|
Major
|
592,097,387 |
flutter
|
[gen_l10n] Automatically handle unimplemented messages
|
We want developers to be able to say AppLocalizations.foo or AppLocalizations.bar("arg1", "arg2"), before foo or bar have been defined in the template arb file. This will allow Flutter developers using the tool to continue writing their applications without requiring that the methods and/or getters are available to them by defining their ARB file.
For a sample workflow:
1. Use `Text(AppLocalizations.someMessage)` in application. `newMessage` does not exist in any ARB file yet. However, building the app is possible, it simply generates some placeholder text, such as "Message not implemented."
2. Define the message in the template ARB file. Something like:
```
{
"someMessage": "Some text to display.",
"@someMessage:" {...}
}
```
3. Run `gen_l10n` tool. Notice that the app now builds the `Text` widget with "Some text to display." in the application.
|
c: new feature,framework,a: internationalization,c: proposal,P3,team-framework,triaged-framework
|
low
|
Minor
|
592,100,047 |
pytorch
|
[JIT] Tracer bakes List[Tensor] attribute as constant
|
```
class TryTracingListClass(torch.nn.Module):
def __init__(self):
super().__init__()
self.classes = [
torch.rand(3, 4) for _ in range(3)]
def forward(self):
# return torch.ops._TorchScriptTesting.take_a_list2(self.classes)
return torch.cat(self.classes)
traced = torch.jit.trace(TryTracingListClass(), ())
print(traced.graph)
```
```
graph(%self : __torch__.TryTracingListClass):
%2 : Double(3, 4) = prim::Constant[value=<Tensor>]() # test/test_jit.py:5937:0
%3 : Double(3, 4) = prim::Constant[value=<Tensor>]() # test/test_jit.py:5937:0
%4 : Double(3, 4) = prim::Constant[value=<Tensor>]() # test/test_jit.py:5937:0
%5 : Tensor[] = prim::ListConstruct(%2, %3, %4)
%6 : int = prim::Constant[value=0]() # test/test_jit.py:5937:0
%7 : Double(9, 4) = aten::cat(%5, %6) # test/test_jit.py:5937:0
return (%7)
```
cc @suo
|
oncall: jit,triaged
|
low
|
Minor
|
592,133,187 |
vscode
|
Display Language doesn't update automatically
|
- Open argv file and change the display language.
- No restart notification is shown
Re #93771
* Change Display Language in one instance, it requires restarting the app
* The second instance doesn't show any notification after sync update
* Restart the second instance uses the new display language
|
feature-request
|
low
|
Minor
|
592,151,861 |
rust
|
rustdoc: Indicate that an item is in the prelude
|
**Describe the problem you are trying to solve**
It sucks when you have been importing an item manually your whole life and then you find out that it's in the prelude.
**Describe the solution you'd like**
It would be very nice if there was like a little check next to the item's name that says "_✓prelude_", when it's in the crate's prelude.
|
T-rustdoc,A-resolve,C-feature-request
|
low
|
Minor
|
592,152,082 |
create-react-app
|
Support worker-loader and report errors in child loaders
|
### Is your proposal related to a problem?
If you try to use `worker-loader` in your webpack config, either by ejecting or by patching it into `react-scripts/config/webpack.conf.js` by other means, and there are any eslint errors in your worker files, your build will pretend to succeed but it will produce no build files other than a copy of the `public` directory.
This is because `react-scripts` doesn't report errors in child loaders. [A PR was created to address this a year ago but it still hasn't been merged.](https://github.com/facebook/create-react-app/pull/6286). [This was originally reported as a problem with using `workerize-loader` in your project](https://github.com/facebook/create-react-app/issues/8014). That issue seems to have been closed prematurely because the related PR never got merged. I haven't used `workerize-loader`, but I've heard there are some issues with using it in your webpack config as well.
### Describe the solution you'd like
Merge the PR. Handle errors in child loaders.
Also, it'd be nice if I could use `worker-loader` without ejecting or patching `react-scripts`. I mention that in the same issue because it's a concrete repro case for the issue I'm talking about here. I don't know what a solution to this would look like, but if there could be some blessed way to add plugins without ejecting that'd be great. Or `create-react-app` could ship with configuration for common ones, and enable that configuration if the relevant packages happened to be installed.
### Describe alternatives you've considered
I'm already patching this change into my own project because it's really bad for a build system to think it succeeded but not produce a build. I used `patch-package` to edit the build script without ejecting. I also used `patch-package` to add `worker-loader` to my webpack config.
|
issue: proposal,needs triage
|
low
|
Critical
|
592,162,843 |
pytorch
|
Linking error after marking an op `manual_kernel_registration: True`
|
## 🐛 Bug
If you specify `manual_kernel_registration: True` for an operator in `native_functions.yaml`, libtorch won't build unless you expose implementation of your operator for exactly one backend, arbitrarily chosen.
## To Reproduce
1. checkout https://github.com/pytorch/pytorch/pull/35829
1. `python setup.py develop`
1. The following linking error happens:
```
/home/wbaranowski/miniconda3/envs/pytorch-cuda-dev/bin/../lib/gcc/x86_64-conda_cos6-linux-gnu/7.3.0/../../../../x86_64-conda_cos6-linux-gnu/bin/ld: /home/wbaranowski/git/Quansight/vanilla/build/lib/libtorch_cpu.so: undefined reference to `at::native::resize_(at::Tensor&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat>)'
collect2: error: ld returned 1 exit status
ninja: build stopped: subcommand failed.
```
## Expected behavior
Require/allow either all implementations to be exported in `at::native::`, or none.
## Environment
```
$ python collect_env.py
Collecting environment information...
PyTorch version: N/A
Is debug build: N/A
CUDA used to build PyTorch: N/A
OS: Ubuntu 18.04.3 LTS
GCC version: (crosstool-NG 1.23.0.450-d54ae) 7.3.0
CMake version: version 3.16.2
Python version: 3.6
Is CUDA available: N/A
CUDA runtime version: 10.1.243
GPU models and configuration:
GPU 0: TITAN RTX
GPU 1: TITAN RTX
Nvidia driver version: 440.33.01
cuDNN version: /usr/local/cuda-10.2.89/targets/x86_64-linux/lib/libcudnn.so.7
Versions of relevant libraries:
[pip] numpy==1.17.5
[pip] torch==1.5.0a0+3927147
[pip] torchvision==0.6.0a0+30538b5
[pip] torchviz==0.0.1
[conda] cudatoolkit 10.0.130 0 anaconda
[conda] magma-cuda101 2.5.1 1 pytorch
[conda] mkl 2019.4 243
[conda] mkl-include 2019.4 243
[conda] numpy 1.17.5 py36h95a1406_0 conda-forge
[conda] torch 1.5.0a0+c219d53 pypi_0 pypi
[conda] torchvision 0.6.0a0+30538b5 pypi_0 pypi
[conda] torchviz 0.0.1 pypi_0 pypi
```
## Additional context
Once you add `manual_kernel_registration` in `native_functions.yaml`, you can no longer use `dispatch` with different function names for different backends. But the C++ prototype is still generated in `torch/include/ATen/NativeFunctions.h` with `CAFFE2_API` and called from `at::TypeDefault` so you have to expose exactly one implementation, regardless of how many implementations there actually are.
This issue came up while I was working on https://github.com/pytorch/pytorch/issues/33941 and tried to fix it with https://github.com/pytorch/pytorch/pull/35679. I have commented about it on https://github.com/pytorch/pytorch/pull/30649#pullrequestreview-384469484 and then discussed with @ezyang. cc @smessmer
|
module: build,triaged
|
low
|
Critical
|
592,178,223 |
PowerToys
|
Synchronize taskbar icon order
|
There should be a way to enable/disable as part of my profile synchronization of the taskbar settings. I have this issue because I use multiple machines, and my winkey+number shortcuts don't always work properly; because the taskbar shortcuts aren't in the same order on both my machines.
If it could only sync the first 10 that would also work, but if it can sync entire taskbar configuration (placement, size, etc) that would be great. I realize that this would mean that the applications installed have to be synchronized as well and there is no real way to do that - but perhaps an option to enable / disable a specific icon; for example "pin to task bar" could mean "pin and sync" or "pin only to this machine" etc.
Not really sure how it could be done or even if its possible.
Stay safe!
|
Idea-New PowerToy
|
low
|
Minor
|
592,192,600 |
flutter
|
Add option to avoid pod install during "flutter build ios"
|
<!-- 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 am trying to setup a flutter iOS build on AppCenter, the project depends on a few company internal pods. I normally commit all installed pods together with the project so never need to do a Pod Install on CI portal.
However, "flutter build ios" will force a Pod Install, have been looking for couple days for ways to disable "Pod Install" but didn't find anything.
Have been trying with Flutter 1.12.13+hotfix.8 • channel stable • and Flutter 1.15.17 • channel beta •
Thanks
<!--
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
Have an option for "flutter build ios" to disable Pod Install, or allow customize Pod command (such as Pod update instead of Pod Install)
Any alternatives are welcome as well.
<!--
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.
-->
|
platform-ios,tool,a: build,P3,team-ios,triaged-ios
|
medium
|
Critical
|
592,192,990 |
flutter
|
Clang 6e00e3fcb082c3f19309ebc6f184ab326eec3328 fails on Android x86 builds
|
Rolling to https://chrome-infra-packages.appspot.com/p/fuchsia/third_party/clang/mac-amd64/+/git_revision:6e00e3fcb082c3f19309ebc6f184ab326eec3328 (the latest GOMA enabled toolchain as of this writing) fails on lld for Android x86 (and maybe x64) targets, with errors like:
> contains a compressed section, but zlib is not available
For some gcc libraries in the NDK. I found https://github.com/android/ndk/issues/888, which is Windows specific and has similar errors, but I believe the problem here is with the Fuchsia clang toolchain we're using rather than with the Android NDK. Rolling to a slightly earlier revision of clang works (e.g. 7e9747b50bcb1be28d4a3236571e8050835497a6).
/cc @chinmaygarde
|
platform-android,engine,P2,team-android,triaged-android
|
low
|
Critical
|
592,196,061 |
go
|
cmd/gofmt: allow same level indent for block comments
|
Given the following code:
```go
/*
someDisabledCode()
*/
```
It will be reformatted as:
```go
/*
someDisabledCode()
*/
```
(notice the indent it injects)
This is frustrating for users who need to maintain patches since it means that there is now a potentially massive diff when trying to comment large blocks of code in a patch.
Would it be possible for gofmt to stop trying to indent everything within the block comments?
\cc @griesemer
|
NeedsInvestigation
|
low
|
Major
|
592,202,619 |
flutter
|
Support golden tests in C++ unit tests for the engine
|
Right now, engine unit tests do not have a good way to run golden file tests on multiple platforms, particularly when there are features like text/gradients/shadows/etc. that differ slightly across platforms.
It seems like the ideal solution would be to make the gtest fixtures use Skia gold if possible. /cc @piinks @kjlubick for input on that.
/cc @chinmaygarde
|
a: tests,engine,c: proposal,P2,team-engine,triaged-engine
|
low
|
Minor
|
592,262,073 |
godot
|
.fnt reload doesn't clear removed kerning pairs
|
**Godot version:** 3.2.0
**OS/device including version:** Windows 10, Alienware Aurora 15(?)
**Issue description:** When editing a .fnt file, Godot will normally detect the change and reload the font data. If you remove a kerning pair, the old values for that pair will remain.
**Steps to reproduce:**
1. Generate a bitmap font with BMFont or similar. Export to the text .fnt format.
2. Create/open a UI Scene
3. Add a text box using that font and type in some text that would typically trigger kerning, e.g. "Aviation"
4. Edit the .fnt file and remove one of the pairs present in the text (e.g. for "Av", the line containing `first=65 second=118`) Save.
5. Focus the Godot window. Observe that the kerning remains unchanged.
6. Restart Godot and open the scene with the text box. The kerning should be different.
**Minimal reproduction project:** Will create upon request.
|
bug,topic:import
|
low
|
Minor
|
592,289,240 |
flutter
|
SelectableText inside InkWell absorbs Tap and nothing happens
|
Placing a `SelectableText` inside `InkWell` makes the `onTap` callback stop working. The ink splash is very minimal and the `SelectableText` selects all the text if I long press for a while.
The expected behavior would be that a tap would execute the `onTap` callback, the ink splash would be wide and visible as normal and a long press should be forwared to the `SelectableText` if the `InkWell` doesn't define a `onLongPress` callback, which would in turn select all the text at once.
|
a: text input,framework,f: material design,f: gestures,has reproducible steps,P2,found in release: 3.3,found in release: 3.7,team-design,triaged-design
|
low
|
Major
|
592,298,089 |
flutter
|
3 point arc in custom painter
|
Not a bug report, my flutter works just fine.
## Use case Trying to create the below shape with custom painter
https://rive.app/a/Pshah/files/flare/suggestion/preview
//https://dribbble.com/shots/10853019-Youtube-Music-Player: Notice the swipe gestures, trying to right a package for that
I was trying to form this shape with custom painter, here's my code: https://github.com/ppshah2023/3-Point-Arc, but I found out that I can't do that. When you create one arc, that arc is upright. Trying to create another arc that's backwards to that arc is backwards, and flutter doesn't allow that.
## Proposal 3 Point arc
I want to create a backward arc, but I thought that creating a technical 3 point arc would be better. Where the current point is stored, and you provide 2 points to define the semicircle of that arc.
I found that such an option is offered in many design vector programs, and this would be a nice touch to add to flutter.
It could be package where there's a thousand more additions for neat vector graphic stuff, but this sort of thing should just be built into flutter. So that it's maintained and not a second choice.
|
framework,c: proposal,P3,team-framework,triaged-framework
|
low
|
Critical
|
592,312,592 |
material-ui
|
Theme interoperability with styled-components and emotion
|
<!-- Provide a general summary of the issue in the Title above -->
Here is the type:
```
export interface CSSProperties extends BaseCSSProperties {
// Allow pseudo selectors and media queries
// `unknown` is used since TS does not allow assigning an interface without
// an index signature to one with an index signature. This is to allow type safe
// module augmentation.
// Technically we want any key not typed in `BaseCSSProperties` to be of type
// `CSSProperties` but this doesn't work. The index signature needs to cover
// BaseCSSProperties as well. Usually you would use `BaseCSSProperties[keyof BaseCSSProperties]`
// but this would not allow assigning React.CSSProperties to CSSProperties
[k: string]: unknown | CSSProperties;
}
```
<!--
Thank you very much for contributing to Material-UI by creating an issue! ❤️
To avoid duplicate issues we ask you to check off the following list.
-->
<!-- Checked checkbox should look like this: [x] -->
- [x] The issue is present in the latest release.
- [x] I have searched the [issues](https://github.com/mui-org/material-ui/issues) of this repository and believe that this is not a duplicate.
## Current Behavior 😯
When I try to spread something like `theme.typography.h5` into an object that Emotion.js will consume, I get a type error.
## Expected Behavior 🤔
Yes, I can pick off individual values fine, but I'd rather be able to use whole objects and spread them in.
## Steps to Reproduce 🕹
Try to spread `useTheme().typography.h5` into the css prop from Emotion.js.
https://codesandbox.io/s/funny-kepler-6u22j
## Context 🔦
<!--
What are you trying to accomplish? How has this issue affected you?
Providing context helps us come up with a solution that is most useful in the real world.
-->
I use Material-UI for components and theming. I use Emotion for adding additional styles to Material-UI components and other components.
Sometimes, I want to use an entire `useTheme().typography[key]` set of styles to override something. For instance, I want a Typography component to have an `h2` variant (have h2 in the dom, for semantic reasons), but I want it to have `h5` styles for visual reasons.
## Your Environment 🌎
| Tech | Version |
| ----------- | ------- |
| Material-UI | v4.9.8 |
| React | 16.13.1 |
| TypeScript | 3.8.3 |
| Emotion.js | 10.0.28 |
|
breaking change,typescript
|
low
|
Critical
|
592,327,044 |
flutter
|
Retain canvas state for high performance drawing
|
Hi,
I would like to be able to continuously draw to a canvas and be able to reflect the canvas state on the screen.
Currently with CustomPainter the paint callback receives a canvas that cannot be set, therefore many developers in the community are relying on passing a points array as state to the custom painter, and then reiterate through the ever growing array redrawing all the points over and over again.
One possible solution: perform the drawing operation into a canvas (member of a widget for ex) and then be able to draw that canvas onto the other canvas in the paint callback of a custom painter by passing it to the custom painter. (HTML5 Canvas allows for drawing one canvas onto another as a reference / Swift allows for drawing CGContext onto another as another reference)
I recently created the following issue: #53804 - which gives a good background to this request.
Thank you for creating this awesome framework!
M
|
engine,c: performance,c: proposal,perf: speed,P2,team-engine,triaged-engine
|
low
|
Major
|
592,352,768 |
opencv
|
null-pointer-dereference and buffer overflow: OpenCV 3.4
|
##### System information (version)
- OpenCV => 3.4.9
- Operating System / Platform => Win64
##### Detailed description
There appear to be several places where a nullptr dereference can occur in core(persistence). These have been listed as security vulnerabilities in https://nvd.nist.gov/vuln/detail/CVE-2019-14493
It appears they were fixed in OpenCV4 but not in OpenCV 3.4 (see issue https://github.com/opencv/opencv/issues/15127 and pull request https://github.com/opencv/opencv/pull/15145).
Similarly, another security bug is listed here https://nvd.nist.gov/vuln/detail/CVE-2019-5063 where a buffer overflow can occur. It appears that https://github.com/opencv/opencv/pull/15868 fixes all locations where buffer overflow can happen in master, but in the 3.4 branch and in 3.4 releases, the fixes were not added (ie https://github.com/opencv/opencv/blob/3.4/modules/core/src/persistence_xml.cpp#L409 in 3.4 vs https://github.com/opencv/opencv/blob/master/modules/core/src/persistence_xml.cpp#L631 in master) . It should be noted that this error is potentially not reproducible in OpenCV 3.4 (see issue https://github.com/opencv/opencv/issues/15857)
As these are both bugs, can we have these fixes back-ported to Open CV 3.4 (if applicable)?
|
priority: low,category: core,affected: 3.4
|
low
|
Critical
|
592,401,308 |
youtube-dl
|
Support https://www.lumni.fr
|
- [x ] I'm reporting a new site support request
- [x ] I've verified that I'm running youtube-dl version **2020.03.24**
- [x ] I've checked that all provided URLs are alive and playable in a browser
- [ x] I've checked that none of provided URLs violate any copyrights
- [ x] I've searched the bugtracker for similar site support requests including closed ones
## Example URLs
- Single video: http://www.lumni.fr/video/la-guerre-froide
## Description
Videos by the France 4, which is a French national channel. There are lived during the day on tv and are provided to help pupils to learn at home because of the covid-19.
|
site-support-request
|
low
|
Critical
|
592,416,784 |
pytorch
|
CUDNN_STATUS_INTERNAL_ERROR with GPU RTX 8000
|
Bug:🐛
cuDNN error: CUDNN_STATUS_INTERNAL_ERROR
## To Reproduce
Steps to reproduce the behavior:
Python crnn.py
<!-- If you have a code sample, error messages, stack traces, please provide it here as well -->
## Expected behavior
It works fine on GPU RTX6000,V100
So it should works fine on RTX8000 too.
## Environment
-Python 3.6
-Pytorch 1.3.1
-Torchvision 0.4
- OS Linux (Ubuntu)
- How you installed PyTorch : pip3
cc @ngimel @csarofeen @ptrblck
|
needs reproduction,module: cudnn,module: cuda,triaged
|
medium
|
Critical
|
592,422,129 |
flutter
|
[espresso] : Unable to run espresso test locally
|
I followed all the steps mentioned in `readme` to run a simple espresso test to check text of a button and then click on it, but getting below error :
```
> Task :app:compileDebugAndroidTestJavaWithJavac FAILED
/Users/deeptibelsare/Documents/DarshanUdacity/Flutter/flutter_app_demo/android/app/src/androidTest/java/com/example/MainActivityTest.java:27: error: cannot find symbol
ActivityScenario.launch(MainActivity.class);
^
symbol: class MainActivity
location: class MainActivityTest
1 error
FAILURE: Build failed with an exception.
* What went wrong:
Execution failed for task ':app:compileDebugAndroidTestJavaWithJavac'.
> Compilation failed; see the compiler error output for details.
```
I've this screen for validation:
And below is a simple test:
```
@Before
public void setUp() throws Exception {
ActivityScenario.launch(MainActivity.class);
}
@Test
public void tapButton() {
onFlutterWidget(withValueKey("buttonText")).check(matches(withText("Send")));
onFlutterWidget(withValueKey("buttonText")).perform(click());
}
```
Appreciate any help in resolving this issue.
Also, what are the other actions, assertions and matchers that I can leverage to write a comprehensive UI test ? Currently I see `click`, `withText`, `withTooltip` and `withValueKey`
|
package,team-ecosystem,P2,p: espresso,triaged-ecosystem
|
low
|
Critical
|
592,439,572 |
TypeScript
|
Provide a ReadonlyRegExp type
|
## Search Terms
readonlyregexp readonly regexp
## Suggestion
Provide a new `ReadonlyRegExp` type that represents a read only `RegExp`.
## Use Cases
This is similar to other readonly types, e.g. `ReadonlyArray` -- I prefer my code to be immutable by default. I have a `DeepReadonly` helper that could map `RegExp` to `ReadonlyRegExp`.
Note that `Readonly<RegExp>` does not work:
```typescript
'a'.replace(/a/ as Readonly<RegExp>, 'b')
/*
No overload matches this call.
The last overload gave the following error.
Argument of type 'Readonly<RegExp>' is not assignable to parameter of type 'string | RegExp'.
Type 'Readonly<RegExp>' is missing the following properties from type 'RegExp': [Symbol.match], [Symbol.replace], [Symbol.search], [Symbol.split](2769)
*/
```
## Examples
```typescript
const emailRegExp = /[a-z]@[a-z\.]/ as ReadonlyRegExp;
export function suppressPII(message: string) {
return message.replace(emailRegExp, '');
}
```
## 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.)
* [x] This feature would agree with the rest of [TypeScript's Design Goals](https://github.com/Microsoft/TypeScript/wiki/TypeScript-Design-Goals).
|
Suggestion,Awaiting More Feedback
|
low
|
Critical
|
592,470,744 |
pytorch
|
torch.autograd.set_detect_anomaly(True) does not exist in C++?
|
## 🐛 Bug
I got the following error when calling `optimizer.backward()`.
`one of the variables needed for gradient computation has been modified by an inplace operation: [CPUFloatType [50, 40]], which is output 0 of TBackward, is at version 2; expected version 1 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True). after train`
## To Reproduce
Not sure, since I can't see which operation that makes my code crash.
<!-- If you have a code sample, error messages, stack traces, please provide it here as well -->
## Expected behavior
I expected the function that was listed in the error to exist.
## Environment
- PyTorch Version (e.g., 1.0): Libtorch 1.4
- OS (e.g., Linux): Windows 7 64 bit
- How you installed PyTorch (`conda`, `pip`, source): source
- Build command you used (if compiling from source): none
- Python version: 3.6
- CUDA/cuDNN version: none
- GPU models and configuration: none
- Any other relevant information: none
cc @yf225
|
module: cpp,triaged
|
low
|
Critical
|
592,508,188 |
vscode
|
[semantic tokens] provide code assist for tokenModifiers
|
refs #93804:
no code assist for tokenModifiers:
<img width="672" alt="2020-04-02_11-59-20" src="https://user-images.githubusercontent.com/1898161/78236003-b6b10080-74d9-11ea-8103-cec53834ae19.png">
I did get code assist in other cases:
<img width="803" alt="2020-04-02_11-59-50" src="https://user-images.githubusercontent.com/1898161/78236058-cc262a80-74d9-11ea-9040-0b876db62f69.png">
|
feature-request,themes
|
low
|
Minor
|
592,556,408 |
godot
|
Weird Light2D mask behaviour when using `light()` in shader
|
**Godot version:**
3.2.1
**OS/device including version:**
Linux 5.5.13-1-MANJARO
**Issue description:**
When using a Light2D as a mask and modifying its appearance in a shader using `light()` the colors are off.
**Steps to reproduce:**
1.Create a Light2D as a mask
2.In the masked sprite add a shader which resamples the original texture and pass it to the `LIGHT.rgb`.
(see reproduction project)
**Minimal reproduction project:**
https://framadrive.org/s/a7PrNRKJ47XGtyr
|
discussion,topic:rendering,documentation
|
low
|
Minor
|
592,560,721 |
opencv
|
OpenCV4.2.0 VideoCapture bug
|
##### System information (version)
- OpenCV => 4.2.0
- Operating System / Platform => Windows 10 64 Bit (One virtual machine)
- Compiler => Visual Studio 2019
when i use 2.4.9, everything works fine. but if i use opencv4.2.0, i got the following image.
```.cpp
#include <iostream>
#include "opencv2/opencv.hpp"
#include <Windows.h>
#include <time.h>
using namespace std;
using namespace cv;
int main(int argc, char* argv[])
{
cv::VideoCapture capture(0);
if (!capture.isOpened())
{
std::cout << "video not open." << std::endl;
return -1;
}
if (!capture.isOpened()) // 如果视频不能正常打开则返回
return -1;
Mat frame;
while (true)
{
time_t now = time(NULL);
capture >> frame; // 等价于cap.read(frame);
if (frame.empty()) // 如果某帧为空则退出循环
{
Sleep(20);
continue;
}
cv::imwrite(to_string(now)+".jpg", frame);
imshow("video", frame);
waitKey(20); // 每帧延时20毫秒
}
capture.release(); // 释放资源
return 0;
}
```
|
priority: low,category: videoio(camera),incomplete,platform: win32,needs investigation
|
low
|
Critical
|
592,617,700 |
electron
|
session option to do not cache authentications
|
### 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
We have a product where we run multiple webpages in parallel, we use `session.setProxy()` to set the proxy. Our proxy provider give the possibility to have different IP on each request. To authenticate on that proxy we have smth `http://customer-x-session-${session}-country-us:[email protected]:port` then use `webContents.on('login')` to make the basic authentication.
The problem here is with the `session` auth cache, so for each request we need to make again the auth with different username which contains that sessions.
So, I found a workaround with `session.clearAuthCache()` but when we have 10 web pages loading in the same session, on the each page navigation we call first `session.clearAuthCache()` but is canceling the pages which are still loading.
### Proposed Solution
Maybe add some config to the session to not cache the basic auth because it's really a problem when it's come to using different proxy username or different endpoints which requires basic auth.
Thank you.
|
enhancement :sparkles:
|
low
|
Minor
|
592,636,753 |
PowerToys
|
Current monitor sinalyzer
|
Hi!
# Summary of the new feature/enhancement
There's something which annoys me on using multiple monitors: sometimes, I start to type on the wrong one and mess with code or trigger an unintended keyboard shortcut.
Is it possible to change the color of the toolbar of the current monitor? I need a way to easily know which one is the focused window.
Thank you!
|
Idea-New PowerToy,Product-Window Manager
|
low
|
Minor
|
592,688,876 |
TypeScript
|
Include CompilerHost in TransformationContext
|
## Search Terms
compiler host, CompilerHost,
transformation context, TransformationContext,
transformer, custom transformer
## Suggestion
I'd like to access the `CompilerHost` instance from my custom transformers (via `TransformationContext`).
As far as I can see this wouldn't expose any "risky" patterns like #25147.
## Use Cases
My specific use-case is resolving import declarations. This can be done using `ts.resolveModuleName` (https://github.com/microsoft/TypeScript/issues/28276#issuecomment-435016356), but this requires a `CompilerHost` instance.
## Example
An example for my specific use-case would be:
```ts
const transformer: ts.TransformerFactory<ts.SourceFile> = context => {
return file => {
return ts.visitEachChild(
file,
node => {
if (ts.isImportDeclaration(node)) {
const resolved = ts.resolveModuleName(
(node.moduleSpecifier as ts.StringLiteral).text,
file.fileName,
context.getCompilerOptions(),
context.getCompilerHost() // this is currently impossible
)
// do something interesting with resolved
}
return node;
},
context
)
}
}
```
## Workaround
Luckily there is a workaround for this:
```ts
const host = ts.createCompilerHost(context.getCompilerOptions());
```
But this feels redundant and somewhat "wrong".
## 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.)
* [x] This feature would agree with the rest of [TypeScript's Design Goals](https://github.com/Microsoft/TypeScript/wiki/TypeScript-Design-Goals).
|
Suggestion,In Discussion
|
low
|
Minor
|
592,693,030 |
pytorch
|
Error when loading jit traced FasterRCNN model in C++
|
## 🐛 Bug
<!-- A clear and concise description of what the bug is. -->
I used the torch.jit.trace() in python to trace the fasterrcnn_resnet50_fpn model provided in the latest torchvision (I installed the torchvision from source). However, an unhandled exception occured when I used libtorch API to load this model.
## To Reproduce
Steps to reproduce the behavior:
1.Use Python API to trace model:
model = torchvision.models.detection.maskrcnn_resnet50_fpn(pretrained=True)
script_model = torch.jit.script(model)
script_model.save("./data/rcnn.pt")
2. Load "rcnn.pt" in Libtorch:
torch::jit::script::Module module = torch::jit::load("./data/rcnn.pt");
3. Then the exception occured:
Unhandled exception at 0x00007FFF5D1CA839 in RCNN_Pytorch_Demo.exe: Microsoft C++ exception: torch::jit::script::ErrorReport at memory location 0x00000037EDCEEF30.
<!-- If you have a code sample, error messages, stack traces, please provide it here as well -->
## Expected behavior
<!-- A clear and concise description of what you expected to happen. -->
It should load the traced model without exception.
## Environment
Visual Studio 2017
PyTorch version: 1.4.0+cu92
Is debug build: No
CUDA used to build PyTorch: 9.2
OS: Microsoft Windows 10 Professional
GCC version: Could not collect
CMake version: version 3.16.0-rc3
Python version: 3.6
Is CUDA available: Yes
CUDA runtime version: 9.2.148
GPU models and configuration:
GPU 0: GeForce GTX 1080 Ti
GPU 1: GeForce GTX 1080 Ti
Nvidia driver version: 432.00
cuDNN version: C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v9.2\bin\cudnn64_7.dll
Versions of relevant libraries:
[pip] efficientnet-pytorch==0.5.1
[pip] numpy==1.17.4
[pip] numpydoc==0.8.0
[pip] torch==1.4.0+cu92
[pip] torchfile==0.1.0
[pip] torchnet==0.0.4
[pip] torchvision==0.5.0+cu92
[conda] blas 1.0 mkl defaults
[conda] efficientnet-pytorch 0.5.1 <pip>
[conda] mkl 2018.0.2 1 defaults
[conda] mkl-service 1.1.2 py36h57e144c_4 defaults
[conda] mkl_fft 1.0.1 py36h452e1ab_0 defaults
[conda] mkl_random 1.0.1 py36h9258bd6_0 defaults
[conda] numpy 1.14.3 py36h9fa60d3_1 defaults
[conda] numpy 1.17.4 <pip>
[conda] numpy-base 1.14.3 py36h555522e_1 defaults
[conda] numpydoc 0.8.0 py36_0 defaults
[conda] torch 1.4.0+cu92 <pip>
[conda] torchfile 0.1.0 <pip>
[conda] torchnet 0.0.4 <pip>
[conda] torchvision 0.6.0a0 <pip>
Libtorch 1.4.0
cc @peterjc123 @maxluk @nbcsm @guyang3532 @gunandrose4u @mszhanyi @skyline75489 @gmagogsfm @suo
|
oncall: jit,triaged
|
medium
|
Critical
|
592,755,683 |
PowerToys
|
Profiles in Keyboard Manager
|
Have both predefined profiles (e.g. mac/linux/etc) available to user, and also allow the user to create custom profiles.
Provide a quick way to switch between profiles.
|
Idea-Enhancement,Product-Keyboard Shortcut Manager
|
high
|
Critical
|
592,768,287 |
vscode
|
Timeline providers that don't need URI's should be able to render on empty workbench
|
#84297
Currently when the workbench has no ActiveTextEditor, the timeline view won't load anything. But there is a category of "workspace timeline" events that should be able to work without any active editor.
For instance:
Timeline of past/future sprints/iterations
Timeline counting down days till code freeze/GA/etc.
Timeline of git tags/releases
I think the provideTimeline should give an optional uri, and if an extension needs the URI they simply return undefined to be excluded from the timeline. But if not, they can elect to be shown even when theres no active editor.
|
feature-request,timeline
|
low
|
Minor
|
592,799,758 |
vscode
|
Better support for case sensitive macOS
|
Issue Type: <b>Bug</b>
I have found a case where a GIT repo has two folders with different casing.
but VSCode only shows one directory
related to https://github.com/microsoft/vscode/issues/90983
VS Code version: Code - Insiders 1.44.0-insider (20b88fa0694cae8e46ffd305a1450ba58ae6065b, 2020-04-01T08:34:11.087Z)
OS version: Darwin x64 19.4.0
<details>
<summary>System Info</summary>
|Item|Value|
|---|---|
|CPUs|Intel(R) Core(TM) i7-8850H CPU @ 2.60GHz (12 x 2600)|
|GPU Status|2d_canvas: enabled<br>flash_3d: enabled<br>flash_stage3d: enabled<br>flash_stage3d_baseline: enabled<br>gpu_compositing: enabled<br>metal: disabled_off<br>multiple_raster_threads: enabled_on<br>oop_rasterization: disabled_off<br>protected_video_decode: unavailable_off<br>rasterization: enabled<br>skia_renderer: disabled_off_ok<br>video_decode: enabled<br>viz_display_compositor: enabled_on<br>viz_hit_test_surface_layer: disabled_off_ok<br>webgl: enabled<br>webgl2: enabled|
|Load (avg)|3, 3, 3|
|Memory (System)|16.00GB (0.80GB free)|
|Process Argv|-psn_0_5612890|
|Screen Reader|no|
|VM|0%|
</details><details><summary>Extensions (22)</summary>
Extension|Author (truncated)|Version
---|---|---
vscode-base64|ada|0.1.0
markdown-toc|Ala|1.5.6
inline-hasher|ale|1.0.0
spellright|ban|3.0.50
better-toml|bun|0.3.2
gotemplate|cas|0.3.0
hadolint|exi|0.3.0
github-actions|for|0.0.1
terraform|mau|1.4.0
markdown-shortcuts|mdi|0.12.0
vscode-docker|ms-|1.0.0
github-actions-vscode|Oma|2.7.0
vscode-yaml-sort|Pas|2.3.0
errorlens|Phi|1.1.3
vscode-yaml|red|0.7.2
vscode-gitweblinks|red|1.5.0
code-settings-sync|Sha|3.4.3
vscode-nginx|sha|0.6.0
shellcheck|tim|0.9.0
sort-lines|Tyr|1.9.0
vscode-icons|vsc|10.0.0
HCL|who|0.0.5
</details>
<!-- generated by issue reporter -->
|
feature-request,file-explorer,macos,file-io
|
high
|
Critical
|
592,807,484 |
TypeScript
|
JSDoc Object.<key, value> Syntax doesn't suport uppercase key type
|
<!-- ⚠️⚠️ Do Not Delete This! bug_report_template ⚠️⚠️ -->
<!-- Please read our Rules of Conduct: https://opensource.microsoft.com/codeofconduct/ -->
<!-- Please search existing issues to avoid creating duplicates. -->
<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: https://code.visualstudio.com/insiders/ -->
<!-- Use Help > Report Issue to prefill these. -->
- VSCode Version: 1.43.1, fe22a9645b44368865c0ba92e2fb881ff1afce94, x64
- OS Version: Ubuntu 18.04.4 LTS
Steps to Reproduce:
1. Use the type `Object.<String, Number>` in JSDoc.
2. Intellisense parses the type as `any`.
3. Change `String` to `string` (which should be equivalent; at least it is on all other occurrences)
4. Intellisense parses the type correctly as `{ [string]: number }`.
<!-- Launch with `code --disable-extensions` to check. -->
Does this issue occur when all extensions are disabled?: Yes
|
Bug,Domain: JSDoc
|
low
|
Critical
|
592,828,281 |
TypeScript
|
Parsing issue for call expression with type arguments following left-shift
|
**TypeScript Version:** 3.8.3
**Search Terms:** left shift, type parameter
**Expected behavior:** Syntax highlighting stays intact.
**Actual behavior:** Syntax highlighting breaks, even with ts-ignore.
**Code**
```ts
function alignof<T>(): number {
return 1;
}
// works
function foo<T>(): number {
return 1 >> alignof<T[]>();
}
// breaks
function bar<T>(): number {
return 1 << alignof<T[]>();
}
// still breaks, does not ignore
function baz<T>(): number {
// @ts-ignore
return 1 << alignof<T[]>();
}
```
Looks like a rescan issue on potentially ambiguous `<` tokens, attempting to interpret as `1 < <type>`.
**Playground Link:** [Provided](https://www.typescriptlang.org/play/index.html#code/GYVwdgxgLglg9mABAQwDYwOZjsAPAFQD4AKASgC5EwQBbAIwFMAnRAbwFgAoRRJhqEEyQBGANxcAvly4B6GYgDucJgGsAzl1CRYCRMDhwCJClVqMWHbr36CRiQoRTosOAgG0AusfGcpnWfJ0fMjqmuDQ8Eh0yExGZJTU9MxsXDx8AkKIwoi4uE6Y2Hj4nt6S0pxyiGqwqKiIQQwhagA0iAAmcAxqVHBQiAXKDGHakfXIAF5xJonmKVaVAAJQagC0A3yp1hl2ufkuRSVkPhJAA)
|
Bug,Help Wanted,PursuitFellowship
|
low
|
Minor
|
592,832,627 |
PowerToys
|
[FancyZones] Priority grid zone positions are not consistent
|
<!--
**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.
-->
# Environment
```
Windows build number: 19041.172
PowerToys version: Master/0.16.0
PowerToy module for which you are reporting the bug (if applicable): FancyZones
```
# Steps to reproduce
With the _'During zone layout changes, windows assigned to a zone will match new size/position'_ setting enabled: Use the Priority Grid template layout with two zones. Add a window to the 'priority' zone (the largest zone). Add a new zone. The window will no longer be in the priority zone and will now be in the non-priority zone to the left. This occurs whenever a new zone is added to the left of the priority zone.
# Expected behavior
Windows added to the priority zone should remain in the priority zone when the number of zones are increased or decreased.
In general, when the zone count is increased or decreased windows should stay in the zone that most closely matches the position they're currently in, as is when increasing/decreasing zones they move around unintuitively.
This is because zone numbering is simply done from top-left to bottom-right without considering what position these zones are in when the zone count is higher or lower.
# Actual behavior
See the below screenshots, I am using the priority grid layout with 2 zones, I add a third zone, Visual Studio moves from the priority zone to the leftside non-priority zone. File explorer, not in the priority zone moves into the priority zone.
# Screenshots
2 zones:
3 zones:
# Proposed fix
The layout info for the priority grid is defined in ZoneSet.cpp, here is a code snippet for 2 and 3 zones:
```C++
/* 2 */
JSONHelpers::GridLayoutInfo(JSONHelpers::GridLayoutInfo::Full{
.rows = 1,
.columns = 2,
.rowsPercents = { 10000 },
.columnsPercents = { 6667, 3333 },
.cellChildMap = { { 0, 1 } } }),
/* 3 */
JSONHelpers::GridLayoutInfo(JSONHelpers::GridLayoutInfo::Full{
.rows = 1,
.columns = 3,
.rowsPercents = { 10000 },
.columnsPercents = { 2500, 5000, 2500 },
.cellChildMap = { { 0, 1, 2 } } }),
```
The `cellChildMap` for each layout should be changed such that zone positions stay consistent across zone counts. In this example the `cellChildMap` for 3 zones should be `.cellChildMap = { { 2, 0, 1 } }`. This would keep the windows in the priority zone in the priority zone, windows on the right hand side of the screen would remain on the right, and a new zone would be added on the left.
As well as this change a new overload method for `ZoneSet::AddZone` would need adding. Currently this method simply sets the zone's Id in a top-left to top-right fashion:
```C++
IFACEMETHODIMP ZoneSet::AddZone(winrt::com_ptr<IZone> zone) noexcept
{
m_zones.emplace_back(zone);
// Important not to set Id 0 since we store it in the HWND using SetProp.
// SetProp(0) doesn't really work.
zone->SetId(m_zones.size());
return S_OK;
}
```
The new method should also take in the `cellChild` value and use that for the Id instead, something like this:
```C++
IFACEMETHODIMP ZoneSet::AddZone(winrt::com_ptr<IZone> zone, integer cellChild) noexcept
{
m_zones.emplace_back(zone);
zone->SetId(cellChild + 1);
return S_OK;
}
```
|
Issue-Bug,FancyZones-Dragging&UI,Product-FancyZones,Priority-3
|
low
|
Critical
|
592,855,090 |
rust
|
An std::fmt syntax to propagate formatting parameters
|
When implementing custom `fmt::Debug`/`Display` formatters for simple types such as in the following example, the straightforward implementation does not propagate formatting parameters:
```rust
pub struct Rectangle {
pub origin: Point,
pub size: Size,
}
impl fmt::Display for Rectangle {
fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
write!(formatter, "Rectangle({} at {})", size, origin)
}
}
```
The implementation above unfortunately does not propagate formatting parameters.
```rust
// Prints each field ignoring formatting option, although we asked to show
// only a certain number of digits.
println!("rectangle = {:.2}", my_rectangle);
```
Would it make sense to allow the formatter to preserve its parameters through some syntax or another?
For example:
```rust
impl fmt::Display for Rectangle {
fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
// ~ or some other symbol informs the formatter to use formatting parameters
// from the parent invocation.
write!(formatter, "Rectangle({~} at {~})", size, origin)
}
}
```
|
T-libs-api,C-feature-request
|
low
|
Critical
|
592,859,694 |
flutter
|
Deprecate legacy release tag parsing logic
|
Once it's safe to (full audit of the code would be necessary to determine this), remove the legacy flutter release tag parsing logic introduced in https://github.com/flutter/flutter/pull/53715.
|
team,tool,P3,c: tech-debt,team-tool,triaged-tool
|
low
|
Minor
|
592,881,895 |
flutter
|
Support adding certain widgets inside CustomScrollViews aka, loosen Sliver requirements.
|
I have a CustomScrollView and I want to wrap my widgets with certain widgets that are not slivers, but don't really affect the rendering of the widgets, e.g. `GestureDetector`. I know it's not a sliver, but its child is. I have to then wrap it with a `SliverToBoxAdaptor` and then change its child to a non-sliver component.
Is this something I need to get over or is it something we can add to Flutter?
|
c: new feature,framework,f: scrolling,c: proposal,P3,team-framework,triaged-framework
|
low
|
Major
|
592,889,941 |
angular
|
Inline CSS from Angular Animations lost when reattaching Route Snapshot
|
# 🐞 bug report
### Affected Package
@angular/animations
@angular/router
### Is this a regression?
Behavior started with angular 9
### Description
When using a Custom Route Reuse strategy the css that had been applied via an angular animation disappears leaving the element in neither "animation state"
## 🔬 Minimal Reproduction
https://github.com/IanNMarshall/angular9-animations-example
#### Steps to reproduce
```
$ ng serve
```
1. Click the "animation" button to be routed to /animation
2. Toggle the animation any number of times or none at all
3. Click home to navigate away
4. Click "animation" again and the animation css has been removed, it is in neither the open or closed state.
5. Additionally if you now toggle the animation it works for all future routing
#### Working example
Same src as above example but pre angular 9, works as expected, keeps animation state.
https://github.com/IanNMarshall/angular8-animation-example
## 🌍 Your Environment
```
Angular CLI: 9.1.0
Node: 12.13.0
OS: darwin x64
Angular:
...
Ivy Workspace:
Package Version
------------------------------------------------------
@angular-devkit/architect 0.901.0
@angular-devkit/core 9.1.0
@angular-devkit/schematics 9.1.0
@schematics/angular 9.1.0
@schematics/update 0.901.0
rxjs 6.5.4
```
**Anything else relevant?**
Could not mimic this behavior with [style] binding or setAttribute("style") seems to only happen with animations.
|
type: bug/fix,area: animations,freq2: medium,area: router,state: needs more investigation,P3
|
low
|
Critical
|
592,898,860 |
rust
|
Cloned sockets are inherited by child processes on windows
|
Given this test code running on Windows 10 compiled with rustc 1.41.0:
```rust
use std::{net::UdpSocket, os::windows::io::AsRawSocket, process::Command};
use winapi::{
shared::minwindef::DWORD,
um::{handleapi, winbase, winnt},
};
fn main() -> std::io::Result<()> {
let socket1 = UdpSocket::bind("0.0.0.0:9999")?;
let socket2 = socket1.try_clone()?;
let mut flags: DWORD = 0;
unsafe {
handleapi::GetHandleInformation(socket1.as_raw_socket() as winnt::HANDLE, &mut flags);
}
println!(
"socket 1 HANDLE_FLAG_INHERIT is set: {}",
((flags & winbase::HANDLE_FLAG_INHERIT) != 0)
);
unsafe {
handleapi::GetHandleInformation(socket2.as_raw_socket() as winnt::HANDLE, &mut flags);
}
println!(
"socket 2 HANDLE_FLAG_INHERIT is set: {}",
((flags & winbase::HANDLE_FLAG_INHERIT) != 0)
);
Command::new("notepad.exe").spawn()?;
Ok(())
}
```
I would expect UDP port 9999 to no longer be listening after the code terminates and I would expect both `socket1` and `socket2` to report `false` on the state of their `HANDLE_FLAG_INHERIT`.
Instead port 9999 remains in a listening state until I close the spawned child process `notepad` and `socket2` shows that its `HANDLE_FLAG_INHERIT` is set.
This appears to have been introduced in rust 1.38.0 via PR #60260 which now sets the `WSA_FLAG_NO_HANDLE_INHERIT` in `WSASocketW` instead of explicitly clearing the `HANDLE_FLAG_INHERIT` with `SetHandleInformation`. Just to confirm, I have compiled this against `nightly-2019-07-26-x86_64-pc-windows-msvc` (before PR) and `nightly-2019-07-27-x86_64-pc-windows-msvc` (after PR) and this code behaves as expected before the PR.
On the surface that PR looks solid and I would totally expect it to work. However it seems (surprisingly) that setting the `WSA_FLAG_NO_HANDLE_INHERIT` flag on a duplicated socket is not effective. I have googled until I could google no longer and alas have found no mention of this behavior.
As a workaround, doing something like:
```
let in_socket = socket.try_clone()?;
unsafe {
handleapi::SetHandleInformation(in_socket.as_raw_socket() as winnt::HANDLE, winbase::HANDLE_FLAG_INHERIT, 0)
};
```
fixes my use case.
Perhaps the correct fix is for [`duplicate`](https://github.com/rust-lang/rust/blob/537ccdf3ac44c8c7a8d36cbdbe6fb224afabb7ae/src/libstd/sys/windows/net.rs#L189) to call [`set_no_inherit`](https://github.com/rust-lang/rust/blob/537ccdf3ac44c8c7a8d36cbdbe6fb224afabb7ae/src/libstd/sys/windows/net.rs#L356) unless it is a UWP app. However I would love it if someone had deeper knowledge as to why this behavior is as it is.
|
O-windows,T-libs-api,C-bug,A-io,A-process
|
low
|
Minor
|
592,925,509 |
flutter
|
[gen_l10n] Allow developer to select subset of localizations to add to application
|
Currently, all localizations files are generated in a Flutter application using the `gen_l10n` tool. We should consider including ways that the Flutter developer can select which localizations to include in their application. The only way right now is to manually modify the lookup method in the localizations delegate class and remove unneeded Dart files from their app, but it would be nice to have some built-in functionality to do this.
For example:
1. Flutter app contains localizations for 15 locales. However, it only wants 2-3 locales for a specific version of their app.
2. Developer can run `gen_l10n` tool, then remove localizations files they do not want and modify generate localizations delegate class to only return the localization delegates they're interested in.
|
c: new feature,framework,a: internationalization,c: proposal,P3,team-framework,triaged-framework
|
low
|
Minor
|
592,937,970 |
go
|
cmd/go: poor error message when resolving an import path that is missing from the main module
|
### What version of Go are you using (`go version`)?
<pre>
$ go version
go version go1.14 windows/amd64
</pre>
### Does this issue reproduce with the latest release?
Yes, as of 1.14
### What operating system and processor architecture are you using (`go env`)?
<details><summary><code>go env</code> Output</summary><br><pre>
$ go env
set GO111MODULE=
set GOARCH=amd64
set GOBIN=
set GOCACHE=REDACTED
set GOENV=REDACTED
set GOEXE=.exe
set GOFLAGS=
set GOHOSTARCH=amd64
set GOHOSTOS=windows
set GOINSECURE=
set GONOPROXY=
set GONOSUMDB=
set GOOS=windows
set GOPATH=REDACTED
set GOPRIVATE=
set GOPROXY=https://proxy.golang.org,direct
set GOROOT=c:\go
set GOSUMDB=sum.golang.org
set GOTMPDIR=
set GOTOOLDIR=c:\go\pkg\tool\windows_amd64
set GCCGO=gccgo
set AR=ar
set CC=gcc
set CXX=g++
set CGO_ENABLED=1
set GOMOD=REDACTED
set CGO_CFLAGS=-g -O2
set CGO_CPPFLAGS=
set CGO_CXXFLAGS=-g -O2
set CGO_FFLAGS=-g -O2
set CGO_LDFLAGS=-g -O2
set PKG_CONFIG=pkg-config
set GOGCCFLAGS=-m64 -mthreads -fmessage-length=0 -fdebug-prefix-map=REDACTED=/tmp/go-build -gno-record-gcc-switches
</pre></details>
### What did you do?
Given a `module example.com/repo` defined in `./go.mod`, if directory `./a` does not contain any Go files and a package `./a/b` contains `import "example.com/repo/a"`, `go build ./a/b` will attempt to fetch `example.com/repo/a` even though `./a` is clearly within a module.
Reproduced here: https://github.com/firelizzard18/golang-issue-38224
### What did you expect to see?
<pre>
$ go build ./a/b
./a is not a package (it contains no .go files)
</pre>
### What did you see instead?
<pre>
$ go build ./a/b
go: finding module for package github.com/firelizzard18/golang-issue/a
a\b\b.go:4:2: no matching versions for query "latest"
</pre>
<details><summary><code>go build -c -v ./a/b</code> Output</summary><br><pre>
$ go build -x -v ./a/b
WORK=REDACTED\go-build863876662
go: finding module for package github.com/firelizzard18/golang-issue/a
# get https://proxy.golang.org/github.com/@v/list
# get https://proxy.golang.org/github.com/firelizzard18/@v/list
# get https://proxy.golang.org/github.com/firelizzard18/golang-issue/a/@v/list
# get https://proxy.golang.org/github.com/@v/list: 410 Gone (0.136s)
# get https://proxy.golang.org/github.com/firelizzard18/@v/list: 410 Gone (0.244s)
# get https://proxy.golang.org/github.com/firelizzard18/golang-issue/a/@v/list: 410 Gone (0.250s)
# get https://github.com/?go-get=1
mkdir -p REDACTED_GOPATH\pkg\mod\cache\vcs # git3 https://github.com/firelizzard18/golang-issue
# lock REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6.lock# REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6 for git3 https://github.com/firelizzard18/golang-issue
cd REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6; git ls-remote -q origin
# get https://github.com/?go-get=1: 200 OK (0.249s)
0.541s # cd REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6; git ls-remote -q origin
cd REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6; git -c log.showsignature=false log -n1 '--format=format:%H %ct %D' aae63bbecdc4e4b7f7eb0bde21b827dbb2e3fe49 --
0.038s # cd REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6; git -c log.showsignature=false log -n1 '--format=format:%H %ct %D' aae63bbecdc4e4b7f7eb0bde21b827dbb2e3fe49 --
cd REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6; git for-each-ref --format %(refname) refs/tags --merged aae63bbecdc4e4b7f7eb0bde21b827dbb2e3fe49
0.038s # cd REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6; git for-each-ref --format %(refname) refs/tags --merged aae63bbecdc4e4b7f7eb0bde21b827dbb2e3fe49
cd REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6; git for-each-ref --format %(refname) refs/tags --merged aae63bbecdc4e4b7f7eb0bde21b827dbb2e3fe49
0.037s # cd REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6; git for-each-ref --format %(refname) refs/tags --merged aae63bbecdc4e4b7f7eb0bde21b827dbb2e3fe49
cd REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6; git -c log.showsignature=false log -n1 '--format=format:%H %ct %D' aae63bbecdc4 --
0.038s # cd REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6; git -c log.showsignature=false log -n1 '--format=format:%H %ct %D' aae63bbecdc4 --
cd REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6; git cat-file blob aae63bbecdc4e4b7f7eb0bde21b827dbb2e3fe49:a/go.mod
0.037s # cd REDACTED_GOPATH\pkg\mod\cache\vcs\e2313c495adbffd1d40784db88daaf99dcbe9402845107a75ca579ef28ad5de6; git cat-file blob aae63bbecdc4e4b7f7eb0bde21b827dbb2e3fe49:a/go.mod
a\b\b.go:4:2: no matching versions for query "latest"
</pre></details>
|
NeedsFix,modules
|
low
|
Critical
|
592,946,376 |
flutter
|
[web] add other locales to a11y enable button
|
The button to enable a11y only has aria-label in english.
We should internationalize the aria-label to make sure that a11y instruction is understandable.
|
c: new feature,engine,a: accessibility,platform-web,P2,team-web,triaged-web
|
low
|
Minor
|
592,952,013 |
godot
|
5.1 AudioStreams not playing all channels
|
**Godot version:**
b0eca5828
**OS/device including version:**
Windows/Linux
**Issue description:**
Trying to play a 5.1 ogg file in godot causes it to only play a few channels, this results in the audio playback not being what it should be.
**Steps to reproduce:**
Use a 5.1 ogg audio file and try to play it through AudioStreamPlayer
I do not have any 5.1 vorbis audio file that I can share on me, so if anyone could provide it for testing that would be great.
|
bug,topic:audio
|
low
|
Minor
|
592,970,142 |
flutter
|
CanvasKit backend should collect "text_layout" benchmark data
|
We currently track this for the DOM backend. We should do the same for CanvasKit.
/cc @mdebbar @hterkelsen
|
engine,platform-web,e: web_canvaskit,P3,team-web,triaged-web
|
low
|
Minor
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.