id
int64 393k
2.82B
| repo
stringclasses 68
values | title
stringlengths 1
936
| body
stringlengths 0
256k
β | labels
stringlengths 2
508
| priority
stringclasses 3
values | severity
stringclasses 3
values |
|---|---|---|---|---|---|---|
542,507,562 | terminal | 'MODE CON LINES=100' kills command prompt | # Environment
```none
Windows build number: Microsoft Windows [Version 10.0.18363.535]
Windows Terminal version (if applicable): 0.7.3451.0
```
# Steps to reproduce
Open command prompt ('CMD'), and type 'MODE CON LINES=100'
# Expected behavior
The screen is cleared, and prompt is shown at upper left corner.
# Actual behavior
No prompt is show (at least characters of prompt are invisible).
| Product-Conpty,Area-Server,Issue-Bug,Priority-2 | low | Minor |
542,513,769 | flutter | FadeInImage performs fade-out and fade-in simultaneously | ## Proposal
FadeInImage fade-out the placeholder, then fade-in the target image.
In other words, at the middle point, there is a moment when neither is displayed.
I feel this as a flickering image.
I think that the fade-out of the placeholder and the fade-in of the target image should be performed simultaneously.
| c: new feature,framework,would be a good package,c: proposal,P3,team-framework,triaged-framework | low | Minor |
542,525,091 | flutter | I need buffer update event instead of ValueNotifier in video_player | currently ,VideoPlayerController implements ValueNotifier, but as a listener, I am more interested in designated change such as buffer update;
could this feature be added;
through we can use
```
VideoPlayerPlatform.instance.videoEventsFor(_controller.textureId).listen(eventListener, onError: errorListener);
```
any textureId is exported for only for test;
and also there are so many same buffer update events;
```
I/flutter (25101): [DurationRange(start: 0:00:00.000000, end: 0:00:57.152000)]
I/flutter (25101): VideoEventType.bufferingUpdate
I/flutter (25101): [DurationRange(start: 0:00:00.000000, end: 0:00:57.152000)]
I/flutter (25101): VideoEventType.bufferingUpdate
I/flutter (25101): [DurationRange(start: 0:00:00.000000, end: 0:00:57.152000)]
I/flutter (25101): VideoEventType.bufferingUpdate
I/flutter (25101): [DurationRange(start: 0:00:00.000000, end: 0:00:57.152000)]
```
so I request two things, 1: designated events; 2: no duplicated events; | c: new feature,p: video_player,package,c: proposal,team-ecosystem,P3,triaged-ecosystem | low | Critical |
542,561,049 | terminal | Terminal window is moved off-screen in 2 screen setup after a single-screen RDP session | # Environment
```none
Windows build number: 10.0.18363.535
Windows Terminal version (if applicable): 0.7.3451.0
```
# Steps to reproduce
1. Have a dual-screen setup with primary screen on the right, secondary on the left.
2. Open the Windows Terminal on a primary (right) screen.
3. Lock the workstation, go somewhere and connect via RDP using 1 screen configuration.
4. WT windows are observed on the RDP session single screen.
5. End RDP session, go back.
6. All WT windows are moved off-screen, past the right edge of the primary (right) screen.
# Expected behavior
WT windows should stay on the primary (right) screen.
# Actual behavior
WT windows are displayed off-screen. It is possible to drag them back via ALT+Space -> Move but is very annoying. | Help Wanted,Needs-Repro,Issue-Bug,Area-UserInterface,Product-Terminal,Priority-3 | low | Minor |
542,568,069 | flutter | [macOS] Allow using integrated GPU to save battery | Even after setting `NSSupportsAutomaticGraphicsSwitching` to `YES` in `macos/Runner/Info.plist` the app always uses the discrete GPU.
```
[β] Flutter (Channel master, v1.13.6-pre.23, on Mac OS X 10.15.2 19C57, locale en-CH)
β’ Flutter version 1.13.6-pre.23 at /Users/ben/flutter
β’ Framework revision 6e72facad1 (2 days ago), 2019-12-24 10:54:04 -0800
β’ Engine revision 33813929e3
β’ Dart version 2.8.0 (build 2.8.0-dev.0.0 886615d0f9)
[β] Android toolchain - develop for Android devices (Android SDK version 28.0.3)
β’ Android SDK at /Users/ben/Library/Android/sdk
β’ Android NDK location not configured (optional; useful for native profiling support)
β’ Platform android-29, build-tools 28.0.3
β’ Java binary at: /Applications/Android Studio.app/Contents/jre/jdk/Contents/Home/bin/java
β’ Java version OpenJDK Runtime Environment (build 1.8.0_202-release-1483-b49-5587405)
β’ All Android licenses accepted.
[β] Xcode - develop for iOS and macOS (Xcode 11.2.1)
β’ Xcode at /Applications/Xcode.app/Contents/Developer
β’ Xcode 11.2.1, Build version 11B53
β’ CocoaPods version 1.8.3
[β] Chrome - develop for the web
β’ Chrome at /Applications/Google Chrome.app/Contents/MacOS/Google Chrome
[β] Android Studio (version 3.5)
β’ Android Studio at /Applications/Android Studio.app/Contents
β’ Flutter plugin version 42.1.1
β’ Dart plugin version 191.8593
β’ Java version OpenJDK Runtime Environment (build 1.8.0_202-release-1483-b49-5587405)
[β] VS Code (version 1.41.1)
β’ VS Code at /Applications/Visual Studio Code.app/Contents
β’ Flutter extension version 3.7.1
[β] Connected device (3 available)
β’ macOS β’ macOS β’ darwin-x64 β’ Mac OS X 10.15.2 19C57
β’ Chrome β’ chrome β’ web-javascript β’ Google Chrome 79.0.3945.88
β’ Web Server β’ web-server β’ web-javascript β’ Flutter Tools
β’ No issues found!
``` | c: new feature,engine,c: performance,platform-mac,a: desktop,perf: energy,P3,team-macos,triaged-macos | low | Major |
542,577,223 | PowerToys | Integrate clang-tidy linter | # Summary of the new feature/enhancement
[clang-tidy](https://clang.llvm.org/extra/clang-tidy) is a great C++ linter with many useful automated refactorings. We may integrate it to our pipeline, starting with [the formatting script](https://github.com/microsoft/PowerToys/blob/master/format_sources.ps1). Unlike clang-format, it requires a [compilation database](https://clang.llvm.org/docs/JSONCompilationDatabase.html) which we need to provide somehow, [e.g. using this](https://github.com/CoatiSoftware/vs-sourcetrail). | Idea-Enhancement,Area-Quality,Cost-Medium | low | Major |
542,599,302 | TypeScript | Union type checking during assignment fails for boolean and passes for other primitives | **TypeScript Version:** 3.7.2
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:**
number string union
**Expected behavior:**
Compiler should be giving an error on assigning `demo1`, like it does with `demo2`.
**Actual behavior:**
Doesn't seem to properly discriminate between the `string` and other types that are conceptually unbound: `number`/`RegExp` etc.
**Related Issues:**
#20863
https://github.com/microsoft/TypeScript/issues/30506
https://github.com/microsoft/TypeScript/issues/9254
**Code**
```ts
type t1 = {
a: number;
b: string;
} | {
a: string;
};
// no error. Invalid assignment is possible
const demo1: t1 = {
a: '123',
b: 'asd',
};
type t2 = {
a: boolean;
b: string;
} | {
a: string;
};
// error Error: TS2322: Type '{ a: string; b: string; }' is not assignable to type 't2'. Object literal may only specify known properties, and 'b' does not exist in type '{ a: string; }'.
const demo2: t2 = {
a: '123',
b: 'asd',
};
```
<details><summary><b>Output</b></summary>
```ts
"use strict";
// no error
const demo1 = {
a: '123',
b: 'asd',
};
// error
const demo2 = {
a: '123',
b: 'asd',
};
```
</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/?ssl=25&ssc=14&pln=1&pc=1#code/C4TwDgpgBMCMUF4oG8BQUNQIYC4oDsBXAWwCMIAnAbnU1LwGdgKBLfAcxoF8oAfFWhlxQmrDtxqoA9FIIB7KJQpyKqAMZz8TKABMIxObDxxEAzNjwByWACYAzJYA0gqPSiWsDHU9RdJoSBgbUzRzYVI5OQAbCCx8GnM3UTZOXz4zTGFk8V9JGUUKZVV1TW09AxtjYKRQzKtbB2dEq09vZz8gA)
| Bug | low | Critical |
542,658,467 | neovim | "q." command (history of dot-repeatable actions) | ### Actual behaviour
I like the functionality of `q /` and `q :` where it will display a list of past executed searches/commands. There are many times where I am doing some edits and would like to repeat that action with `.` later in normal mode somewhere else, however often times I end up performing some other edit action so `.` would no longer be relevant but if I could type `q .` then retrieve a list of most recently used editing actions then I would have at my fingertips all past actions and it would be awesome to use anyone at my command
### Expected behaviour
Typing `q .` would bring up a list of past editing actions that I can press enter to execute immediately just like how `q /` would bring up a list of past searches or `q :` would bring up a list of past executed commands.
| enhancement,core,multicursor | low | Minor |
542,664,464 | go | crypto/tls: improve server-side (lack of) renegotiation docs and error message | <!-- Please answer these questions before submitting your issue. Thanks! -->
### What version of Go are you using (`go version`)?
<pre>
$ go version
go version go1.13.5 linux/amd64
</pre>
### Does this issue reproduce with the latest release?
yes
### What operating system and processor architecture are you using (`go env`)?
`amd64/linux`
### What did you do?
<!--
If possible, provide a recipe for reproducing the error.
A complete runnable program is good.
A link on play.golang.org is best.
-->
I am unable to renegotiate successfully using TLS 1.2 to a Go server I wrote. https://play.golang.org/p/Q39eh8PVAOP. After starting the server, I connect using openssl:
`openssl s_client -connect 172.16.28.75:10000 -tls1_2`. Following a successful handshake, and some test messages going through, I start a renegotiation by sending `R` in the openssl client session. The renegotiation fails because the Go server expects a `helloRequestMessage` as the first message in the renegotiation while the`openssl` client sends a `clientHelloMessage` as the first message in the renegotiation.
### What did you expect to see?
A successful renegotiation as demonstrated when running the same server using `openssl s_server -key /home/vagrant/easy-rsa/easyrsa3/pki/private/server.key -cert /home/vagrant/easy-rsa/easyrsa3/pki/issued/server.crt -accept 10000`.
And output from renegotiation on the client:
```
R
RENEGOTIATING
depth=0 CN = server
verify error:num=20:unable to get local issuer certificate
verify return:1
depth=0 CN = server
verify error:num=21:unable to verify the first certificate
verify return:1
```
### What did you see instead?
On the server:
```
$ go run test.go
2019/12/26 15:30:25 New connection
2019/12/26 15:30:26 tls: received unexpected handshake message of type *tls.clientHelloMsg when waiting for *tls.helloRequestMsg
exit status 1
```
On the client:
```
---
R
RENEGOTIATING
4684979820:error:140043F2:SSL routines:CONNECT_CR_SRVR_HELLO:sslv3 alert unexpected message:/BuildRoot/Library/Caches/com.apple.xbs/Sources/libressl/libressl-47.11.1/libressl-2.8/ssl/ssl_pkt.c:1200:SSL alert number 10
4684979820:error:140040E5:SSL routines:CONNECT_CR_SRVR_HELLO:ssl handshake failure:/BuildRoot/Library/Caches/com.apple.xbs/Sources/libressl/libressl-47.11.1/libressl-2.8/ssl/ssl_pkt.c:585:
```
| Documentation,help wanted,NeedsFix | low | Critical |
542,690,442 | vscode | Auto save & new line prevents undo from working | - Version: 1.41.0
- Commit: 9579eda04fdb3a9bba2750f15193e5fafe16b959
- Date: 2019-12-11T17:58:38.338Z
- Electron: 6.1.5
- Chrome: 76.0.3809.146
- Node.js: 12.4.0
- V8: 7.6.303.31-electron.0
- OS: Darwin x64 19.2.0
Steps to Reproduce:
1. set `"files.trimFinalNewlines": true` and `"files.autoSave": "afterDelay"`; _edit: instead set `"files.insertFinalNewline": true` to follow POSIX conventions_
2. move the caret to the very end of the file
3. insert a string at the very end of a file
4. a new line is automatically added
5. try to undo step 3 will remove the new line, which is then added again
Does this issue occur when all extensions are disabled?: Yes | bug,undo-redo | low | Major |
542,705,939 | rust | panic=abort testing / subprocess testing | This is the tracking issue for `-Zpanic-abort-tests` and related functionality.
Specifically, this flag switches the default test strategy of (loosely) a thread per test to a process per test, and implements the associated functionality of collecting process failures and treating that as passing/failing the test. It also includes a currently unstable option in test binaries, `--force-run-in-process`.
This was originally added in https://github.com/rust-lang/rust/pull/64158 and Cargo support landed in https://github.com/rust-lang/cargo/pull/7460. | T-libs-api,C-tracking-issue,Libs-Tracked | low | Critical |
542,728,344 | flutter | Remove "dartpad" mode in snippet generator, and just put dartpad for all snippets | cc @RedBrogdon @gspencergoog
While we're at it, we should probably rename "snippet" and "sample" to be more intuitive. I always think of "snippet" as being something too small to compile, and "sample" as being an example that can be tested, which is the opposite of reality as far as I can tell.
| team,framework,d: api docs,P2,team-framework,triaged-framework | low | Minor |
542,729,664 | godot | Addon zipped with windows 10 unable to load because of required ZIP version | **Godot version:**
Godot_v3.1.1-stable_win64
**OS/device including version:**
WIn10
**Issue description:**
editor/editor_asset_installer.cpp:175 - Condition ' !dir_map.has(ppath) ' is true. Continuing..:
**Steps to reproduce:**
1. create addon
2. use right click -> send to -> Compressed (zipped) folder
3. AssetLib -> Import
**Minimal reproduction project:**
[test.zip](https://github.com/godotengine/godot/files/4004017/test.zip)
Related Issue: the imported addon will be added to the root project folder
1. download addon
https://godotengine.org/asset-library/asset/457
https://github.com/samdze/godot-modifiers-plugin/archive/4e88b9812c26d35673d6c12198aeb2f8cb2d9425.zip
2. import addon | bug,topic:core,topic:editor,confirmed,topic:thirdparty | low | Major |
542,762,051 | godot | Transparent + no culling + deep parallax depth crashes editor | **Godot version:**
Godot 3.2 beta 4
**OS/device including version:**
Windows 10
**Issue description:**
Transparent + no culling + deep parallax depth combination in SpatialMaterial crashes editor. I *think* it displays correctly in editor's main viewport and runtime, but if you try to see mesh preview (not material preview) in inspector, editor crashes without any error message.
**Steps to reproduce:**
- Add SpatialMaterial to a mesh
- Set transparent `true`, select cull mode `disable`
- Enable depth, and enable Deep Parallax and add depth texture
- Try to see mesh preview in the inspector panel
**Minimal reproduction project:**
[SpatialCrash.zip](https://github.com/godotengine/godot/files/4004441/SpatialCrash.zip)
| bug,topic:rendering,confirmed,crash | low | Critical |
542,771,629 | youtube-dl | https://study.com/ | <!--
######################################################################
WARNING!
IGNORING THE FOLLOWING TEMPLATE WILL RESULT IN ISSUE CLOSED AS INCOMPLETE
######################################################################
-->
## Checklist
<!--
Carefully read and work through this check list in order to prevent the most common mistakes and misuse of youtube-dl:
- First of, make sure you are using the latest version of youtube-dl. Run `youtube-dl --version` and ensure your version is 2019.12.25. If it's not, see https://yt-dl.org/update on how to update. Issues with outdated version will be REJECTED.
- Make sure that all provided video/audio/playlist URLs (if any) are alive and playable in a browser.
- Make sure that site you are requesting is not dedicated to copyright infringement, see https://yt-dl.org/copyright-infringement. youtube-dl does not support such sites. In order for site support request to be accepted all provided example URLs should not violate any copyrights.
- Search the bugtracker for similar site support requests: http://yt-dl.org/search-issues. DO NOT post duplicates.
- Finally, put x into all relevant boxes (like this [x])
-->
- [ ] I'm reporting a new site support request
- [ ] I've verified that I'm running youtube-dl version **2019.12.25**
- [ ] I've checked that all provided URLs are alive and playable in a browser
- [ ] I've checked that none of provided URLs violate any copyrights
- [ ] I've searched the bugtracker for similar site support requests including closed ones
## Example URLs
<!--
Provide all kinds of example URLs support for which should be included. Replace following example URLs by yours.
-->
- Single video: https://study.com/academy/topic/the-evolution-of-organizational-behavior.html
- Single video: https://study.com/academy/topic/the-evolution-of-organizational-behavior.html
- Playlist: https://study.com/academy/course/anatomy-physiology-course.html
## Description
<!--
Provide any additional information.
If work on your issue requires account credentials please provide them or explain how one can obtain them.
-->
WRITE DESCRIPTION HERE
| site-support-request | low | Critical |
542,777,198 | pytorch | update embedding at indices, other than those passed as input, in the case of sparse tensors | when we use,
```
x = nn.Embedding(10, 5, sparse=True)
```
then, only the embeddings at indices that are passed as input get updated, for example,
```
y = x(torch.LongTensor([0])).sum()
y.backward()
for param in x.parameters():
print(param.grad)
```
gives
```
tensor(indices=tensor([[0]]),
values=tensor([[1., 1., 1., 1., 1.]]),
size=(10, 5), nnz=1, layout=torch.sparse_coo)
```
but if we want to update embedding of words in our vocabulary that our model predicted, then there is no way to do that, or changing requires_grad for a word that is not passed as input is also not possible, that is doing something like,
```
x.index[1].requires_grad = True
x.index[8].requires_grad = True
```
cannot be done, and only embedding of indices that are passed as input to this sparse tensor get updated.
there is no way to update embedding of a word that is not passed as input when using sparse embedding, and if we set sparse=False, then it will update all word embedding.
cc @vincentqb | module: sparse,module: nn,triaged,enhancement | low | Major |
542,799,541 | flutter | Allow specifying --user-data-dir for chrome launch | i want to use --user-data-dir save my user data ,how to do ?
like this
`
flutter run -d chrome --web-port=8686 --user-data-dir=~/Documents/chrome
`
| c: new feature,tool,platform-web,P3,team-web,triaged-web | medium | Major |
542,842,503 | pytorch | pytorch out of memory when calculation squared difference of unfold | ## π Bug
> RuntimeError: CUDA out of memory. Tried to allocate 90.73 GiB (GPU 0; 1.96 GiB total capacity; 12.11 MiB already allocated; 1.13 GiB free; 1.89 MiB cached)
## To Reproduce
Steps to reproduce the behavior:
```
import torch
x = torch.rand(1,3,1000,1000).cuda()
print(x.shape)
t = torch.rand(1,3,100,100).cuda()
print(t.shape)
y = x.unfold(dimension=2, size=t.shape[2], step=1).unfold(dimension=3, size=t.shape[3], step=1)
print(y.shape)
ssd = torch.einsum('xcijkl,xckl->xcij', y, t)
ssd *= - 2
ssd += torch.einsum('xcijkl, xcijkl->xcij', y, y)
ssd += torch.einsum('xcij, xcij', t, t)
```
## Environment
> Collecting environment information...
> PyTorch version: 1.2.0
> Is debug build: No
> CUDA used to build PyTorch: 10.0.130
>
> OS: Ubuntu 19.10
> GCC version: (Ubuntu 9.2.1-9ubuntu2) 9.2.1 20191008
> CMake version: Could not collect
>
> Python version: 3.7
> Is CUDA available: Yes
> CUDA runtime version: Could not collect
> GPU models and configuration: GPU 0: GeForce 940MX
> Nvidia driver version: 435.21
> cuDNN version: Could not collect
>
> Versions of relevant libraries:
> [pip3] numpy==1.16.3
> [pip3] pytorch-ignite==0.1.2
> [pip3] torch==1.2.0
> [pip3] torchsummary==1.5.1
> [pip3] torchvision==0.4.0
> [conda] Could not collect
When I try to calculate the sum of squared difference between an image and a template I get the error out of memory. Even on cpu its not working.
> RuntimeError: [enforce fail at CPUAllocator.cpp:64] . DefaultCPUAllocator: can't allocate memory: you tried to allocate 97416120000 bytes. Error code 12 (Cannot allocate memory)
I can do the same thing with opencv with no problems.
```
import numpy as np
import cv2
x = np.random.uniform(size=(1000,1000,3))
x = np.array(x*255., dtype=np.uint8)
print(x.shape)
t = np.random.uniform(size=(100,100,3))
t = np.array(t*255., dtype=np.uint8)
print(t.shape)
ssd = cv2.matchTemplate(x, t, cv2.TM_SQDIFF)
print(ssd.shape)
```
cc @fmassa | module: memory usage,triaged | low | Critical |
542,842,582 | kubernetes | Namespace deletion fails because object does not implement protobuf marshalling interface | <!-- Please use this template while reporting a bug and provide as much info as possible. Not doing so may result in your bug not being addressed in a timely manner. Thanks!
If the matter is security related, please disclose it privately via https://kubernetes.io/security/
-->
**What happened**:
Namespace stuck in deletion because API resource does not implement the protobuf marshalling interface.
**What you expected to happen**:
kube-controller-manager to be able to terminate the namespace.
**How to reproduce it (as minimally and precisely as possible)**:
1. Setup [email protected] running on Kubernetes v1.16.4 (minikube provider) using the minikube walkthrough guide - https://github.com/kubernetes/sample-apiserver/blob/v0.16.4/docs/minikube-walkthrough.md.
2. Create namespace
```
$ k create ns foo
```
3. Delete the namespace and ensure the deletion fails
```
$ k delete ns foo
$ k get ns foo -o yaml
apiVersion: v1
kind: Namespace
metadata:
creationTimestamp: "2019-12-15T19:41:31Z"
deletionTimestamp: "2019-12-15T19:41:33Z"
name: foo
resourceVersion: "798"
selfLink: /api/v1/namespaces/foo
uid: e5b47622-6789-4f41-a2bd-eb04aadc6d90
spec:
finalizers:
- kubernetes
status:
conditions:
- lastTransitionTime: "2019-12-15T19:41:38Z"
message: All resources successfully discovered
reason: ResourcesDiscovered
status: "False"
type: NamespaceDeletionDiscoveryFailure
- lastTransitionTime: "2019-12-15T19:41:38Z"
message: All legacy kube types successfully parsed
reason: ParsedGroupVersions
status: "False"
type: NamespaceDeletionGroupVersionParsingFailure
- lastTransitionTime: "2019-12-15T19:41:38Z"
message: 'Failed to delete all resource types, 1 remaining: object *v1alpha1.FlunderList
does not implement the protobuf marshalling interface and cannot be encoded
to a protobuf message'
reason: ContentDeletionFailed
status: "True"
type: NamespaceDeletionContentFailure
phase: Terminating
```
```
$ k logs wardle-server-lgkcm -n wardle wardle-server
E1227 11:06:58.528242 1 writers.go:105] apiserver was unable to write a JSON response: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:06:58.909046 1 writers.go:105] apiserver was unable to write a JSON response: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:06:59.321062 1 writers.go:105] apiserver was unable to write a JSON response: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:07:00.314837 1 writers.go:105] apiserver was unable to write a JSON response: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:07:01.743636 1 writers.go:105] apiserver was unable to write a JSON response: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:07:04.592075 1 writers.go:105] apiserver was unable to write a JSON response: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:07:09.755761 1 writers.go:105] apiserver was unable to write a JSON response: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:07:20.189966 1 writers.go:105] apiserver was unable to write a JSON response: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:07:40.774322 1 writers.go:105] apiserver was unable to write a JSON response: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:08:21.927852 1 writers.go:105] apiserver was unable to write a JSON response: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:09:43.951662 1 writers.go:105] apiserver was unable to write a JSON response: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:11:57.372764 1 writers.go:105] apiserver was unable to write a JSON response: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
```
```
$ k logs -n kube-system kube-controller-manager-minikube
E1227 11:07:01.906469 1 namespace_controller.go:148] deletion of namespace foo failed: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:07:04.611506 1 namespace_controller.go:148] deletion of namespace foo failed: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:07:09.862402 1 namespace_controller.go:148] deletion of namespace foo failed: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:07:20.210405 1 namespace_controller.go:148] deletion of namespace foo failed: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:07:40.807497 1 namespace_controller.go:148] deletion of namespace foo failed: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:08:21.965230 1 namespace_controller.go:148] deletion of namespace foo failed: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:09:44.008476 1 namespace_controller.go:148] deletion of namespace foo failed: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:11:57.431584 1 namespace_controller.go:148] deletion of namespace foo failed: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
E1227 11:16:57.455757 1 namespace_controller.go:148] deletion of namespace foo failed: object *v1alpha1.FlunderList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message
```
**Anything else we need to know?**:
k8s.io/sample-apiserver types do not implement the protobuf marshalling interface but the kube-controller-manager does not fallback to json.
Is it a hard requirement for Kubernetes v1.16 the api resources to implement the protobuf marshaling interface or I miss some configuration of kube-contoroller-manager?
**Environment**:
- Kubernetes version (use `kubectl version`):
```
$ k version --short
Client Version: v1.16.1
Server Version: v1.16.3
```
- Cloud provider or hardware configuration:
- OS (e.g: `cat /etc/os-release`):
- Kernel (e.g. `uname -a`):
- Install tools: minikube
- Network plugin and version (if this is a network-related bug):
- Others:
| kind/bug,priority/important-soon,sig/api-machinery,lifecycle/frozen | medium | Critical |
542,868,783 | TypeScript | Support for co-dependently conditionally typed arguments | <!-- π¨ STOP π¨ π¦π§π’π£ π¨ πΊπ»πΆπ· π¨
Half of all issues filed here are duplicates, answered in the FAQ, or not appropriate for the bug tracker.
Please help us by doing the following steps before logging an issue:
* Search: https://github.com/Microsoft/TypeScript/search?type=Issues
* Read the FAQ, especially the "Common Feature Requests" section: https://github.com/Microsoft/TypeScript/wiki/FAQ
-->
## Search Terms
- "co dependent arguments"
- "co-dependent arguments"
- "codependent arguments"
- "codependently arguments"
- "codependently typed arguments"
<!-- List of keywords you searched for before creating this issue. Write them down here so that others can find this suggestion more easily -->
## Suggestion
It should be easier to conditionally type a function argument based on the type/value (in the case of enums / string literal types) of a previous argument in the function. Having a more specific way to do this than currently possible should improve editor intelligence and performance.
<!-- A summary of what you'd like to see added or changed -->
## Use Cases
It is quite common to have an API where the type of the second argument depends on the type and/or value of the first. For example, take a look at the following JavaScript code. (It's a fairly trivial example, but it should illustrate the use case.
```js
function reducer(actionType, args) {
switch (actionType) {
case "add":
console.log(args.a + args.b);
break;
case "concat":
console.log(args.firstArr.concat(args.secondArr));
break;
}
}
reducer("add", { a: 1, b: 3 }); // 4
reducer("concat", { firstArr: [1,2], secondArr: [3, 4] }); // [1,2,3,4]
```
It is currently possible to type this, but it is extremely convulted and the editor somewhat struggles to understand this at the call site. This method is based on the fantastic article over at https://artsy.github.io/blog/2018/11/21/conditional-types-in-typescript/
```ts
enum ActionTypeEnum {
add,
concat
}
type Action =
| { type: ActionTypeEnum.add, a: number, b: number }
| { type: ActionTypeEnum.concat, firstArr: any[], secondArr: any[] };
type ActionType = Action["type"];
type ExcludeTypeKey<K> = K extends "type" ? never : K
type ExcludeTypeField<A> = { [K in ExcludeTypeKey<keyof A>]: A[K] }
type ExtractActionParameters<A, T> = A extends { type: T }
? ExcludeTypeField<A>
: never
function reducer<T extends ActionType>(actionType: T, args: ExtractActionParameters<Action, T>) {
const properlyTypedArgs = { ...args, type: actionType } as Action;
switch (properlyTypedArgs.type) {
case ActionTypeEnum.add:
console.log(properlyTypedArgs.a + properlyTypedArgs.b);
break;
case ActionTypeEnum.concat:
console.log(properlyTypedArgs.firstArr.concat(properlyTypedArgs.secondArr));
break;
}
}
reducer(ActionTypeEnum.add, { a: 1, b: 3 }); // 4
reducer(ActionTypeEnum.concat, { firstArr: [1, 2], secondArr: [3, 4] }); // [1,2,3,4]
```
There are four main issues with this approach.
The first, fairly obvious one, is that for the type checker to understand that type of args inside the switch, we're having to merge the two arguments into a single object, cast it, and then switch based on that:
```ts
const properlyTypedArgs = { ...args, type: actionType } as Action;
```
This is an example of the type system leaking into the emitted javascript and is arguably inefficient.
The second is that the editor doesn't really understand what's going on. It understands enough to tell you you've made an error when the type check doesn't validate but also doesn't understand well enough to properly autocomplete. The error could also be easier to understand. See below.
See above how it's suggesting firstArr and secondArr as fields in the second object, even though including them fails the type checker. The only suggestions should be a & b.
The third issue is that it's entirely undiscoverable for all but the most advanced in TypeScript. This is a semi-common use case, and I had no idea how to properly type it. Without this singular article detailing this approach, I never would have discovered it. This should be built into the language to fix this.
Fourthly, and similarly, it's very difficult to read and understand unless you really, really understand generics. I am sure that somebody smarter than I would be able to come up with a more expressive way of detailing the co-dependence of these two types.
<!--
What do you want to use this for?
What shortcomings exist with current approaches?
-->
## Examples
I think I've properly explained this in the use case above.
<!-- Show how this would be used and what the behavior would be -->
## 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,Needs Proposal | low | Critical |
542,896,176 | pytorch | Use of Sequence collections for abstract classes in Dataset | ## π Feature
Using the sequence (or any other similar Python abstract class) for the `Dataset` class in order to tackle the note in lines 30-32 of `torch.utils.data.dataset`:
```python
# No `def __len__(self)` default?
# See NOTE [ Lack of Default `__len__` in Python Abstract Base Classes ]
# in pytorch/torch/utils/data/sampler.py
```
By using the `Sequence` abstract classes, we can see in the official Python documentation that they have the exact abstract methods that the `Dataset` class expects, i.e. `__getitem__` and `__len__`.
## Motivation
My motivation is three-fold: (1) tackling the note mentioned above left by the PyTorch contributors, (2) let the code be auto-documentedβsince accessing this code states explicitly that you need to define those two methods and, (3) using Python abstract classes instead of inheritting from object (which is a rather deprecated practice for when Python did not have abstract classes capabilities).
## Pitch
Implement the Dataset class using new Python abstract classes capabilities, or at least discuss about it.
## Alternatives
I currently do not have alternative proposals.
## Additional context
`None`.
cc @SsnL | feature,module: dataloader,triaged | low | Minor |
542,951,253 | pytorch | Cuda error 59 : device-side assert triggered | Hello,
I'm using an PyTorch implementation of Yolo for a deep learning model, running on AWS EC2 instance (g3s.xlarge with Tesla M60 GPU).
I have an issue when the training process: 8 classes with around 7000 images (for the moment to estimate the results), batch size = 64 (128 also possible), the error is: **Cuda error 59 : Runtime Error: device-side assert triggered**
I found a lot of subject on the same error, but no one answer really clear to the issue.
- I try to increase the number of images (in case the error is because there is not enought targets)
- I try to decrease the batch size
- I decrease the number of classes (a try with 2 classes, same number of images and try with 2 classes and around 2k images)
- I try to change PyTorch and Torchvision versions (using torch==1.0.1.post2 and torchvision==0.2.2.post3)
- I also try to restart all the process on fresh install of my instance
I attach a screenshoot of the error using the CUDA_LAUNCH_BLOCKING=1 to get more accuracy stacktrace, in case it can help
PS: The error come after the first epoch, before run the test function on images
Do you have any idea to fix this ?
Thanks for your help,
Regards,
Florian
cc @ngimel | needs reproduction,module: cuda,triaged | low | Critical |
542,968,290 | TypeScript | commonjs export binding produces invalid code for increment/decrement in PrefixUnaryExpression | <!-- π¨ STOP π¨ π¦π§π’π£ π¨ πΊπ»πΆπ· π¨
Half of all issues filed here are duplicates, answered in the FAQ, or not appropriate for the bug tracker. Even if you think you've found a *bug*, please read the FAQ first, especially the Common "Bugs" That Aren't Bugs section!
Please help us by doing the following steps before logging an issue:
* Search: https://github.com/Microsoft/TypeScript/search?type=Issues
* Read the FAQ: https://github.com/Microsoft/TypeScript/wiki/FAQ
Please fill in the *entire* template below.
-->
<!--
Please try to reproduce the issue with the latest published version. It may have already been fixed.
For npm: `typescript@next`
This is also the 'Nightly' version in the playground: http://www.typescriptlang.org/play/?ts=Nightly
-->
**TypeScript Version:** 3.7.2
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:** export increment decrement syntax error
**Code**
```ts
// @module: commonjs
let foo = 1;
export { foo };
while (!--foo) { }
```
**Expected behavior:**
```js
while (!(exports.foo = --foo)) { }
```
**Actual behavior:**
Because the module-transform uses `onSubstituteNode`, it doesn't use the auto-parenthesizing that's done by the factory API.
```js
while (!exports.foo = --foo) { } // SyntaxError: `!exports.foo` is no valid assignment target
```
**Playground Link:** https://www.typescriptlang.org/play/?target=99&module=1&ssl=3&ssc=19&pln=1&pc=1#code/DYUwLgBAZg9jEF4IEYDcAoEAPADjATpAN7RwQC+GA7gBYCWoEAFAIQC0bsMAlBCeUA
**Related Issues:** #33044
| Bug,Help Wanted | low | Critical |
542,970,055 | TypeScript | ObjectAssignmentRest causes "Property 'foo' does not exist on type '{}'" | <!-- π¨ STOP π¨ π¦π§π’π£ π¨ πΊπ»πΆπ· π¨
Half of all issues filed here are duplicates, answered in the FAQ, or not appropriate for the bug tracker. Even if you think you've found a *bug*, please read the FAQ first, especially the Common "Bugs" That Aren't Bugs section!
Please help us by doing the following steps before logging an issue:
* Search: https://github.com/Microsoft/TypeScript/search?type=Issues
* Read the FAQ: https://github.com/Microsoft/TypeScript/wiki/FAQ
Please fill in the *entire* template below.
-->
<!--
Please try to reproduce the issue with the latest published version. It may have already been fixed.
For npm: `typescript@next`
This is also the 'Nightly' version in the playground: http://www.typescriptlang.org/play/?ts=Nightly
-->
**TypeScript Version:** 3.7.2
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:** rest destructuring assignment does not exist on type
**Code**
```ts
let { foo = {}, ...rest } = {}; // no error in ObjectBindingPattern
({ foo = {} } = {}); // works in ObjectAssignmentPattern without rest property
({ foo = {}, ...rest } = {}); // error: Property 'foo' does not exist on type '{}'
```
**Expected behavior:**
all or none should have an error
**Actual behavior:**
Rest in ObjectAssignmentPattern causes error on unrelated property.
**Playground Link:** https://www.typescriptlang.org/play/?target=99&ssl=3&ssc=30&pln=1&pc=1#code/DYUwLgBA3hBmD28IF5oF8A0EB0uBOIAzpGiugNwBQAFDAkqlKaY2gJRW1yJlNa7YCxCC3QcgA
| Bug | low | Critical |
542,972,081 | TypeScript | commonjs export binding not updated in ForInOrOfStatement | <!-- π¨ STOP π¨ π¦π§π’π£ π¨ πΊπ»πΆπ· π¨
Half of all issues filed here are duplicates, answered in the FAQ, or not appropriate for the bug tracker. Even if you think you've found a *bug*, please read the FAQ first, especially the Common "Bugs" That Aren't Bugs section!
Please help us by doing the following steps before logging an issue:
* Search: https://github.com/Microsoft/TypeScript/search?type=Issues
* Read the FAQ: https://github.com/Microsoft/TypeScript/wiki/FAQ
Please fill in the *entire* template below.
-->
<!--
Please try to reproduce the issue with the latest published version. It may have already been fixed.
For npm: `typescript@next`
This is also the 'Nightly' version in the playground: http://www.typescriptlang.org/play/?ts=Nightly
-->
**TypeScript Version:** 3.7.2
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:** commonjs export for of
**Code**
```ts
// @module: commonjs
// @target: esnext
let foo = '';
export { foo };
for (foo in {}) { }
for (foo of []) {}
for ({foo} of [{foo}]) {}
```
**Expected behavior:**
`for...in` and `for...of` should update the export binding.
```js
for (_a in {}) {
exports.foo = foo = _a;
}
for (_b of []) {
exports.foo = foo = _b;
}
for (_c of [{foo}]) {
exports.foo = foo = _c.foo;
}
```
**Actual behavior:**
Binding is not updated
**Playground Link:** https://www.typescriptlang.org/play/?target=99&module=1#code/DYUwLgBAZg9jEF4IHJkG4BQIAeAHGATpAN7RwQC+msBEAFLPAJYB2ExFAlO5RjfYwgwoEANoBdbhz6F6xRhSEjR8uBUnsKQA
| Bug | low | Critical |
542,972,466 | godot | 3.x: Materials generated from .gltf file are too big in filesize | **Godot version:**
3.1.2
**OS/device including version:**
Windows 10 v1809
**Issue description:**
When you import fully textured .gltf scenes Godot automatically generates .material files for you. Unfortunately there's a flaw in it, because, for example - a door material is 13 mb in size while all the textures are barely 300kb in size. Also, the exported gltf is only 2,6 mb in size. Something's totally not right here.
**Steps to reproduce:**
Make a model in blender, make materials, export to gltf and finally import to godot.
**Minimal reproduction project:**
[material-bug.zip](https://github.com/godotengine/godot/files/4006125/material-bug.zip)
here's a screenshot of all of the wooden floor material files - they're only 718 kb in file size, while the material is 9,2 mb (!)
| discussion,confirmed,documentation,topic:import,topic:3d | medium | Critical |
542,995,379 | TypeScript | import is cut after "transpileModule" compilation when object spread is used (TypeScript 3.7) | **TypeScript Version:** 3.7.0-dev.20190820
**Search Terms:**
- transpileModule
- cut import
- object spread
**Code**
```ts
import { additionalInfo } from './additionalInfo';
async function getObject() {
return {
key: 'value',
}
}
async function main() {
const obj = await getObject();
return {
...obj,
additionalInfo,
};
}
```
**Expected behavior:**
import should stay in compiled version even when `transpileModule` is used.
**Actual behavior:**
import is cut from `transpileModule` compilation output. I met this behaviour during usage of [ts-loader](https://github.com/TypeStrong/ts-loader) with [`transpileOnly: true` option used](https://github.com/TypeStrong/ts-loader#transpileonly). It uses TS's `transpileModule` underneath.
Occurs also on Nightly version of `typescript` (3.8.0-dev.20191224 at the moment).
Error occurs first on 3.7.0-dev.20190820 version. No error on 3.7.0-dev.20190817
More specifically, error occurs after applying this PR - https://github.com/microsoft/TypeScript/pull/32969. So I'm guessing, `obj` somehow gets `unknown` type during compilation. Spreading is not allowed on it and import is cut off from the compilation output.
import will stay after compilation in these cases:
- manual unapplying of that PR on `typescript` source;
- writing return type of `getObject` function;
- manual wrapping of the object with `Promise.resolve()`;
- making `getObject` function not async;
- use normal `tsc` compilation;
**Playground Link:**
Sadly, not reproducible in Typescript Playground.
I constructed example in this repository: https://github.com/sorochanalex/typescript-bug
**Related Issues:**
I have found _something_ similar here - https://github.com/microsoft/TypeScript/issues/15469 - but this issue was resolved.
| Bug,Help Wanted | low | Critical |
543,001,669 | PowerToys | Visual / Text reminder when PowerToys is elevated | From the twitters, two people have recommended we make it clear when we are elevated.
Issue we face is most stuff is extremely transient or works without UX (example FancyZones)
Ideas:
- Append to title like Command prompt and others do
- Change systray icon to visually show we are elevated as well. | Idea-Enhancement | low | Major |
543,102,232 | electron | Support BrowserWindow.maximizable on Linux | ### 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
Setting `maximizable` to false when creating a `BrowserWindow` has no effect on Linux.
For my specific scenario, I've created a borderless window with a maximum width and height. Resizing the window respects the maximum width and height correctly, but if I drag the window to the top of the screen, the window becomes maximized, making it larger than the maximum width and height. I'd like to prevent the window from being maximized.
### Proposed Solution
If `maximizable` is set to false, the `BrowserWindow` should not be maximizable.
### Alternatives Considered
None.
### Additional Information
I originally created #21595 before I discovered this current behaviour is by design. π€¦ββ
Adding this feature _might_ also fix #21508, but I'm not 100% confident of that.
| enhancement :sparkles:,platform/linux,component/BrowserWindow | low | Minor |
543,267,289 | neovim | `starstar` globbing searches through 'wildignore' paths | <!-- Before reporting: search existing issues and check the FAQ. -->
- `nvim --version`: v0.4.3
- `vim -u DEFAULTS` (version: 8.1, patches 1-2268) behaves differently? No, same behaviour.
- Operating system/version: Arch Linux
If a path pattern for a directory is in `'wildignore'`, files under that directory won't be returned in the results of functions like `glob()`. However, the directory will still be searched for. This can be problematic when the ignored directory is very large: functions using path expansions (for example, autocompleting `:edit **<tab>`) could take a very long time to complete, even though none of the files from the ignored folder should be considered.
It would be nice if the glob expansion function (`ExpandOne` in `ex_getln.c` I think?) could consider `'wildignore'` before diving into sub-directories.
See also:
https://github.com/vim/vim/issues/2132
https://groups.google.com/forum/#!topic/vim_dev/SEPEM44lsl8
### Steps to reproduce using `nvim -u NORC`
In this example, `big` should be ignored (I use a symlink to `/usr/share` to easily get a folder large enough to trigger a delay for demonstration purposes.)
```
$ tree
.
βββ src
βββ apples
βΒ Β βββ models.py
βΒ Β βββ views.py
βββ big
βΒ Β βββ data -> /usr/share
βΒ Β βββ models.py
βΒ Β βββ views.py
βββ oranges
βββ models.py
βββ views.py
$ nvim -u NONE -c 'set wildignore=**/big/*'
```
### Actual behaviour
Entering `:e **/m` on the command line and pressing `<tab>` for autocompletion, there will be a large delay before results are displayed in the `wildmenu` (and `src/big/models.py` is correctly not included in those results.)
Similar for `:echo glob('src/**/models.py')` (compare with `glob('src/*/models.py')`, which returns almost immediately.)
### Expected behaviour
No unnecessary delay.
| bug-vim,completion | low | Major |
543,270,808 | flutter | Modify default Gradle file or "Build and release an Android app" document | I have an open source project (https://github.com/bernaferrari/color-studio) and I was setting up the signing keys to release in Play Store. The issue is, the [documentation](https://flutter.dev/docs/deployment/android) only covers how to add your own key (replacing `signingConfigs.debug` with `signingConfigs.release` and then setting it up). It doesn't cover how to keep both: the release keys if available, or the debug otherwise. I want people to download my project and call `flutter run --release` without trouble.
So, I got it by making the following modifications:
```groovy
def propertyExists = keystorePropertiesFile.exists() // new attribute added
if (propertyExists) {
keystoreProperties.load(new FileInputStream(keystorePropertiesFile))
}
signingConfigs {
if (propertyExists) { // check if file exists
release {
keyAlias keystoreProperties['keyAlias']
keyPassword keystoreProperties['keyPassword']
storeFile file(keystoreProperties['storeFile'])
storePassword keystoreProperties['storePassword']
}
}
}
buildTypes {
release { // provide both
signingConfig propertyExists ? signingConfigs.release : signingConfigs.debug
}
}
```
And I was thinking, there are three possibilites:
1. Modify the document with my instructions (5 lines of difference)
2. Detail in the document what should do if you want to have both signing configs available when release key is not present.
3. Just modify the Android sample when Flutter creates a project and include them with a fake key.properties file. That way, anyone releasing an Android app would only have to modify that file instead of creating that file, looking at documentation and modifying gradle.
I'm actually in favor of option 3, but option 1 and 2 are also good. | c: new feature,tool,t: gradle,a: release,P3,team-tool,triaged-tool | low | Critical |
543,272,311 | pytorch | broadcast randperm with dim specification | ## π Feature
<!-- A clear and concise description of the feature proposal -->
Currently randperm uses only int as input
It would be great to provide a tensor and specify a dimension to make a permutation on.
## Motivation
That would be an example of how to do right now
```
idx = torch.stack([torch.randperm(n) for _ in range(28*28)]).t().long() # permute pixel of an image
data_shuffled = torch.zeros(10, 784).scatter_(0, idx, data.view(10, -1)).view(10, 28, 28) # permute each image independently but in a vectorized way.
```
It isn't neither nice nor optimized.
<!-- Please outline the motivation for the proposal. Is your feature request related to a problem? e.g., I'm always frustrated when [...]. If this is related to another GitHub issue, please link here too -->
## Pitch
<!-- A clear and concise description of what you want to happen. -->
## Alternatives
<!-- A clear and concise description of any alternative solutions or features you've considered, if any. -->
## Additional context
<!-- Add any other context or screenshots about the feature request here. -->
| triaged,function request | low | Minor |
543,276,134 | flutter | Add Support for tvOS | _Latest status: https://github.com/flutter/flutter/issues/47928#issuecomment-1366960093_
----
*@lesnitsky commented on Dec 24, 2019, 4:08 PM UTC:*
I'm trying to build `Flutter.framework` for tvOS and getting this error. Is there a way to workaround this issue?
*This issue was moved by [devoncarew](https://github.com/devoncarew) from [dart-lang/sdk#39918](https://github.com/dart-lang/sdk/issues/39918).* | c: new feature,tool,engine,c: proposal,P3,team-engine,triaged-engine | medium | Critical |
543,281,415 | TypeScript | Can't infer unique symbol types via another variable | Literals such as 0 can be inferred correctly but unique symbols are not.
<!--
Please try to reproduce the issue with the latest published version. It may have already been fixed.
For npm: `typescript@next`
This is also the 'Nightly' version in the playground: http://www.typescriptlang.org/play/?ts=Nightly
-->
**TypeScript Version:** 3.7.x-dev.20191224
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:**
**Code**
```ts
let revision: symbol;
const rev = revision = Symbol();
rev === revision;
```
**Expected behavior:**
Type rev is unique symbol.
**Actual behavior:**
Type rev is symbol. Can't even declare that rev is a unique symbol.
**Playground Link:** https://www.typescriptlang.org/play/index.html?ts=3.8.0-dev.20191224&ssl=1&ssc=1&pln=3&pc=18#code/DYUwLgBATiBuCWBneB7AdgLgogngWwCMVgBuAKAGN1FIZYIBeaOJVNRiAZXyOAAoAlOTqMGTOq3QkgA
**Related Issues:** <!-- Did you find other bugs that looked similar? -->
| Bug | low | Critical |
543,296,866 | rust | Remove `HirId` from MIR | The [`source_scopes`](https://doc.rust-lang.org/nightly/nightly-rustc/rustc/mir/struct.Body.html#structfield.source_scopes) field of `mir::Body` contains a `lint_root` `HirId` for figuring out the correct lint levels when reporting lints in MIR processing. Not only does this break our abstractions between HIR and MIR, this makes inlining functions from other crates super weird, because right now we just outright stop reporting any lints on the code from the other crates, no matter who wrote any kind of lint level.
We should instead figure out a system that stores all the lint level information directly in the MIR. To prevent duplication between mir bodies, we should consider some interning scheme (e.g. a query taking a `DefId` and returning an interned lint level information struct that may have a `parent` field pointing to the parent lint level datastructure).
cc @spastorino @wesleywiser @eddyb | C-cleanup,A-lints,T-compiler,A-HIR,A-MIR | low | Minor |
543,302,683 | go | archive/zip: FileHeader.SetMode: character files encoded as regular files | ```
package main
import (
"archive/zip"
"log"
"os"
)
const charDeviceFile = "/dev/null"
func main() {
fi, err := os.Stat(charDeviceFile)
if err != nil {
panic(err)
}
fh, err := zip.FileInfoHeader(fi)
if err != nil {
panic(err)
}
if isChar(fi) != isChar(fh.FileInfo()) {
log.Printf("file mode inconsistent after zip.FileInfoHeader: got %q, want %q", fh.FileInfo().Mode(), fi.Mode())
}
}
func isChar(fi os.FileInfo) bool { return fi.Mode() & os.ModeCharDevice != 0 }
```
This reproducer program [fails unexpectedly](https://play.golang.org/p/aS7W0qzdRPH) with the error:
file mode inconsistent after zip.FileInfoHeader: got "-rw-rw-rw-", want "Dcrw-rw-rw-"
The data loss is due to a logically impossible bit match in `fileModeToUnixMode`. The fix:
```
--- src/archive/zip/struct.go 2019-12-29 02:26:54.000000000 +0100
+++ src/archive/zip/struct.go 2019-12-29 02:27:05.000000000 +0100
@@ -337,12 +337,10 @@
m = s_IFIFO
case os.ModeSocket:
m = s_IFSOCK
+ case os.ModeDevice | os.ModeCharDevice:
+ m = s_IFCHR
case os.ModeDevice:
- if mode&os.ModeCharDevice != 0 {
- m = s_IFCHR
- } else {
- m = s_IFBLK
- }
+ m = s_IFBLK
}
if mode&os.ModeSetuid != 0 {
m |= s_ISUID
``` | NeedsInvestigation | low | Critical |
543,309,796 | rust | Remove uses of `_` in defaults | Quoting myself from the referenced thread:
I feel like the internal use of `_` as a default name for things is a bit undisciplined, and it's leaking into diagnostics in annoying ways. There is also this: https://github.com/rust-lang/rust/pull/66679
I feel like internally, we should keep a strong separation between names that we get from the source code being compiled and names being generated by the compiler...
_Originally posted by @mark-i-m in https://github.com/rust-lang/rust/issues/67565#issuecomment-569458176_
cc @Centril @matthewjasper @pnkfelix | C-cleanup,C-enhancement,A-diagnostics,T-compiler | low | Minor |
543,311,457 | opencv | Invalid parameters description in Java documentation for findContours | https://docs.opencv.org/master/javadoc/org/opencv/imgproc/Imgproc.html#findContours(org.opencv.core.Mat,java.util.List,org.opencv.core.Mat,int,int)
> hierarchy - Optional output vector (e.g. std::vector<cv::Vec4i>), containing information about the image topology. It has as many elements as the number of contours. For each i-th contour contours[i], the elements hierarchy[i][0] , hierarchy[i][1] , hierarchy[i][2] , and hierarchy[i][3] are set to 0-based indices in contours of the next and previous contours at the same hierarchical level, the first child contour and the parent contour, respectively. If for the contour i there are no next, previous, parent, or nested contours, the corresponding elements of hierarchy[i] will be negative.
This is not the case. The Java hierarchy output is a matrix with 1 row, and N columns (for N contours), where each element contains a 4-element double array. The double array holds the indices of the next, previous, child, and parent contours.
Eg: hierarchy.get(0, k) gives a double[4]
hierarchy.get(1, k) throws "java.lang.NullPointerException: Attempt to read from null array" | category: documentation,category: java bindings,Hackathon | low | Minor |
543,314,188 | pytorch | What is the significance of torchvision._is_tracing()? | ## What is the significance of torchvision._is_tracing()? β
### 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/)
cc @fmassa | triaged,module: vision | low | Minor |
543,338,054 | TypeScript | Exact generic constraints | <!-- π¨ STOP π¨ π¦π§π’π£ π¨ πΊπ»πΆπ· π¨
Half of all issues filed here are duplicates, answered in the FAQ, or not appropriate for the bug tracker.
Please help us by doing the following steps before logging an issue:
* Search: https://github.com/Microsoft/TypeScript/search?type=Issues
* Read the FAQ, especially the "Common Feature Requests" section: https://github.com/Microsoft/TypeScript/wiki/FAQ
-->
## Search Terms
<!-- List of keywords you searched for before creating this issue. Write them down here so that others can find this suggestion more easily -->
- Generics
- Constraints
- HOCs
- Injected types
## Suggestion
<!-- A summary of what you'd like to see added or changed -->
I would like to write a function that wraps a child function and provides a subset of its required arguments.
Here's an example: https://www.typescriptlang.org/play/index.html#code/PTAEGMAsEsBsBMAUBKUAnApgRwK7UwM6jwCGALiaNAHbiw7w0DmoABia6CdfGwEasAdACgAZjlploAe2oQYCRKQoAuUAG8uagmTTMANKD7bdzUAF9kagG7TovdcNDOIsgtNgZBsaUyXkSQRJDZUC+ZGFzAG5hYRBQSGlwFFACHAAHdNhoDCJ2VkNuXkwyHDRqIkpxSRk5MkhyUFlYAE90bDxCfiFhMhb0jFAAQXhGKVlQAF4NLVTTahZosQlwcblE8AAeABVQDAAPMgweIhGx2oA+RCg4eDV-VVBt1EmL0Ft7K1AHkjUAeQAttAyDtDABrDAtaSiYajYGXF5vD4OJwuaqrWqgADuaBImQwSFC-yBIO24Mh0Nh51kF1QjhcDPkt0QmkEbNChTUAHI8VkMFyLMgYgzzLEGSUynIcbyCTFRcJwG4yNjcfjeNMNtcFPAhcJpWqWUZuXxpC0BZYokA.
However, this errors out because the child function's generic argument `H` could be defined to be _more_ specific than just `a: string` (for example, it could be defined to be `a: 'foo' | 'bar'`). This is covered here: https://github.com/Microsoft/TypeScript/issues/29049, and explained here: https://stackoverflow.com/a/56701587.
It would be useful to have a way to set a constraint on the generic `H` where it _must_ contain exactly the fields that it is "extend"ing from. For example:
```ts
// Note that `contains` here is a made up new term. We can choose another name if preferable.
const hoc = <T contains {a: string}>(child: (arg: T) => void): (arg: Omit<T, 'a'>) => void => {
const wrapper = (arg: Omit<T, 'a'>) => {
child({ ...arg, a: 'apple' });
};
return wrapped;
};
hoc((arg: {}) => {}); // Error, arg does not contain {a: string}
hoc((arg: {a: string}) => {}); // Works!
hoc((arg: {a: string, b: string}) => {}); // Works!
hoc((arg: {a: 'foo' | 'bar'}) => {}); // Error, a must be `string`, not `'foo' | 'bar'`.
```
I don't fully understand why, but a similar example in Flow works without issue (possibly due to the use of exact types?): https://flow.org/try/#0PQKgBAAgZgNg9gdzCYAoVxhgMYAsCWMAJgBQCUYATgKYCOArvjQM5hECGALu2PgHbYY9IvwDmYAAbsJYdnyKSARhIB0qKPQGd8cPjgLESHbgC4wAbwA+ss806UxAGjCLb9sWEsBfMmYBucPgK5qhgYTi6zHAw1CrwokZc7CrszsbJimSoXgDc6JhguHDY5GDM9AAOFTD41KxSEs5yCjSc9JR8rDwaWjp6nLhcYLowAJ5UdIwsSqrqmtjauoXFADwAKmbmNmXufOJeAHwkeIREZommYGsUALwHYAFBvmAX7GYAJAAi+FBQ685WbZ2Bx7TyHW73R7BULhHoLPpgBCUdhVaikdIfb6-f4WaxvHYg8TeA4UELhcn6U4kLYqWnpJpmADkKOq1EZYB8eXJXnQ5Na7T0SJZaLyPNQ2EinERyNRChuyxKJ2IZDyQtl1NcYEZijgo0ZnKAA.
## Use Cases
<!--
What do you want to use this for?
What shortcomings exist with current approaches?
-->
My motivating use case is building a well-typed middleware framework. Middleware often add fields to the `context` of an HTTP request before or after the handler is called. I would like for a Middleware function to exactly specify which fields it adds to the handler. Applied to the playground code above, `child` is the HTTP handler, and `hoc` is one such middleware function.
## Examples
<!-- Show how this would be used and what the behavior would be -->
As above.
## 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).
- In particular, this helps make TypeScript more composable, as per "Produce a language that is composable and easy to reason about.".
| Discussion | medium | Critical |
543,344,045 | pytorch | Parallelization: more balanced work distribution among workers | ## π Feature
Parallelization: more balanced work distribution among workers
## Motivation
I recently checked the code for `at::parallel_for` method and this is what I observed.
Suppose there are `N` indices and `T` workers, then worker `i (0 <= i < T)` receives
indices `[i * ceil(N/T), min(N, (i+1) * ceil(N/T)) )`.
Now, suppose `N=10, T=4`, then this scheme will have the following distribution : `3|3|3|1`.
Definitely, something like `3|3|2|2` looks better. The issues is that the current approach is biased against the last worker(s) - it (them) always receive(s) the least work, which **could be even zero**. To elaborate, the whole situation becomes even worse if, for example, `N=100, T=40`. Then work of size 3 is scheduled for each worker, meaning that only 34 workers are going to do something useful, while 6 workers do nothing. And, in general, the situation gets worse if the number of workers scales up with the size of an input.
## Pitch
What about a slightly different distribution? This one **uses all the workers**!
For a worker `i: 0 <= i < T` let
```
begin(i) = ceil(N*i/T),
end(i) = begin(i+1),
```
and the worker receives indices `[begin(i), end(i) )`
It can be shown that for any `N >= T >= 1`, any `i,j: 0 <= i,j < T`:
```
|(end(i) - begin(i)) - (end(j) - begin(j))| <= 1,
end(i) - begin(i) >= 1, begin(i) < N,
```
so this new scheme is optimally balanced, and each worker will do at least something!
The only issue I see is a more likely chance of overflow in computing `begin`
cc @yf225 @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @rohan-varma @gqchen @aazzolini @xush6528 | module: cpp,feature,triaged,module: data parallel | low | Minor |
543,361,874 | pytorch | DataLoader does not consider default floating point type | ## π Bug
The default collate function used in `torch.utils.data.DataLoader` ignores the default floating point type set via `torch.set_default_dtype()`. The created type for a python `float` is always `torch.float64`. However, I am not sure if this is a deliberate design decision.
## To Reproduce
Steps to reproduce the behavior:
1. Set the dtype to float32 with torch.set_default_dtype(torch.float32)
2. Use DataLoader on a dataset which has python scalar floats
```
>>> torch.set_default_dtype(torch.float32)
>>> torch.utils.data._utils.collate.default_collate((1.0,)).dtype
torch.float64
```
<!-- If you have a code sample, error messages, stack traces, please provide it here as well -->
## Expected behavior
The default collate function should create for python `float` types the dtype that is set with `torch.set_default_dtype()`.
## Environment
```
PyTorch version: 1.3.1
Is debug build: No
CUDA used to build PyTorch: 10.0.130
OS: Debian GNU/Linux bullseye/sid
GCC version: (Debian 8.3.0-26) 8.3.0
CMake version: Could not collect
Python version: 3.7
Is CUDA available: Yes
CUDA runtime version: 10.1.168
GPU models and configuration: GPU 0: GeForce RTX 2070 SUPER
Nvidia driver version: 430.64
cuDNN version: Could not collect
Versions of relevant libraries:
[pip3] numpy==1.16.5
[conda] _pytorch_select 0.2 gpu_0
[conda] blas 1.0 mkl
[conda] mkl 2019.4 243
[conda] mkl-service 2.3.0 py37he904b0f_0
[conda] mkl_fft 1.0.15 py37ha843d7b_0
[conda] mkl_random 1.1.0 py37hd6b4f25_0
[conda] pytorch 1.3.1 cuda100py37h53c1284_0
[conda] torchvision 0.4.2 cuda100py37hecfc37a_0
```
cc @SsnL | module: dataloader,triaged | low | Critical |
543,364,314 | TypeScript | [Compiler API] function which check if a class implements an interface | <!-- π¨ STOP π¨ π¦π§π’π£ π¨ πΊπ»πΆπ· π¨
Half of all issues filed here are duplicates, answered in the FAQ, or not appropriate for the bug tracker.
Please help us by doing the following steps before logging an issue:
* Search: https://github.com/Microsoft/TypeScript/search?type=Issues
* Read the FAQ, especially the "Common Feature Requests" section: https://github.com/Microsoft/TypeScript/wiki/FAQ
-->
## Search Terms
<!-- List of keywords you searched for before creating this issue. Write them down here so that others can find this suggestion more easily -->
Compiler API, InterfaceDeclaration, ClassDeclaration, Checker
## Suggestion
<!-- A summary of what you'd like to see added or changed -->
A function to check whether a `ClassDeclaration` implements an `InterfaceDeclaration`
## Use Cases
Hello, I wanted to check whether a `ClassDeclaration` of `a.ts` file implements an `InterfaceDeclaration` from `b.ts` file using Compiler API.
But I couldn't find a method or a function for it.
## Examples
```ts
function isClassImplementInterface(
ts.ClassDeclaration: classDeclaration,
ts.InterfaceDeclaration: interfaceDeclaration
): boolean {
// return true if classDeclaration implements interfaceDeclaration correctly
}
```
## 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,API,Awaiting More Feedback | low | Critical |
543,365,752 | godot | Light2D is unreliable when reparenting it | **Godot version:**
3.2 beta4
**Issue description:**
Soo I have a rather complicated light setup to workaround #26507 and recently (?) I noticed that it's unreliable. What I do is keep whole game in a ViewportContainer and reparent any Light2D that appears to the parent of this container and then draw the whole screen masked with lights. Yep, that's complicated.
My problem is that lights sometimes don't work. Completely randomly. Like, check this screenshot. The character has light around it.
But when I run the game multiple times, 70% of the time the light does not appear.
This happens after reparenting. I tried debugging it to see what's the difference when it works and when doesn't and there doesn't seem to be any. The light has same parent/position in parent in both cases, all properties have the same values in remote inspector. Disabling/enabling light or toggling visibility doesn't fix it. Here's a GIF of the bug in action.
So uh, I have no idea how to reproduce it outside my game, but it's clearly some problem with the engine.
**Minimal reproduction project:**
I haven't managed to create a minimal project :/ Simply reparenting doesn't seem to cause the bug, it's just a trigger. | bug,topic:rendering,confirmed,topic:2d | low | Critical |
543,400,227 | TypeScript | Give better error when using private identifier in parameter property | <!-- π¨ STOP π¨ π¦π§π’π£ π¨ πΊπ»πΆπ· π¨
Half of all issues filed here are duplicates, answered in the FAQ, or not appropriate for the bug tracker. Even if you think you've found a *bug*, please read the FAQ first, especially the Common "Bugs" That Aren't Bugs section!
Please help us by doing the following steps before logging an issue:
* Search: https://github.com/Microsoft/TypeScript/search?type=Issues
* Read the FAQ: https://github.com/Microsoft/TypeScript/wiki/FAQ
Please fill in the *entire* template below.
-->
<!--
Please try to reproduce the issue with the latest published version. It may have already been fixed.
For npm: `typescript@next`
This is also the 'Nightly' version in the playground: http://www.typescriptlang.org/play/?ts=Nightly
-->
**TypeScript Version:** 3.8.0-dev.20191228
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:** parameter property private
**Code**
```ts
class C {
constructor(private #prop: number) {}
}
```
**Expected behavior:**
Maybe `Parameter properties cannot have a private name.`
**Actual behavior:**
```
A parameter property is only allowed in a constructor implementation.(2369)
Parameter '(Missing)' implicitly has an 'any' type.(7006)
Identifier expected.(1003)
';' expected.(1005)
Unexpected token. A constructor, method, accessor, or property was expected.(1068)
```
**Playground Link:** https://www.typescriptlang.org/play/?ts=3.8.0-dev.20191228&ssl=3&ssc=2&pln=1&pc=1#code/MYGwhgzhAEDC0G8BQ1XWAewHYQC4CcBXYXDfACgAd8BLANzFwFNoBiajSgLmi0IFsARk3wBKRAF8kEoA
**Related Issues:** <!-- Did you find other bugs that looked similar? -->
| Bug | low | Critical |
543,428,166 | vscode | Pseudoterminal handleInput does not catch Terminal.sendText | Info:
- VSCode Version: 1.41.1
- OS Version: Windows 10
Issue:
This code creates a pseudoterminal and tries to write a string to it:
```
const writeEmitter = new vscode.EventEmitter<string>();
const pty: vscode.Pseudoterminal = {
onDidWrite: writeEmitter.event,
open: () => writeEmitter.fire('Opening terminal!'),
close: () => {},
handleInput: data => writeEmitter.fire(data === '\r' ? '\r\n' : data)
};
let terminal = vscode.window.createTerminal({ name: "test", pty });
terminal.sendText("test");
```
However the handleInput does not catch the sendText command and no text is printed.
The docs do mention that it catches sendText():
>handleInput(data: string): void
>Implement to handle incoming keystrokes in the terminal or when an extension calls Terminal.sendText. data contains the keystrokes/text serialized into their corresponding VT sequence representation.
Typing in the pseudoterminal does trigger handleInput. | bug,api,confirmed,terminal-input | low | Major |
543,467,880 | vscode | Improve task debugging by giving more context | <!-- 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. -->
Most of this text comes from an email. The context is that I got on a phone call with a customer who was experiencing issues debugging an Azure Function App.
#### The issue
F5 debugging of a Function App would show:
```
> Executing task: func host start <
```
And then something saying that the last exit code was 1.
It wasnβt clear what the problem was, why it was failing, and we found that this would fail on `func extensions install` and `func host install`β¦ running the func commands in the PowerShell Integrated Console or outside of VS Code in any terminal worked just fineβ¦
#### The resolution
Then we realized that the customer's default shellβ¦β¦.. was WSL.
As you know, tasks use the default shell to run in by default. Which means that the `func` commands were being run in WSL which didnβt have the Core Tools installed in it so the Task threw an exit code of 1 because the command wasnβt found.
Once Aleksandar changed his default shell back to PowerShell, everything started working again.
#### The problem
The lack of warning on tasks as to what went wrong is very hard to debug⦠see above. It took an hour and a half to make this realization... And even more of the customer's time to get to the point where they asked to get on a call...
#### The ask
Show the shell being used to run tasks or give more than just an exit code if possible. We need more context as what's actually going on when a task is being run.
I think this is a generic problem with any extension that relies on tasks and not just the Azure Functions extension. | feature-request,tasks | medium | Critical |
543,475,634 | flutter | Material Colors stored in a Map are not const and cannot be used as constructor defaults | ## Use Case
**Reproduction**
Code up a ```Widget``` which uses a default value for a ```Color```. For example
```Dart
class DisplayArray extends StatelessWidget {
final Color onColor;
DisplayArray({
this.onColor = Colors.red.shade900,
// and so on
```
**Expected**
The code complies because the ```Color``` constructor is a ```const``` constructor. So, ```Colors.red.shade900``` should likewise be ```const```.
**Actual**
The code does not compile. The color ```Colors.red.shade900``` is stored in a ```Map```. The complier seems to be unable to work out that ```Colors.red.shade900``` is ```const```.
**Workaround**
I found that I had to redefine the ```Color``` from the ```Colors``` class in my own project as as
```Dart
const Color RED_SHADE_900 = const Color(0xFFB71C1C);
```
This approach seems a bit messy to me. Code is duplicated simply for the sake of making Colors.red.shade900 appear as ```const```.
## Proposal
Define the material colors as ```const```.
**Futter Doctor**
```
β ~ flutter doctor
Doctor summary (to see all details, run flutter doctor -v):
[β] Flutter (Channel stable, v1.12.13+hotfix.5, on Mac OS X 10.15.2 19C57,
locale en-CA)
[β] Android toolchain - develop for Android devices (Android SDK version 28.0.3)
[β] Xcode - develop for iOS and macOS (Xcode 11.3)
[β] Android Studio (version 3.5)
[β] IntelliJ IDEA Community Edition (version 2019.3)
[!] VS Code (version 1.32.3)
β Flutter extension not installed; install from
https://marketplace.visualstudio.com/items?itemName=Dart-Code.flutter
[β] Connected device (1 available)
! Doctor found issues in 1 category.
```
| framework,f: material design,c: proposal,team-design,triaged-design | low | Minor |
543,509,088 | angular | Property getter/setter behavor difference when using @angular/elements | # π bug report
### Affected Package
The issue is related to the package @angular/elements
### Is this a regression?
No
### Description
In the demo below you will see a (common?) pattern of initializing data in an angular component. There are many ways to do this but essentially it is common (for me at least) to initialize data in a property setter. This works fine in angular components, however, when the component is used in a custom element via angular/elements (as you will see) the property setter is not called if the attribute value is the same as the value returned from the getter.
Putting it another way... In angular components, during initialization, the property value is set regardless of the current property value. However, in angular elements, the value is first retrieved and only set if the new value is different from the current value.
Obviously, there are many ways to refactor the code to avoid this issue but is surprizing at a minimum and very frustrating if you don't know what's going on (my experience).
## π¬ Minimal Reproduction
https://stackblitz.com/edit/angular-elements-demo-fnulni
## π₯ Exception or Error
No error. This can result in unexpected behavior for angular element users that expect property setters to be called during initialization.
## π Your Environment
**Angular Version:**
Tested in both v8 and v9-rc.
| type: bug/fix,area: elements,state: confirmed,P3 | low | Critical |
543,533,654 | flutter | Show back button instead of the menu button when both are used. | <!-- 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
On iOS there is no way to go back when you both have a route below and a drawer, as the menu button overwrites the back button.
<!--
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
Show the back button instead of the menu button in the app bar when both a drawer and a route below is used. The drawer can always be opened with a swipe gesture, but you cannot always go back on iOS as there are no hardware/software back buttons.
<!--
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.
-->
| framework,f: cupertino,c: proposal,team-design,triaged-design | low | Critical |
543,582,988 | TypeScript | Useless scary elaboration on all overloads in JSX |
```
git clone https://github.com/React95/React95/
git checkout 57dc66d8a087792fd8fa0f1cfb7c7560767e8520
yarn
```
1. Open up VS Code in the base
1. Open up the file packages/core/components/Checkbox.jsx
1. Go to line 100
1. Make a change
```
- <Label {...style} checked={checked} disabled={disabled}>
+ <Label style={style} checked={checked} disabled={disabled}>
```
Observe the error message:
```
No overload matches this call.
Overload 1 of 2, '(props: Pick<Pick<Pick<DetailedHTMLProps<LabelHTMLAttributes<HTMLLabelElement>, HTMLLabelElement>, "form" | ... 255 more ... | "htmlFor"> & { ...; }, "form" | ... 256 more ... | "htmlFor"> & Partial<...>, "form" | ... 256 more ... | "htmlFor"> & { ...; } & { ...; }): ReactElement<...>', gave the following error.
Type '{ children: Element[]; style: any; checked: any; disabled: any; }' is not assignable to type 'IntrinsicAttributes & Pick<Pick<Pick<DetailedHTMLProps<LabelHTMLAttributes<HTMLLabelElement>, HTMLLabelElement>, "form" | ... 255 more ... | "htmlFor"> & { ...; }, "form" | ... 256 more ... | "htmlFor"> & Partial<...>, "form" | ... 256 more ... | "htmlFor"> & { ...; } & { ...; }'.
Property 'checked' does not exist on type 'IntrinsicAttributes & Pick<Pick<Pick<DetailedHTMLProps<LabelHTMLAttributes<HTMLLabelElement>, HTMLLabelElement>, "form" | ... 255 more ... | "htmlFor"> & { ...; }, "form" | ... 256 more ... | "htmlFor"> & Partial<...>, "form" | ... 256 more ... | "htmlFor"> & { ...; } & { ...; }'.
Overload 2 of 2, '(props: StyledComponentPropsWithAs<"label", any, {}, never>): ReactElement<StyledComponentPropsWithAs<"label", any, {}, never>, string | ... 1 more ... | (new (props: any) => Component<...>)>', gave the following error.
Type '{ children: Element[]; style: any; checked: any; disabled: any; }' is not assignable to type 'IntrinsicAttributes & Pick<Pick<Pick<DetailedHTMLProps<LabelHTMLAttributes<HTMLLabelElement>, HTMLLabelElement>, "form" | ... 255 more ... | "htmlFor"> & { ...; }, "form" | ... 256 more ... | "htmlFor"> & Partial<...>, "form" | ... 256 more ... | "htmlFor"> & { ...; } & { ...; }'.
Property 'checked' does not exist on type 'IntrinsicAttributes & Pick<Pick<Pick<DetailedHTMLProps<LabelHTMLAttributes<HTMLLabelElement>, HTMLLabelElement>, "form" | ... 255 more ... | "htmlFor"> & { ...; }, "form" | ... 256 more ... | "htmlFor"> & Partial<...>, "form" | ... 256 more ... | "htmlFor"> & { ...; } & { ...; }'.ts(2769)
```
| Suggestion,Experience Enhancement | low | Critical |
543,586,701 | flutter | [in_app_purchase] Support for test billing on Android? | I'm trying to submit a test purchase with id of 'android.test.purchased' but receiving an error:
_Unhandled Exception: PlatformException(NOT_FOUND, Details for sku android.test.purchased are not available. Has this ID already been fetched?, null)_
Is this use case supported?
https://developer.android.com/google/play/billing/billing_testing
[β] Flutter (Channel stable, v1.12.13+hotfix.5, on Microsoft Windows [Version 10.0.18362.535], locale en-US)
β’ Flutter version 1.12.13+hotfix.5 at D:\_Development\_mobile\flutter
β’ Framework revision 27321ebbad (3 weeks ago), 2019-12-10 18:15:01 -0800
β’ Engine revision 2994f7e1e6
β’ Dart version 2.7.0
[β] Android toolchain - develop for Android devices (Android SDK version 28.0.3)
β’ Android SDK at C:\Users\Shawn\AppData\Local\Android\sdk
β’ Android NDK location not configured (optional; useful for native profiling support)
β’ Platform android-29, build-tools 28.0.3
β’ Java binary at: C:\Program Files\Android\Android Studio New\jre\bin\java
β’ Java version OpenJDK Runtime Environment (build 1.8.0_202-release-1483-b03)
β’ All Android licenses accepted.
[β] Android Studio (version 3.5)
β’ Android Studio at C:\Program Files\Android\Android Studio New
β’ Flutter plugin version 40.2.2
β’ Dart plugin version 191.8593
β’ Java version OpenJDK Runtime Environment (build 1.8.0_202-release-1483-b03)
[β] VS Code (version 1.41.1)
β’ VS Code at C:\Users\Shawn\AppData\Local\Programs\Microsoft VS Code
β’ Flutter extension version 3.7.1
[β] Connected device (1 available)
β’ AOSP on IA Emulator β’ emulator-5554 β’ android-x86 β’ Android 9 (API 28) (emulator)
β’ No issues found!
| c: new feature,platform-android,p: in_app_purchase,package,c: proposal,dependency: android,P3,team-android,triaged-android | low | Critical |
543,605,686 | TypeScript | Can't evaluate equality of symbols made by Symbol.for | This is also an obstacle of branded types and ghost types. I think resolving this problem is easy because these types are just a new literal type similar to string literal types.
<!--
Please try to reproduce the issue with the latest published version. It may have already been fixed.
For npm: `typescript@next`
This is also the 'Nightly' version in the playground: http://www.typescriptlang.org/play/?ts=Nightly
-->
**TypeScript Version:** 3.7.x-dev.20191228
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:**
**Code**
```ts
const a = Symbol.for('');
const b = Symbol.for('');
const c: typeof b = Symbol.for(''); // error
a === b; // error
```
**Expected behavior:**
no error
**Actual behavior:**
Type 'typeof b' is not assignable to type 'typeof a'.
This condition will always return 'false' since the types 'typeof a' and 'typeof b' have no overlap.
**Playground Link:** https://www.typescriptlang.org/play/index.html?ts=3.8.0-dev.20191228&ssl=1&ssc=1&pln=4&pc=18#code/MYewdgzgLgBAhjAvDAygTwLYCMQBsB0AZiAE4AUA5BQJQDcAUKJLFkqpjgceVXY+NBjAAXDChoADgFMQhGK2TpseIqUo1aMAPRaYUkiVL0EiU-M069B0kA
**Related Issues:** #35200
| Suggestion,Awaiting More Feedback | medium | Critical |
543,633,032 | angular | When using the Unified Angular Location Service, AngularJS useHash paths have their leading slashes removed | # π bug report
### Affected Package
The issue is caused by package @angular/common/upgrade
### Is this a regression?
I don't think so. The Unified Location Service was introduced in 8.x and I think it's always behaved this way.
### Description
When:
- The Unified Location Service is being used
- The `{ useHash: true }` mode is active
- AngularJS code sets the path via `$location.path()`
- The path has a leading slash, such as `/mypath'
I expect the URL to be:
```
http://mydomain.com/index.html#/mypath
```
But instead, the leading slash is removed before it's applied to the location's hash:
```
http://mydomain.com/index.html#mypath
```
## π¬ Minimal Reproduction
[This stackblitz](https://stackblitz.com/edit/angular-hybrid-uniified-location?file=angularjs.module.ts) shows a matrix of scenarios:
||useHash: true|useHash: false|
|-|-|-|
|`$location.path('/slash')` (AngularJS $locationShim)|β|β
|
|`$locationpath('noslash')` (AngularJS $locationShim)|β
|β
|
|`location.go('/slash')` (Angular Location)|β
|β
|
|`location.go('noslash') (Angular Location)|β
|β
|
## π Your Environment
**Angular Version: 8**
with @angular/upgrade
**Anything else relevant?**
This was reported to me by a UI-Router Hybrid user trying to enable the Unified Angular Location Service. https://github.com/ui-router/angular-hybrid/issues/320
A simple workaround is to continue to use the AngularJS `$location` service for URL management and upgrade it via ngUpgrade to be used in Angular services. | type: bug/fix,area: common,freq1: low,area: upgrade,state: confirmed,P3 | low | Critical |
543,633,696 | TypeScript | Typescript: autocomplete won't work when "Parameters<Fn>" is used and "Fn" is passed through a generic | Issue Type: <b>Bug</b>
Hello everybody, I wasn't entirely sure of where to post this: vscode or typescript? Since Typescript correctly shows the type and the problem lies in the autocompletion itself, I'm thinking there's a higher chance for this problem to be related to vscode.
The problem is: the autocomplete won't work even though Typescript clearly knows the type. This happens when I use the helper "Parameters" from typescript applied to a function whose type has been passed through a generic. Let me show an example:
```typescript
interface Test {
myFunction: (params: { a: number, b: number }) => void
}
type Wrapped<Fn extends (params: any) => any> = {
myFunction: (params: Parameters<Fn>[0]) => ReturnType<Fn>,
}
function wrap<T extends Test>(test: T) {
let result: Wrapped<T['myFunction']>
return result
}
let t: Test
let w = wrap(t)
t.myFunction({})
w.myFunction({})
```
The autocompletion works as expected in `t.myFunction({})`, but it won't work at all in `w.myFunction({})`. The following image exemplifies the behavior:
Typescript correctly assigns the type for both functions, see:
Is this behavior expected? If yes, why two equal types show different behaviors when autocompleting?
I can easily fix this by instead of passing the function type as a generic, I pass the object type, see:
```typescript
interface Test {
myFunction: (params: { a: number, b: number }) => void
}
type Wrapped<T extends Test> = {
myFunction: (params: Parameters<T['myFunction']>[0]) => ReturnType<T['myFunction']>,
}
function wrap<T extends Test>(test: T) {
let result: Wrapped<T>
return result
}
let t: Test
let w = wrap(t)
t.myFunction({})
w.myFunction({})
```
The autocompletion works fine for both `t.myFunction` and `w.myFunction` now. But why doesn't it work when the function type itself is passed to the generic?
I appreciate any help. Thanks.
--------------
EDIT: it seems the problem occurs when using `extends (params: any) => any`. In the second example, where both calls had the autocompletion working, if I change `type Wrapped<T extends Test>` to `type Wrapped<T extends { myFunction: (params: any) => any }>` it goes back to the behavior of my first example. Again, the type gets correctly assigned by Typescript, the autocompletion is the only thing that won't work. If I use `extends any` instead of `extends (params: any) => any` to type the function, it will work in all cases. Pretty weird right?
VS Code version: Code 1.41.0 (9579eda04fdb3a9bba2750f15193e5fafe16b959, 2019-12-11T18:37:42.077Z)
OS version: Windows_NT x64 10.0.18362
<details>
<summary>System Info</summary>
|Item|Value|
|---|---|
|CPUs|Intel(R) Core(TM) i5-8600K CPU @ 3.60GHz (6 x 3600)|
|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<br>surface_control: disabled_off<br>surface_synchronization: enabled_on<br>video_decode: enabled<br>viz_display_compositor: enabled_on<br>viz_hit_test_surface_layer: disabled_off<br>webgl: enabled<br>webgl2: enabled|
|Load (avg)|undefined|
|Memory (System)|15.94GB (10.81GB free)|
|Process Argv||
|Screen Reader|no|
|VM|0%|
</details><details><summary>Extensions (7)</summary>
Extension|Author (truncated)|Version
---|---|---
color-info|bie|0.5.1
vscode-eslint|dba|2.0.11
auto-close-tag|for|0.5.6
vscode-styled-components|jpo|0.0.27
remote-wsl|ms-|0.41.6
markdown-preview-enhanced|shd|0.5.0
code-spell-checker|str|1.7.20
</details>
<!-- generated by issue reporter --> | Bug | low | Critical |
543,662,682 | TypeScript | Support Type Acquisition From Npx Shebang | Originally filed under https://github.com/microsoft/vscode/issues/87834.
## Search Terms
shebang, npx, type acquisition
## Suggestion
When authoring JS scripts that are to be invoked using `npx`, it would be great if the typing information could be inferred from the package that `npx` will call.
## Use Cases
I'm currently building a library, where primary use case is to be consumed as executable scripts, with `npx` as the shebang command (this allows for global execution without the need for prior installation). And since all of the needed functionality is automatically injected into the global scope there is no need for external references. This is intended for portability and ease of authoring.
When authoring these scripts, the desire is to leverage the typings provided by the library called by `npx`. The assumption here is that the toolchain would detect the `npx` invocation in the shebang, and load the appropriate typings (local if library is previously installed, or downloaded if not).
## Examples
```javascript
#!/usr/bin/npx @arcsine/nodesh
stdin
.sort()
.unique()
.stdout
```
In the sample above, `stdin` is a global variable that should be picked up via intellisense, and the methods it exposes should be searchable (with the associated docs).
## Considerations
Alternatively, one thought was to leverage a triple-slash directive, but that is limited as it cannot:
* point to global node modules (without an absolute path)
* reference ephemeral libraries as loaded by npx
## Checklist
My suggestion meets these guidelines:
* [ ] 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 |
543,663,454 | rust | Support TCP_FASTOPEN in TcpStream and TcpListener | ## Background
`TcpStream` and `TcpListener` in `libstd` doesn't support [TCP Fast Open (TFO)](https://en.wikipedia.org/wiki/TCP_Fast_Open). Because (in Linux)
1. `TcpListener` must call `setsockopt` with `TCP_FASTOPEN` after `listen(..)`
2. `TcpStream`
* Kernel >= 4.11, call `setsockopt` with `TCP_FASTOPEN_CONNECT` before `connect`
* Otherwise, call `sendto` with flag `MSG_FASTOPEN`
Since libraries out there (`mio`, `tokio`, `async_std`, ...) use `net::TcpStream` and `net::TcpListener` inside there implementation, supporting TFO in `libstd` would be the best solution for all.
## Cross Platform
1. Linux (as mentioned above)
2. OS X
- `TcpListener` calls `setsockopt` with `TCP_FASTOPEN` after `listen(...)`
- For TCP clients, uses `connectx` instead of `connect`
- `associd`: `SAE_ASSOCID_ANY`
- `flags`: `CONNECT_RESUME_ON_READ_WRITE | CONNECT_DATA_IDEMPOTENT`
3. Windows
- `TcpListener` calls `setsockopt` with `TCP_FASTOPEN` after `listen(...)`
- For TCP clients, uses `ConnectEx`
* Call `setsockopt` with `SO_UPDATE_CONNECT_CONTEXT`, otherwise `getpeername` won't work
> When the ConnectEx function returns, the socket s is in the default state for a connected socket.
> The socket s does not enable previously set properties or options until SO_UPDATE_CONNECT_CONTEXT is set. | T-libs-api,C-feature-request | low | Minor |
543,676,164 | godot | AnimationTree Add3's +add and -add act differently | the documentation isn't particularly detailed but I can't see any reason why +add and -add should act any differently for the same animation. I'm also not sure why it appears to act more like blend than add in this instance, shouldn't it move sideways and up/down with both -1 and 1?
tested on 3.1, 3.1.2 and master in linux X11
[test_anim2.zip](https://github.com/godotengine/godot/files/4008799/test_anim2.zip)
| bug,confirmed,topic:animation | low | Minor |
543,720,047 | godot | Exceptions when changing focus or completing a process in iOS (Xcode) | Godot version: 3.1.1
OS/device including version: ANY
If during application launch, unload the application or change focus to another application - it throws an exception
Friends, my hands are falling. I can not find the source of the error.
I no longer know where to look for the cause of the error
A simple 2D game for little children. There is nothing but a third-party βShareβ module.
An exception is thrown when the focus changes or the application is suddenly unloaded while the application is still loading. (Users do not notice this error, this is a background bug)
TEMPLATE COMPILING IOS:
```
platform=iphone target=release optimize=size tools=no use_lto=no deprecated=no game_center=no minizip=no xaudio2=no disable_advanced_gui=yes disable_3d=yes no_editor_splash=yes builtin_bullet=no builtin_certs=no builtin_enet=no builtin_libvpx=no builtin_libwebp=no builtin_libwebsockets=no builtin_mbedtls=no builtin_miniupnpc=no builtin_opus=no builtin_pcre2=no builtin_recast=no builtin_squish=no builtin_thekla_atlas=no builtin_xatlas=no debug_symbols=no separate_debug_symbols=no module_bullet_enabled=no module_csg_enabled=no module_cvtt_enabled=no module_dds_enabled=no module_enet_enabled=no module_gdnative_enabled=no module_gridmap_enabled=no module_hdr_enabled=no module_jpg_enabled=no module_mbedtls_enabled=no module_mobile_vr_enabled=no module_mono_enabled=no module_opensimplex_enabled=no module_opus_enabled=no module_pvr_enabled=no module_recast_enabled=no module_regex_enabled=no module_squish_enabled=no module_svg_enabled=no module_tga_enabled=no module_thekla_unwrap_enabled=no module_theora_enabled=no module_tinyexr_enabled=no module_upnp_enabled=no module_visual_script_enabled=no module_webm_enabled=no module_webp_enabled=no module_websocket_enabled=no module_xatlas_unwrap_enabled=no config=force
```
| bug,platform:ios,topic:porting,crash | low | Critical |
543,729,424 | rust | Allow deny/forbid level lints in macros to cross crate boundaries | I'm attempting to add an `assert_ffi_safe!` macro to `static_assertions` by creating an `extern "C"` block with a function that references the input types and marking it with `#[forbid(improper_ctypes)]` (see https://github.com/nvzqz/static-assertions-rs/pull/31). However, while the macro works correctly if used within the same crate, it results in false positives when used externally. This is evidenced by `compile_fail` doc tests passing.
I imagine that `rustc` lints are disabled when calling into a macro that is not defined within the calling crate so that they don't become a nuisance to consumers who have no control over the code within the macro. This is a case where ensuring that lints work across crate boundaries would be beneficial. | A-lints,A-macros,T-lang,C-feature-request | low | Minor |
543,775,046 | pytorch | the cmake problem with build from source ? | i want to build pytorch from source in conda env , and follow help on https://github.com/pytorch/pytorch#from-source.
export CMAKE_PREFIX_PATH=${CONDA_PREFIX:-"$(dirname $(which conda))/../"}
when i run 'python setup.py build --cmake-only', i got error "cmake returned non-zero exit status 1."
CMakeFiles/CMakeError.log detail is
"Run Build Command(s):/home/wangzhiwei.yt/anaconda3/envs/py37/bin/ninja cmTC_2c68b
[1/2] Building CXX object CMakeFiles/cmTC_2c68b.dir/src.cxx.o
FAILED: CMakeFiles/cmTC_2c68b.dir/src.cxx.o
/usr/bin/c++ -DCAFFE2_IS_NUMA_AVAILABLE -std=c++14 -o CMakeFiles/cmTC_2c68b.dir/src.cxx.o -c src.cxx
src.cxx:1:18: fatal error: numa.h: No such file or directory
#include <numa.h>
^
compilation terminated.
ninja: build stopped: subcommand failed. "
env detail by get from python collect_env.py
Collecting environment information...
PyTorch version: N/A
Is debug build: N/A
CUDA used to build PyTorch: N/A
OS: Debian GNU/Linux 9.8 (stretch)
GCC version: (Debian 6.3.0-18+deb9u1) 6.3.0 20170516
CMake version: version 3.14.0
Python version: 3.7
Is CUDA available: N/A
CUDA runtime version: 10.1.243
GPU models and configuration:
GPU 0: Tesla T4
GPU 1: Tesla T4
GPU 2: Tesla T4
GPU 3: Tesla T4
Nvidia driver version: 418.67
cuDNN version: /usr/local/cuda-10.1/targets/x86_64-linux/lib/libcudnn.so.7
Versions of relevant libraries:
[pip] numpy==1.17.4
[conda] blas 1.0 mkl
[conda] magma-cuda101 2.5.1 1 pytorch
[conda] mkl 2019.4 243
[conda] mkl-include 2019.4 243
[conda] mkl-service 2.3.0 py37he904b0f_0
[conda] mkl_fft 1.0.15 py37ha843d7b_0
[conda] mkl_random 1.1.0 py37hd6b4f25_0
system : Linux n22-190-017 4.9.0-8-amd64 #1 SMP Debian 4.9.144-3 (2019-02-02) x86_64 GNU/Linux
Could anyone help me? Thanks!!
cc @ezyang | module: binaries,triaged | low | Critical |
543,778,998 | go | runtime: pclntab is too big | This article contains results from applying a visualization tool:
https://www.cockroachlabs.com/blog/go-file-size/
If the analysis is correct, pclntab needs a major redesign.
Separated from #6853 to keep the conversation focused.
| NeedsInvestigation,binary-size,compiler/runtime | medium | Critical |
543,787,196 | vue | Event bubbling does not trigger on older iOS devices | ### Version
2.6.11
### Reproduction link
https://github.com/chenjiahan/vue-event-bubble-issue
### Steps to reproduce
This is a browser quirk on iOS <= 11.3
1. Find an old iOS device. (I reproduced the issue on an iOS 8.4 device)
2. Binding click events on parent and child elements
3. Click child element
4. Event listener for parent element is not firing
### What is expected?
Correct event bubbling behavior
### What is actually happening?
Mobile safari has buggy event.timeStamp implementations, the value of event.timestamp is abnormal when the page is first loaded. But event.timestamp will return to normal value after refreshing the page.
We receive a lot of [feedback from vant users](https://github.com/youzan/vant/issues/3015), it is currently observed that the problem may occur on iOS 8.4γ10.2γ10.3γ11.3 (mobile safari and wechat browser)
Screenshot:
---
<!-- generated by vue-issues. DO NOT REMOVE --> | need repro,browser quirks | low | Critical |
543,877,131 | node | Debugger.setInstrumentationBreakpoint is broken | <!--
Thank you for reporting a possible bug in Node.js.
Please fill in as much of the template below as you can.
Version: output of `node -v`
Platform: output of `uname -a` (UNIX), or version and 32 or 64-bit (Windows)
Subsystem: if known, please specify the affected core module name
If possible, please provide code that demonstrates the problem, keeping it as
simple and free of external dependencies as you can.
-->
* **Version**: v14.0.0-pre, master branch
* **Platform**: Linux
* **Subsystem**: inspector
<!-- Please provide more details below this comment. -->
Newly added `Debugger.setInstrumentationBreakpoint` inspector API doesn't work for me. Changes were introduced for request #24687. The same API works fine in Chrome.
I've created a test #31137. Breakpoints for scripts with sourcemaps only don't work either. If I use the protocol incorrectly, please let me know which additional methods should be called. | inspector | low | Critical |
543,931,838 | opencv | OpenCV on Raspberry Pi Zero W - Floating point exception / black empty images |
[CMake_output.txt](https://github.com/opencv/opencv/files/4012939/CMake_output.txt)
- OpenCV => v4.2.0
- OS (cross-compilation) => macOS Catalina 10.15.1
- Target OS => Raspberry Pi Zero W, Raspbian Buster (10)
- Toolchain: crosstool-ng v1.23.0, armv6-rpi-linux-gnueabi
I am trying to utilise OpenCV 4 on the Raspberry Pi Zero W, running Raspbian Buster (10). When doing any transformations on a captured image, the application exits with a floating-point exception.
I have cross-compiled OpenCV with various options, most notably the ENABLE_VFPV3 option, with no luck. I always get the floating-point exception. I use **crosstool-ng** for the cross-compilation on macOS, and CMake (through Jetbrains CLion, mostly).
OpenCV is configured with the following options before being built:
```
-DCMAKE_TOOLCHAIN_FILE=armv6-rpi-linux-gnueabi-toolchain.cmake -DBUILD_JPEG=ON -DBUILD_PNG=ON -DBUILD_TIFF=ON -DWITH_V4L=ON -DENABLE_VFPV3=ON -DBUILD_TESTS=OFF -DCMAKE_BUILD_TYPE=Debug -DBUILD_SHARED_LIBS=OFF
```
OpenCV builds without any issues, and so do the applications that utilise the libraries. I don't build the shared libs, as I want the compiled application to not be dependent on 3rd party libraries to be installed.
The toolchain file that I use for cross-compilation of both OpenCV and the test application:
```
set(CMAKE_SYSTEM_NAME Linux)
set(CMAKE_SYSTEM_PROCESSOR arm)
set(CMAKE_CXX_FLAGS "-static-libstdc++")
set(CMAKE_SYSROOT /Volumes/crosstool/x-tools/armv6-rpi-linux-gnueabi/armv6-rpi-linux-gnueabi/sysroot)
set(CMAKE_OSX_SYSROOT /Volumes/crosstool/x-tools/armv6-rpi-linux-gnueabi/armv6-rpi-linux-gnueabi/sysroot)
set(CMAKE_C_COMPILER "/Volumes/crosstool/x-tools/armv6-rpi-linux-gnueabi/bin/armv6-rpi-linux-gnueabi-gcc")
set(CMAKE_CXX_COMPILER "/Volumes/crosstool/x-tools/armv6-rpi-linux-gnueabi/bin/armv6-rpi-linux-gnueabi-g++")
```
I have successfully compiled and run various applications with no issues up to this point, using this same setup.
A simple test application to reproduce the error:
```
#include <iostream>
#include "opencv2/imgproc/imgproc.hpp"
#include "opencv2/video/video.hpp"
#include "opencv2/videoio/videoio.hpp"
using namespace cv;
using namespace std;
int main( int argc, char** argv )
{
VideoCapture cap(0); //capture the video from web cam
if ( !cap.isOpened() ) // if not success, exit program
{
cout << "Cannot open the web cam" << endl;
return -1;
}
while (true)
{
Mat image;
bool bSuccess = cap.grab();
if (!bSuccess) //if not success, break loop
{
cout << "Cannot read a frame from video stream" << endl;
break;
}
cap.retrieve(image);
cvtColor(image, image, COLOR_BGR2GRAY);
}
return 0;
}
```
The _CMakefiles.txt_:
```
cmake_minimum_required(VERSION 3.14)
project(vision)
set(CMAKE_CXX_STANDARD 17)
set(CMAKE_CXX_FLAGS "-static-libstdc++")
message( STATUS "Profile: ${PROFILE}" )
if ("${PROFILE}" STREQUAL "macos")
include_directories(
/usr/local/include/opencv4
)
link_directories(
/usr/local/lib
/usr/local/lib/opencv4/3rdparty
)
elseif ("${PROFILE}" STREQUAL "arm")
include_directories(
../opencv-arm/build/install/include/opencv4
)
link_directories(
../zlib/install/lib
../opencv-arm/build/install/lib
../opencv-arm/build/install/lib/opencv4/3rdparty
)
else()
message( ERROR "unsupported build profile" )
return()
endif()
add_executable(vision main.cpp)
target_link_libraries(
vision
# video
opencv_videoio
opencv_video
# gui
opencv_highgui
# image processing
opencv_imgcodecs
opencv_imgproc
opencv_core
libjpeg-turbo
libjasper
libtiff
libpng
libwebp
ittnotify
dl
z
)
if ("${PROFILE}" STREQUAL "arm")
target_link_libraries(
vision
tegra_hal
pthread
)
endif()
```
After attempting the color transformation (or any other transformation) on the RPi Zero, the application exits with a `Floating point exception`. Compiling and running on macOS works without any issues (`-DPROFILE=macos`).
My only thoughts at this stage are that the OpenCV compilation needs to do something special for the RPi Zero for the floating-point calculations to work. Any help/direction will be greatly appreciated.
I did make some progress in finding a workaround. Using remote GDB, I was able to pinpoint a divide-by-zero in the OpenCV parallelism implementation (`modules/core/src/parallel.cpp`, line 316). It seems ParallelLoopBodyWrapperContext.nstripes is zero. Following [this post](https://stackoverflow.com/a/56281760/3177406), I re-compiled OpenCV with parallelism disabled, which allows the processing to succeed. I've only managed to utilise the color conversion operation, though. Both _blur_ and _Canny_ produce empty/black images, although not causing the floating point exception.
The CMake configuration options that I now utilised are:
```
-DCMAKE_TOOLCHAIN_FILE=armv6-rpi-linux-gnueabi-toolchain.cmake -DBUILD_JPEG=ON -DBUILD_PNG=ON -DBUILD_TIFF=ON -DBUILD_TESTS=OFF -DCMAKE_BUILD_TYPE=Debug -DBUILD_SHARED_LIBS=OFF -DWITH_IPP=OFF -DWITH_TBB=OFF -DWITH_OPENMP=OFF -DWITH_PTHREADS_PF=OFF
```
A deeper investigation will probably be needed into why the parallelism fails on the Pi Zero. | category: build/install,incomplete | low | Critical |
543,955,996 | pytorch | Computing dot product of columns sliced from large matrix causes illegal memory access in CUDA | ## π Bug
If you compute the dot product of a column sliced from a larger CUDA matrix with itself using `th.mm()`, it will produce an illegal memory access. If you instead compute `(v**2).sum()` (doesn't use CUBLAS), it will not produce the illegal memory access. Additionally, if `v` is allocated separately rather than being sliced, it will not produce the error.
This may be a duplicate of #17897 or just a similar underlying CUBLAS issue. This script can succeed if N*K > 2^31 - 1 (the commented `N = ...` line is bigger but still works) so I'm not sure.
## To Reproduce
```python
import torch as th
N = 16909322
#N = 16909321 # << this one won't fail
K = 127
print(f"Allocating {4*N*K / (1024**3)} GB ...")
U = th.zeros((N, K), device="cuda:0")
print("Done.")
u = U[:, 0:1]
#u = th.zeros_like(u) # << adding this makes the following line succeed for either value of N
v = th.mm(u.permute(1, 0), u)
#v = (u**2).sum() # << this succeeds for either value of N
# Just to force the error to propagate up to Python, otherwise it will silently
# fail and exit.
th.cuda.synchronize()
```
## Environment
- PyTorch Version (e.g., 1.0): 1.3 (also verified still present in latest master)
- OS (e.g., Linux): Linux
- Python version: 3.7
- CUDA/cuDNN version: 9.0.176
- GPU models and configuration: V100
cc @ezyang @gchanan @zou3519 @csarofeen @ptrblck @ngimel | module: dependency bug,module: cuda,triaged,module: 64-bit,module: cublas | low | Critical |
543,959,414 | go | cmd/go: reject unenforced 'go' versions (older than 1.6) in 'go.mod' files | If I have a very simple Go package that doesn't depend on any language features added since the original Go 1 release, I naturally thought I could put in my go.mod file:
```
go 1
```
But that doesn't parse:
```
$ go install
go: errors parsing go.mod:
/home/me/proj/go.mod:3: usage: go 1.23
```
But `go 1.0` is weird, because we never called it that. Is that correct? Docs don't say.
`go 1.1` is also weird, because it implies that I'm depending on something that was added in Go 1.1, but I'm not.
So I guess `go 1.11` for when modules were introduced?
/cc @ianlancetaylor @bcmills @jayconrod @dmitshur | Documentation,help wanted,NeedsFix | low | Critical |
543,965,318 | youtube-dl | Unsupported URL: planeta-online.tv | <!--
######################################################################
WARNING!
IGNORING THE FOLLOWING TEMPLATE WILL RESULT IN ISSUE CLOSED AS INCOMPLETE
######################################################################
-->
## Checklist
<!--
Carefully read and work through this check list in order to prevent the most common mistakes and misuse of youtube-dl:
- First of, make sure you are using the latest version of youtube-dl. Run `youtube-dl --version` and ensure your version is 2019.12.25. If it's not, see https://yt-dl.org/update on how to update. Issues with outdated version will be REJECTED.
- Make sure that all provided video/audio/playlist URLs (if any) are alive and playable in a browser.
- Make sure that site you are requesting is not dedicated to copyright infringement, see https://yt-dl.org/copyright-infringement. youtube-dl does not support such sites. In order for site support request to be accepted all provided example URLs should not violate any copyrights.
- Search the bugtracker for similar site support requests: http://yt-dl.org/search-issues. DO NOT post duplicates.
- Finally, put x into all relevant boxes (like this [x])
-->
- [x] I'm reporting a new site support request
- [x] I've verified that I'm running youtube-dl version **2019.12.25**
- [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
<!--
Provide all kinds of example URLs support for which should be included. Replace following example URLs by yours.
-->
- Single video: http://www.planeta-online.tv/media/video/show/8254.html
- Single video: http://www.planeta-online.tv/media/video/show/8234.html
- Single TV-stream: http://www.planeta-online.tv/channels/6/live.html
- Single TV-stream: http://www.planeta-online.tv/channels/134/live.html
- Single TV-stream: http://www.planeta-online.tv/channels/2/live.html
## Description
<!--
Provide any additional information.
If work on your issue requires account credentials please provide them or explain how one can obtain them.
-->
All video and TV-stream HLS-examples above plays fine in Google Chrome browser.
| site-support-request | low | Critical |
543,968,245 | godot | Scrollbars do not reappear after disabling and re-enabling ScrollContainer scrolling. | **Godot version:**
master
**OS/device including version:**
Windows 10
**Issue description:**
If you disable horizontal or vertical scrolling on a ScrollContainer node, the scrollbar will not reappear upon re-enabling scrolling.
**Steps to reproduce:**
Create a ScrollContainer node with some content. Add the following script:
```
extends ScrollContainer
func _input(event):
if event is InputEventKey:
if event.pressed and not event.echo:
scroll_horizontal_enabled = not scroll_horizontal_enabled
scroll_vertical_enabled = not scroll_vertical_enabled
# uncomment for workaround
# set_visible(false)
# set_visible(true)
```
Run the project and press any key. The scrollbars will be hidden. Press any key again and the scrollbars do not return as expected. (Resizing the window will make them reappear.)
**Minimal reproduction project:**
[ScrollContainer_scrollbar_visibility_bug.zip](https://github.com/godotengine/godot/files/4010580/ScrollContainer_scrollbar_visibility_bug.zip)
| bug,topic:gui | low | Critical |
543,974,915 | go | cmd/go: convert tests using testdata files to script framework | I'd like to convert the cmd/go tests using testdata/src as their GOPATH to the script framework. We don't run those tests in parallel because we don't want concurrent go command runs started by the tests using the same GOPATH. While we can modify the test framework to copy the testdata files to a isolated temp directory GOPATH for those tests, I think it's simpler to just convert those tests to the script framework, and that has the additional benefit that we keep the test logic and data together in the same files making the tests easier to read. | Testing,NeedsFix | high | Critical |
543,991,860 | pytorch | AttributeError: module 'torch.distributed' has no attribute 'init_process_group' on torch 1.3 aarch64 | ## π Bug
I get the error:
> AttributeError: module 'torch.distributed' has no attribute 'init_process_group'
when running the following code (see reproduce). I am using the Jetson Nano and have installed torch through the [PyTorchv1.3.0 pip wheel for linux aarch64 using Python 3.6](https://devtalk.nvidia.com/default/topic/1049071/pytorch-for-jetson-nano-version-1-3-0-now-available/).
## To Reproduce
Steps to reproduce the behavior:
Run the following code:
```
"""run.py:"""
#!/usr/bin/env python
import os
import torch
import torch.distributed as dist
from torch.multiprocessing import Process
def run(rank, size):
""" Distributed function to be implemented later. """
pass
def init_process(rank, size, fn, backend='gloo'):
""" Initialize the distributed environment. """
os.environ['MASTER_ADDR'] = '127.0.0.1'
os.environ['MASTER_PORT'] = '29500'
dist.init_process_group(backend, rank=rank, world_size=size)
fn(rank, size)
if __name__ == "__main__":
size = 2
processes = []
for rank in range(size):
p = Process(target=init_process, args=(rank, size, run))
p.start()
processes.append(p)
for p in processes:
p.join()
```
Which is [just some tutorial code.](https://pytorch.org/tutorials/intermediate/dist_tuto.html)
Output:
```
Process Process-1:
Process Process-2:
Traceback (most recent call last):
File "/usr/lib/python3.6/multiprocessing/process.py", line 258, in _bootstrap
self.run()
File "/usr/lib/python3.6/multiprocessing/process.py", line 93, in run
self._target(*self._args, **self._kwargs)
File "run.py", line 16, in init_process
dist.init_process_group(backend, rank=rank, world_size=size)
AttributeError: module 'torch.distributed' has no attribute 'init_process_group'
Traceback (most recent call last):
File "/usr/lib/python3.6/multiprocessing/process.py", line 258, in _bootstrap
self.run()
File "/usr/lib/python3.6/multiprocessing/process.py", line 93, in run
self._target(*self._args, **self._kwargs)
File "run.py", line 16, in init_process
dist.init_process_group(backend, rank=rank, world_size=size)
AttributeError: module 'torch.distributed' has no attribute 'init_process_group'
```
<!-- If you have a code sample, error messages, stack traces, please provide it here as well -->
## Environment
Output:
```
PyTorch version: 1.3.0
Is debug build: No
CUDA used to build PyTorch: 10.0.326
OS: Ubuntu 18.04.2 LTS
GCC version: (Ubuntu/Linaro 7.4.0-1ubuntu1~18.04.1) 7.4.0
CMake version: Could not collect
Python version: 3.6
Is CUDA available: Yes
CUDA runtime version: Could not collect
GPU models and configuration: Could not collect
Nvidia driver version: Could not collect
cuDNN version: /usr/lib/aarch64-linux-gnu/libcudnn.so.7.3.1
Versions of relevant libraries:
[pip3] numpy==1.17.4
[pip3] torch==1.3.0
[pip3] torchvision==0.5.0a0+2d7c066
[conda] Could not collect
```
cc @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @rohan-varma @gqchen @aazzolini @xush6528 | oncall: distributed,triaged | low | Critical |
544,005,280 | angular | Unified Angular Location Service incorrectly parses the initial URL when in useHash mode | # π bug report
### Affected Package
The issue is caused by package @angular/common/upgrade
### Is this a regression?
probably not
### Description
When:
- The Unified Location Service is being used
- The `{ useHash: true }` mode is active
- During bootstrap, the browser URL contains an initial route in the hash, such as `http://mydomain.com/#/initialRoute`
Then:
I expect the $location path to be `/initialRoute`.
However, the $location `path` is `/` and the `hash` is `/initialRoute`
## π¬ Minimal Reproduction
[This stackblitz](https://stackblitz.com/edit/angular-hybrid-uniified-location-initial-path?file=angularjs.module.ts) demonstrates the problem.
When it loads, click the button labelled "Reload page with an initial path in the hash".
The stackblitz also has a toggle between the `$locationShim` from `@angular/common/upgrade` and the old AngularJS `$location` service showing the difference in behavior parsing the initial URL.
## π Your Environment
**Angular Version: 8**
with @angular/upgrade
**Anything else relevant?**
This issue is related to #34596. | type: bug/fix,area: common,freq1: low,area: upgrade,state: confirmed,P3 | low | Critical |
544,010,348 | TypeScript | When refactoring by renaming destructured arguments, solution is verbose | ## Search Terms
refactor renaming redundant
renaming destructured arguments
## Suggestion
Currently VSCode does an amazing job of renaming variables. However, when renaming, destructuring assignments end up overly wordy / misleading.
For example:
```
function contrivedExample(props: PropType) {
const { prop1, prop2, prop4 } = props;
return { prop3: prop4 };
}
```
When renaming `prop4` to `prop3`, this correctly becomes:
```
function contrivedExample(props: PropType) {
const { prop1, prop2, prop3 } = props;
return { prop3: prop3 };
}
```
but it would be even nicer if it could simply become:
```
function contrivedExample(props: PropType) {
const { prop1, prop2, prop3 } = props;
return { prop3 };
}
```
NOTE: this was [originally requested](https://github.com/microsoft/vscode/issues/87535) of the VSCode team but they seem to feel this is a TypeScript-level feature.
## Use Cases, Examples
Please see above for an illustrative example. This is a convenience only.
## 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 |
544,021,083 | flutter | Hero animation with stateful widget | ## Steps to Reproduce
is it possible to save state when using Hero animation?
1. create statefull widget, for example counter
2. wrap in on Hero widget
3. click on button to change state
4. change page
5. it's not saving state
```dart
import 'package:flutter/material.dart';
void main() => runApp(HeroApp());
class HeroApp extends StatelessWidget {
@override
Widget build(BuildContext context) {
return MaterialApp(
title: 'Transition Demo',
home: MainScreen(),
);
}
}
class MainScreen extends StatelessWidget {
@override
Widget build(BuildContext context) {
return Scaffold(
appBar: AppBar(
title: Text('Main Screen'),
),
body: Column(
children: <Widget>[
Hero(
tag: 'imageHero',
key: ValueKey('imageHero'),
child: HeroWidget(
key: ValueKey('imageHero'),
),
),
FlatButton(
onPressed: () =>
Navigator.push(context, MaterialPageRoute(builder: (_) {
return DetailScreen();
})),
child: Text('navigate'))
],
),
);
}
}
class HeroWidget extends StatefulWidget {
const HeroWidget({
Key key,
}) : super(key: key);
@override
_HeroWidgetState createState() => _HeroWidgetState();
}
class _HeroWidgetState extends State<HeroWidget> {
int count = 1;
@override
Widget build(BuildContext context) {
return Container(
child: RaisedButton(
child: Text('clicked $count'),
onPressed: () {
setState(() {
count++;
});
},
),
);
}
}
class DetailScreen extends StatelessWidget {
@override
Widget build(BuildContext context) {
return Scaffold(
body: Center(
child: Column(
children: <Widget>[
Hero(
tag: 'imageHero',
child: HeroWidget(
key: ValueKey('imageHero'),
),
),
FlatButton(
onPressed: () => Navigator.pop(context), child: Text('back'))
],
),
),
);
}
}
```
**Target Platform:** iOS
**Target OS version/browser:**
**Devices:** Iphone X
```
[β] Flutter (Channel stable, v1.12.13+hotfix.5, on Mac OS X 10.15.2 19C57,
locale en-AM)
β’ Flutter version 1.12.13+hotfix.5 at /Users/narek/development/flutter
β’ Framework revision 27321ebbad (3 weeks ago), 2019-12-10 18:15:01 -0800
β’ Engine revision 2994f7e1e6
β’ Dart version 2.7.0
[β] Android toolchain - develop for Android devices (Android SDK version 29.0.0)
β’ Android SDK at /Users/narek/Library/Android/sdk
β’ Android NDK location not configured (optional; useful for native profiling
support)
β’ Platform android-29, build-tools 29.0.0
β’ Java binary at: /Users/narek/Library/Application
Support/JetBrains/Toolbox/apps/AndroidStudio/ch-0/191.6010548/Android
Studio.app/Contents/jre/jdk/Contents/Home/bin/java
β’ Java version OpenJDK Runtime Environment (build
1.8.0_202-release-1483-b49-5587405)
β’ All Android licenses accepted.
[β] Xcode - develop for iOS and macOS (Xcode 11.3)
β’ Xcode at /Applications/Xcode.app/Contents/Developer
β’ Xcode 11.3, Build version 11C29
β’ CocoaPods version 1.8.3
[β] Android Studio (version 3.5)
β’ Android Studio at /Users/narek/Library/Application
Support/JetBrains/Toolbox/apps/AndroidStudio/ch-0/191.6010548/Android
Studio.app/Contents
β’ Flutter plugin version 42.1.1
β’ Dart plugin version 191.8593
β’ Java version OpenJDK Runtime Environment (build
1.8.0_202-release-1483-b49-5587405)
[β] VS Code (version 1.41.1)
β’ VS Code at /Applications/Visual Studio Code.app/Contents
β’ Flutter extension version 3.7.1
[β] Connected device (1 available)
β’ iPhone 8 β’ 9C589AC5-E580-4478-8D0D-2EC07550E74D β’ ios β’
com.apple.CoreSimulator.SimRuntime.iOS-13-3 (simulator)
β’ No issues found!
```
| c: new feature,framework,f: material design,f: routes,P3,workaround available,team-design,triaged-design | low | Major |
544,029,848 | rust | Incorrect/unexpected behavior with associated type lifetime bounds in lifetime-parametrized traits | ```Rust
trait Bound<'a> {
type Assoc;
}
trait Bounded<'a, C: Bound<'a>>
where
<C as Bound<'a>>::Assoc: 'static,
{
}
impl<'a, T> Bounded<'a, T> for T
where
T: Bound<'a>,
<T as Bound<'a>>::Assoc: 'static,
{
}
```
This fails with ```the associated type `<T as Bound<'_>>::Assoc` may not live long enough
consider adding an explicit lifetime bound `<T as Bound<'_>>::Assoc: 'static```, despite that very bound clearly being present.
This problem is resolved by removing the `'a` parametrization on `Bound` i.e.
```Rust
trait Bound {
type Assoc;
}
```
or by adding the bound explicitly to `Assoc` at definition-site
```Rust
trait Bound {
type Assoc: 'static;
}
```
but not, for some reason, by adding a HRTB and/or bound on the initial type parameter that should (as far as I understand this) be equivalent
```Rust
impl<'a, T> Bounded<'a, T> for T
where
T: Bound<'a> + 'static,
<T as Bound<'a>>::Assoc: 'static,
for<'b> <T as Bound<'b>>::Assoc: 'static
{
}
```
[playground](https://play.rust-lang.org/?version=stable&mode=debug&edition=2018&gist=84929734a411d595b93ed890d17d6ecf) | A-lifetimes,A-trait-system,A-associated-items,T-compiler,C-bug | low | Critical |
544,037,053 | flutter | Tool appears to hang when connected to multiple devices and one's screen has turned off | Steps to reproduce:
1. Connect two devices
2. `flutter run -d all` on some Flutter app (e.g. `flutter create` template app)
3. Turn the screen off on one of the devices.
4. Press "P" to show the performance overlay.
5. Press "h" to show the help.
Expected results: Performance overlay shows on the device whose screen is on. Help screen displays in the console.
Actual results: Nothing happens (assuming the phone you turned off happens to be the one we try to update first). User gets very confused and things everything has hung. | tool,a: quality,P2,team-tool,triaged-tool | low | Major |
544,038,664 | godot | Changing type of parent node makes camera gizmo disappear | Godot 3.2 beta4
Windows 10 64 bits
OpenGL ES 3.0 Renderer: GeForce GTX 1060 6GB/PCIe/SSE2
1) Create a new scene with `Node` as root
2) Add a `Camera` child node
3) Change the type of the root node to be a `Spatial`: observe the camera wireframes disappear, and turn into a box if selected | bug,topic:editor,confirmed | low | Minor |
544,040,596 | flutter | Device name pattern matching doesn't work past first space character ("-d ...") | I have a device with the name `SM G960U`.
`flutter logs -d SM` works fine.
`flutter logs -d G96` does not.
In fact, anything past the space, including the space (`flutter logs -d " "`), does not work. | tool,P2,team-tool,triaged-tool | low | Minor |
544,045,589 | TypeScript | Issue errors on tslib version ranges in `package.json` | Similar in spirit to https://github.com/microsoft/TypeScript/issues/33603
@bterlson and friends on the Azure SDK team mentioned that there was an issue with `tslib` where a local installation might grab `[email protected]`, but could have issues in other environments because their `package.json` specified a minimum version (`1.9.x`) which was too old.
I don't think that we generally dive into `package.json` to issue errors, but this seemed pretty unexpected, and it leads to more obvious behavior than with issuing errors about missing exports.
On the other hand, this might be a problem for slightly more esoteric but legitimate use-cases like alternate implementations/sources of `tslib`. For example, a user could point `tslib` at a local folder or a GitHub URL. | Suggestion,Needs Proposal | low | Critical |
544,056,413 | TypeScript | Support JSX transformations with custom jsxFactory | ## Search Terms
`jsx`, `jsxFactory`, `transformer`
## Suggestion
Allow tapping into the JSX generation process to control `React.createElement` output format when using `--jsxFactory`.
_Specific cases requested (e.g. #15217, #32619) would also benefit from a generalized solution._
## Use Cases
### Emit component children as thunks
Useful to reactive frameworks which attach dependency tracking via the JSX component hierarchy.
The internal [createJsxFactoryExpression()](https://github.com/microsoft/TypeScript/blob/36c87aca8a1b0fd6aa079e378cd176cf1cbf4467/src/compiler/factory.ts#L129-L136) call LHS is controlled using the `--jsxFactory` compiler option:
```ts
// jsx
<div>
<div>a</div>
<div>b</div>
</div>
// current emit (jsxFactory: "h")
h('div', null,
h('div', null, 'a'),
h('div', null, 'b')
)
```
This produces a nested `React.createElement()`/`h()` chain which evaluates eagerly.
With (still synchronous) thunks, frameworks can track each component dependency as they traverse the tree, for fine-grained change detection:
```ts
// thunks (needs transformer)
h('div', null, () => [
h('div', null, () => ['a']),
h('div', null, () => ['b'])
])
```
This requires recursively transforming the JSX emits to a signature of `h()` which accepts some args as a thunk.
I've written a hacky (#16607) transformer for this but had to extract most of the JSX-related internals from the TypeScript compiler, whereas making the change to [createExpressionForJsxElement()](https://github.com/microsoft/TypeScript/blob/36c87aca8a1b0fd6aa079e378cd176cf1cbf4467/src/compiler/factory.ts#L138-L168) is trivial:
```ts
const cs = argumentsList.slice(2);
const argThunk = ts.createArrowFunction(
[],
[],
[],
undefined,
undefined,
ts.createArrayLiteral(cs)
);
const thunkified = [...argumentsList.slice(0, 2), argThunk];
```
According to #34547, plans are already in motion to pass nested component args as arrays instead of varargs. Specifying the `jsxFactory` implies we are assuming control of the composition behavior; we should also control the function signature.
| Suggestion,Awaiting More Feedback | low | Minor |
544,062,504 | vscode | Issue Reporter editor no undo/redo | Issue Type: <b>Bug</b>
It seems we can't run undo/redo in the comment editor of issue reporter.
VS Code version: Code 1.41.1 (26076a4de974ead31f97692a0d32f90d735645c0, 2019-12-18T14:57:51.166Z)
OS version: Darwin x64 19.0.0
<details>
<summary>System Info</summary>
|Item|Value|
|---|---|
|CPUs|Intel(R) Core(TM) i9-9980HK CPU @ 2.40GHz (16 x 2400)|
|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<br>surface_control: disabled_off<br>surface_synchronization: enabled_on<br>video_decode: enabled<br>viz_display_compositor: enabled_on<br>viz_hit_test_surface_layer: disabled_off<br>webgl: enabled<br>webgl2: enabled|
|Load (avg)|2, 2, 2|
|Memory (System)|32.00GB (0.38GB free)|
|Process Argv||
|Screen Reader|no|
|VM|0%|
</details><details><summary>Extensions (1)</summary>
Extension|Author (truncated)|Version
---|---|---
python|ms-|2019.11.50794
(2 theme extensions excluded)
</details>
<!-- generated by issue reporter --> | feature-request,issue-reporter,papercut :drop_of_blood: | low | Critical |
544,063,763 | opencv | Is there alternative function for dnn_registerLayer? | Now, i am changing to python code to java code.
I have tested various opencv versions ( 4.2 4.1 3.4.9 3.4.4)
I have tried to run caffe.model ( using dnn)
In Python code, i used cv2.dnn_registerLayer function
But In opencv sdk, there is no function.
i searched it on Google, but i couldnot found.
Is there alternative function for dnn_registerLayer?
Please help me T.T
<!--
If you have a question rather than reporting a bug please go to http://answers.opencv.org where you get much faster responses.
If you need further assistance please read [How To Contribute](https://github.com/opencv/opencv/wiki/How_to_contribute).
Please:
* Read the documentation to test with the latest developer build.
* Check if other person has already created the same issue to avoid duplicates. You can comment on it if there already is an issue.
* Try to be as detailed as possible in your report.
* Report only one problem per created issue.
This is a template helping you to create an issue which can be processed as quickly as possible. This is the bug reporting section for the OpenCV library.
-->
##### System information (version)
<!-- Example
- OpenCV => 3.1
- Operating System / Platform => Windows 64 Bit
- Compiler => Visual Studio 2015
-->
- OpenCV => :grey_question:
- Operating System / Platform => :grey_question:
- Compiler => :grey_question:
##### Detailed description
<!-- your description -->
##### Steps to reproduce
<!-- to add code example fence it with triple backticks and optional file extension
```.cpp
// C++ code example
```
or attach as .txt or .zip file
-->
| category: dnn | low | Critical |
544,064,023 | godot | Extend Script is inheriting from the wrong class | **Godot version:**
3.2.beta4.mono.offical
**OS/device including version:**
Windows 10
**Issue description:**
The extend script tool appears to be ignoring the attached C# script and instead is extending the base node type.
**Steps to reproduce:**
1. Use the Scene panel to add a new node of type Node.
1. Use the Scene panel to select this node then right-click and and select Attach Script. Use Langauge: `C#` and name: `ScriptA.cs`.
2. Use the Scene panel to select this node then right-click and select Extend Script.
Note that the dialog defaults to inheriting `Node` not `ScriptA.cs`.
| bug,topic:editor,topic:dotnet | low | Minor |
544,068,516 | pytorch | sparse tensor eliminate_zeros | ## π Feature
In `scipy.sparse.coo_matrix`, there is a method [`eliminate_zeros`](https://github.com/scipy/scipy/blob/v1.4.1/scipy/sparse/coo.py#L556-L564). It eliminates all the zero values in `values` and the corresponding indices in `indices`, which I think simple but sometimes useful.
I tried to implement it myself for pytorch, but I found that `indices` and `values` are both private attributes for sparse tensor, and I didn't find any `set` functions for them. Meanwhile, if we get the `indices` and `values`, and eliminate zeros, and then create a new sparse tensor with eliminated ones, this way works but in my opinion inefficient. So I think maybe we need to make method `eliminate_zeros` in the sparseTensor class.
cc @nikitaved @pearu @cpuhrsch @IvanYashchuk @vincentqb | module: sparse,feature,triaged | low | Major |
544,069,630 | godot | 3.2. Object disappearing when using the scrollwheel outside of running window Visibility Rect | **Godot version:** 3.2 beta4
**OS/device including version:** Win7
**Issue description:**
When using the mouse scrollwheel outside of the running window, particles disappear. Moving the mouse back into the window area causes the particles to reappear. The scrollwheel has no affect if mouse is inside the running window. There is no input map assigned to the mouse scrollwheel. No code using the scrollwheel scancode. The issue is present when running the project from the editor as well as when running the exported project.
**Steps to reproduce:**
Run the minimal project and scroll inside vs outside the running window. Edit: Scroll beyond the Visibility Rect!
**Minimal reproduction project:**
[particle_test_scrollwheel.zip](https://github.com/godotengine/godot/files/4011745/particle_test_scrollwheel.zip)
Godot4 Beta4 MRP:
[scrollwheel_visibility_test.zip](https://github.com/godotengine/godot/files/9984247/scrollwheel_visibility_test.zip)
| bug,topic:rendering,needs testing,topic:2d,topic:particles | medium | Major |
544,170,753 | flutter | Equivalent of `InkWell` in Cupertino style | We already have `InkWell` widget for tap effect in material style. Is there any equivalent in Cupertino style?
Example here: https://developer.apple.com/design/human-interface-guidelines/ios/images/Gestures_Tap.mp4
After tapping the **Sounds & Haptics** section, its background color turns into gray immediately. | c: new feature,platform-ios,framework,a: fidelity,f: cupertino,customer: crowd,c: proposal,P2,team-design,triaged-design | low | Critical |
544,208,025 | electron | Remove the size limit of picture in picture video popout | * [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
Electron 7 enabled the ability to toggle picture in picture (pip) mode for videos. It works great but the resulting popout has a a maximum size limitation (around 1/3 of screen). I think that this limitation impacts potential use of this functionality. For instance you cannot move video to another screen and use it's full size.
### Proposed Solution
Remove the maximum size limitation and allow the user to set any video size.
| enhancement :sparkles: | low | Minor |
544,220,736 | storybook | Reduce size of react-syntax-highlighter dependency | When built, the manager + preview are around 2.6 MB excluding the user's code. Even for a development tool, this is quite a hefty amount.
I've been investigating different ways this can be reduced, and by far the largest contributor to the bundle size is `react-syntax-highlighter`. It's an astronomical 1.2 MB π² https://bundlephobia.com/[email protected]
This is a picture of just the manager, broken down into dependencies by size:
`highlight.js` and `refractor` are both dependencies of `react-syntax-highlighter`, and as you can see the configuration for many different languages is loaded.
The docs explain the size can be reduced by using a [light build](https://github.com/conorhastings/react-syntax-highlighter#light-build) or an [async build](https://github.com/conorhastings/react-syntax-highlighter#async-build).
The async build looks ideal and doesn't introduce breaking changes, but I don't know if you guys want to include dynamic imports in storybook. The light build can also work, but might require the user to load the languages or provide configuration of languages to include.
I'm happy to work on this if this is something you guys want. | ui,performance issue | low | Critical |
544,272,149 | vscode | Diff editor: Differentiate unimportant differences | I would like the ability to differentiate unimportant/minor differences in the diff editor. This is somewhat related to #43026, except this is not only about whitespace and ideally _ignore_ should only be one of the ways this differentiation is used (e.g. different coloring). Obviously the semantics for what's "unimportant" is subjective and/or language-dependent. You might be able to use the language grammar (or new semantic API) to determine this in addition to some user configurability. **Beyond Compare** allows this differentiation based on a grammar with user configurability, and it's very useful.
The following are some examples of unimportant differences:
- Whitespace in a whitespace-insensitive language
- Case in a case-insensitive language
- Comments
- Strings?
| feature-request,diff-editor | medium | Minor |
544,275,350 | pytorch | Need a launch utility for Distributed RPC framework. | We need a launch utility similar to what we have for DDP: https://github.com/pytorch/pytorch/blob/master/torch/distributed/launch.py to easily launch multiple RPC workers on the same host. We could either reuse the same utility and extend it to support the Distributed RPC Framework as well.
cc @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @gqchen @aazzolini @rohan-varma @xush6528 @jjlilley | triaged,better-engineering,module: rpc | low | Minor |
544,290,505 | flutter | Need better documentation for DrawShadow (what is elevation?) | Neither the Dart side [`Canvas.drawShadow`](https://github.com/flutter/engine/blob/eb05391962fb74e0152d6bb9827abdf341113de5/lib/ui/painting.dart#L4018) nor the engine side [`Canvas::drawShadow`](https://github.com/flutter/engine/blob/eb05391962fb74e0152d6bb9827abdf341113de5/lib/ui/painting/canvas.cc#L413), i.e. [`flutter::PhysicalShapeLayer::DrawShadow`](http://34.94.103.149/classflutter_1_1_physical_shape_layer.html#ac5d99b79c8a24ca58446d601af248db9) have *any* information on how **`elevation`** works exactly and I could also not find out what values do what exactly at different scales etc.
The only thing I can say for sure is that the Flutter web implementation of the method behaves completely different than the native version.
## Needs better documentation
I think that better documentation for this function is needed in general, but more specifically I would like to know how `elevation` works exactly - from the native code I would have concluded that values from `0` to `600` make sense because [the light source is positioned at `z = 600`](http://34.94.103.149/namespaceflutter.html#af5a46f83182e00e27bf950417af3ea2b), however, that does not exactly line up with my experiments in native (I gave up on web behavior).
Additionally, it would be useful to know where the light source is positioned - I would assume that it is positioned at `(0, 0, 600 * dpr)`, but I have no idea at this point. | engine,d: api docs,P3,team-engine,triaged-engine | low | Minor |
544,294,519 | go | encoding/gob: permit registering duplicate names | In go1.9, type aliases were introduced. The purpose of type aliases is to permit the movement of types from one location to another. This implies:
1) that the _real_ name of the type can change over time
2) that there can be _multiple_ names for the exact same Go type
The use of type aliases interacts poorly with gob-encoded data stored on disk that represent interface values as `gob` was a developed in an era where there was one-to-one relationship between types and names.
Consider the following examples.
At some point in the past, we [gob-encoded is stored on disk using something like](https://play.golang.org/p/dt0pW47u17P):
```go
type Rect struct { ... }
func (r Rect) Area() int { ... }
type Any struct {
Shape interface{ Area() int }
}
func main() {
gob.Register(Rect{})
var b bytes.Buffer
e := gob.NewEncoder(&b)
e.Encode(Shape{Rect{2, 5}})
// assume that b is stored on disk
fmt.Printf("%q\n", b.String())
}
```
On the decoder side, [this works fine with code like](https://play.golang.org/p/GXlOGbcTGvy):
```go
type Rect struct { ... }
func (r Rect) Area() int { ... }
type Any struct {
Shape interface{ Area() int }
}
func main() {
// Assume that the contents of r is loaded from disk as created by:
// https://play.golang.org/p/dt0pW47u17P
r := strings.NewReader(...)
gob.Register(Rect{})
d := gob.NewDecoder(r)
v := new(Shape)
fmt.Println(d.Decode(v)) // got nil; works as expected
}
```
At some point in the future, [suppose we move `Rect`, now decoding fails](https://play.golang.org/p/6vu43aUdTm-):
```go
type Rect = Rect2 // this is now a type alias to a different name
type Rect2 struct { ... }
func (r Rect2) Area() int { ... }
type Any struct {
Shape interface{ Area() int }
}
func main() {
// Assume that the contents of r is loaded from disk as created by:
// https://play.golang.org/p/dt0pW47u17P
r := strings.NewReader(...)
gob.Register(Rect{})
d := gob.NewDecoder(r)
v := new(Shape)
fmt.Println(d.Decode(v)) // got `gob: name not registered for interface: "main.Rect"`, expected nil
}
```
The problem is that the full type name is encoded in the gob wire data. Trying to decode the data at a later date does not know about the movement of the type due to type aliases.
Given that problem, one would think that you could just use `gob.RegisterName` to manually register the old name:
```go
gob.RegisterName(Rect{})
gob.RegisterName("main.Rect", Rect{})
```
However, that panics with:
```
panic: gob: registering duplicate names for main.Rect2: "main.Rect2" != "main.Rect"
```
I propose the following changes:
1) That `gob` be better documented about the interaction of type aliases and serialization of interface values.
2) That `gob.RegisterName` be relaxed to permit duplicate names. However, this can cause confusion in the implementation as to which name to use when marshaling. Alternatively, we could add a `gob.RegisterAlias` that makes it clear that the duplicate names are aliases. | NeedsInvestigation,FeatureRequest | low | Minor |
544,317,777 | go | cmd/compile: teach prove about incrementing, len, and overflow | I have some hot code that looks approximately like this:
```go
func f(a []int, i int) int {
for i < len(a) && a[i] != 0 {
i++
}
return i
}
```
The generated code has a bounds check on every iteration. This is because `i` might be negative. We can rule that out up front:
```go
func f(a []int, i int) int {
if i < 0 {
panic("bad")
}
for i < len(a) && a[i] != 0 {
i++
}
return i
}
```
This could should be able to go without a bounds check. `i` cannot be negative at the beginning. In order to become negative, it must overflow. However, it cannot overflow: It is being incremented, so it must reach `len(a)` before overflowing, in which case we exit the loop before evaluating `a[i]`.
(I actually know a priori in my code that `i` is not negative; I'd be tempted to use #30582 here. I could also switch to `i` being a uint, which does remove the bounds check, but it makes other things in the real code a bit awkward.)
I don't know how hard it is to teach prove about this case.
cc @zdjones @rasky @randall77
| Performance,NeedsInvestigation,compiler/runtime | low | Minor |
544,323,808 | vue | serverPrefetch rejection with no value creates broken response | ### Version
2.6.11
### Reproduction link
[http://sorry-no-can-do.com](Sorry, I don't have time to setup a reproduction case for SSR, that's not a 5 minute thing).
### Steps to reproduce
1. Use SSR
2. Add a `serverPrefetch` with `return Promise.reject()`
### What is expected?
An error to occur due to the rejection in `serverPrefetch`
### What is actually happening?
No error occurs, but the rendered result will not contain the component with the bad `serverPrefetch`.
---
Currently, rejecting a promise in `serverPrefetch` with no value will create a broken response from SSR. By broken I mean that no error will occur with `renderToString`, so it looks like it was successful, but the rendered DOM may be missing components, including the root component all together.
This is because the rejection value from `serverPrefetch` is wired straight through to the `done` callback on `render`, but that callback only considers an error to have occurred if a value is provided. Due to this straight wiring, any `serverPrefetch` which rejects without a value will allow rendering to continue, but will skip the rendering of the component on which the error occurs. If that's the root component, the rendered result has no element for hydration to use in the rendered result, which means there's no way for the client to recover. For a real-life example of this, see vuejs/vue-apollo#901.
Here's the relevant code, the `done` callback for `render`:
https://github.com/vuejs/vue/blob/42fdf3f8378e5dc379808a4fa67b39c4a74ae191/src/server/create-renderer.js#L84-L87
And where the rejection value from `serverPrefetch` is wired through:
https://github.com/vuejs/vue/blob/42fdf3f8378e5dc379808a4fa67b39c4a74ae191/src/server/render.js#L65
A fix could be as easy as expanding that `catch` to check for `err` and provide a default value if one is not provided. I'm punting on what that default would be, so I'm opening this as an issue rather than a PR.
<!-- generated by vue-issues. DO NOT REMOVE --> | improvement,has workaround | low | Critical |
544,334,476 | TypeScript | Missing syntax error for destructuring with private names | <!-- π¨ STOP π¨ π¦π§π’π£ π¨ πΊπ»πΆπ· π¨
Half of all issues filed here are duplicates, answered in the FAQ, or not appropriate for the bug tracker. Even if you think you've found a *bug*, please read the FAQ first, especially the Common "Bugs" That Aren't Bugs section!
Please help us by doing the following steps before logging an issue:
* Search: https://github.com/Microsoft/TypeScript/search?type=Issues
* Read the FAQ: https://github.com/Microsoft/TypeScript/wiki/FAQ
Please fill in the *entire* template below.
-->
<!--
Please try to reproduce the issue with the latest published version. It may have already been fixed.
For npm: `typescript@next`
This is also the 'Nightly' version in the playground: http://www.typescriptlang.org/play/?ts=Nightly
-->
**TypeScript Version:** 3.8.0-dev.20191228
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:**
**Code**
```ts
class C{
#prop = 1;
method() {
const { #prop/*1*/: prop } = this;
({ #prop/*2*/: this.#prop } = this);
}
}
```
**Expected behavior:**
Error at `/*1*/` and `/*2*/` since this syntax is not (yet) allowed. https://github.com/tc39/proposal-class-fields/issues/4
Thanks to @dragomirtitian for pointing this out.
**Actual behavior:**
No error.
**Playground Link:** https://www.typescriptlang.org/play/?ts=3.8.0-dev.20191228&ssl=7&ssc=2&pln=1&pc=1#code/MYGwhgzhAEDCDeAoaLoGIAOAnA9h6AvNAIwDcyqAtgKYAuAFjgCYAUAlNEqt9MDgHYRandNjwAuaGPwBfQtAYBLCOR6oW8UbgySlEAHSZt0OUT1tVqGYhlA
**Related Issues:** #35934 #35935 #35936
| Bug | low | Critical |
544,350,634 | pytorch | DataLoader: Segmentation Fault (core dumped) | ## π Bug
<!-- A clear and concise description of what the bug is. -->
## To Reproduce
Steps to reproduce the behavior:
1. python train.py
1.
```
dataloader = DataLoader(
dataset,
batch_size=1,
num_workers=0,
pin_memory=True)
for index, item in enumerate(dataloader):
```
## Expected behavior
train
<!-- A clear and concise description of what you expected to happen. -->
## Environment
Please copy and paste the output from our
[environment collection script](https://raw.githubusercontent.com/pytorch/pytorch/master/torch/utils/collect_env.py)
(or fill out the checklist below manually).
You can get the script and run it with:
```
wget https://raw.githubusercontent.com/pytorch/pytorch/master/torch/utils/collect_env.py
# For security purposes, please check the contents of collect_env.py before running it.
python collect_env.py
```
- PyTorch Version (e.g., 1.0):
- OS (e.g., Linux):
- How you installed PyTorch (`conda`, `pip`, source):
- Build command you used (if compiling from source):
- Python version:
- CUDA/cuDNN version:
- GPU models and configuration:
- Any other relevant information:
PyTorch version: 1.1.0
Is debug build: No
CUDA used to build PyTorch: 10.0.130
OS: Ubuntu 16.04.6 LTS
GCC version: (Ubuntu 5.4.0-6ubuntu1~16.04.12) 5.4.0 20160609
CMake version: version 3.5.1
Python version: 3.7
Is CUDA available: Yes
CUDA runtime version: 9.0.176
GPU models and configuration: GPU 0: GeForce RTX 2080 Ti
Nvidia driver version: 430.50
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.7.6.4
/usr/local/lib/libcudnn.so.7
Versions of relevant libraries:
[pip3] numpy==1.16.4
[conda] _pytorch_select 0.2 gpu_0 defaults
[conda] blas 1.0 mkl https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
[conda] cuda100 1.0 0 pytorch
[conda] cuda90 1.0 h6433d27_0 pytorch
[conda] mkl 2018.0.3 1 defaults
[conda] mkl-service 1.1.2 py37h90e4bf4_5 defaults
[conda] mkl_fft 1.0.6 py37h7dd41cf_0 defaults
[conda] mkl_random 1.0.1 py37h4414c95_1 defaults
[conda] pytorch 1.1.0 py3.7_cuda10.0.130_cudnn7.5.1_0 pytorch
[conda] torchvision 0.3.0 py37_cu10.0.130_1 pytorch
## Additional context
<!-- Add any other context about the problem here. -->
cc @ezyang @gchanan @zou3519 @SsnL @VitalyFedyunin | high priority,needs reproduction,module: crash,module: dataloader,triaged | medium | Critical |
544,351,374 | pytorch | Allow using digits in names of named tensors | ## π Feature
Currently you cannot use digits in names of tensors. Please allow digits, so that `torch.randn(64, 64, 4, 4, names=("height", "width", "channel_0", "channel_1"))` works rather than throwing an error
## Motivation
<!-- Please outline the motivation for the proposal. Is your feature request related to a problem? e.g., I'm always frustrated when [...]. If this is related to another GitHub issue, please link here too -->
I am studying machine learning models which are hybrid between neural networks and tensor networks. When working with tensor networks, we have a lot of tensors with a lot of dimensions. Sometimes we have multiple dimensions corresponding to, say, channels. Or sometimes, for example in tensor train, we have a lot of bond dimensions, which come in pairs. For example, in the model I am currently working on, at some point I have an intermediate representation of a batch of images as 4 tensors, for which I want dimensions to have the following semantically meaningful names:
```
#core_0 has shape ('batch', 'height', 'width', 'out_quantum_0', 'bond_0', 'bond_1'),
#core_1 has shape ('batch', 'height', 'width', 'out_quantum_1', 'bond_1', 'bond_2'),
#core_2 has shape ('batch', 'height', 'width', 'out_quantum_2', 'bond_2', 'bond_3'),
#core_3 has shape ('batch', 'height', 'width', 'out_quantum_3', 'bond_3', 'bond_0')
```
These 4 tensors represent a tensor by using matrix tensor ring format (or, as physicists call it, matrix product operator). What I want to do with it next is I plug into an einsum expression (`pytorch.einsum` or `opt_einsum.contract`) like this:
```
opt_einsum.contract(
core_0, core_0.names,
core_1, core_1.names,
core_2, core_2.names,
core_3, core_3.names,
["batch", *(f"out_quantum_{i}" for i in range(4)), "height", "width"], # these are dimension names of the resulting tensor
backend="torch",
optimize="optimal"
```
Right now pytorch named tensors don't allow digits in names, so instead of writing `core_0.names` in the snippet above, I have to keep names with digits somewhere else instead of as pytorch names.
Currently my workarounds due to lack of this feature are to:
1. Store names not as `tensor.names`, but separately; or
2. Implement `digits_to_words` and `words_to_digits` functions and use them to convert `quantum_55` to `quantum_fivefive` and vice versa.
## Pitch
Please allow digits without any restrictions in dimensions names. If you wonder what you should when you have a dimension `0` and a dimension `"0"`, I suggest treating them as different names.
This feature will help
* Tensor decompositions folk. These techniques are used for compression, acceleration, and regularization of neural networks.
* Tensor network folk.
## Alternatives
If you think allowing names like `"1"` may confuse people, forbid names which don't have any symbols but digits.
Another, perhaps even better, alternative is to allow arbitrary immutable hashable objects as names.
cc @zou3519 | triaged,enhancement,module: named tensor | low | Critical |
544,365,562 | TypeScript | Infer type when object coerced to string if toString explicitly defined | ## Search Terms
coercion toString computed properties
## Suggestion
Infer type when object coerced to string if `toString` explicitly defined
## Examples
Would be nice if the following would be allowed
```typescript
const action = {
type: 'foo',
toString() { return this.type }
} as const;
const reducerHandler = {
[action]: () => { } // Currently errors: A computed property name must be of type 'string', 'number', 'symbol', or 'any'.
}
```
[Playground](https://www.typescriptlang.org/play/index.html#code/MYewdgzgLgBAhsKBLcMC8MDeAoGMoCeADgKYBcMA5AGYgiUA0u+IAylAE5JgDmAFAEosMDiSgBXDmHwALJBAB0hUjAC+2VfAgxQkKAG5s2XdBEkAJuOAkOACThhzAGxvoszANoJk4ALoVBdAA+YXV1Y3AIEBcFJxB+UUtrOwdnGwF9IA)
## 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 |
544,371,009 | godot | GPU memory leak in editor | **Godot version:** 3.1.2.stable.official
**OS/device including version:** Linux 64b
**Issue description:**
GPU memory in editor is not being released after closing files in which graphical resources are used (possibly related to `AnimatedSprite`).
Example (sequential), values are GPU memory measured by `nvidia-smi`:
* in project manager: **429MiB**
* after opening project with empty scene: **618MiB**
* after opening scene with animation: **918MiB**
* after closing scene with animation and waiting 10 seconds: **918MiB** (results sometimes wary slightly, e.g. sometimes I got 961MiB, but that's still much higher compared to 618MiB)
**Steps to reproduce:**
* Create a new scene with `AnimatedSprite`
* Check GPU mem - baseline
* Create new frames in inspector
* Copy non-tiny animation files (e.g. 200 of PNGs each ~1MB) to project
* Drag and drop to animation (bottom panel)
* Check GPU mem (should be higher for Godot editor process)
* Close scene
* Check GPU mem - doesn't go back to baseline
**Minimal reproduction project:**
Well, not very minimal when judging by size :smile:. It's too big for GH, so I had to upload it somewhere else: https://mega.nz/#!sAI0VKQI!QAW42ZeR-IjHjBeCDmhLH_9Gr7wi83oy3GjFWMmK7HM | bug,topic:rendering,topic:editor,confirmed | low | Major |
544,384,513 | go | cmd/compile: optimize overhead from CPU feature detection | As investigated in https://github.com/golang/go/issues/36196, the overhead of checking for hardware FMA on every iteration of a loop causes it to slow down. @josharian's [CL 212360](https://golang.org/cl/212360), which introduces a `HasCPUFeature ` intrinsic, somewhat alleviates this overhead, but it is still non-negligble. We should look into lowering or in some cases eliminating the overhead for operations that require CPU feature detection, like population count, FMA, rounding, SSE3, etc...
One method is to hoist the check outside the loop. To quote https://github.com/golang/go/issues/15808#issuecomment-568224699
> weβd need an optimization that rewrote:
> ```
> var b bool = ...
> for ... {
> if b {
> ...
> } else {
> ...
> }
> }
> ```
> into
> ```
> var b bool = ...
> if b {
> for ... {
> ...
> }
> } else {
> for ... {
> ...
> }
> }
> ```
> Which is not always going to be an improvement.
For large loops and operations that permit > 2 implementations, the above optimization could result in inflated binaries, but it works well for smaller loops.
Another method is to set a function pointer to the preferred implementation on program initialization, so that all invocations incur an indirect function call overhead, with the benefit that the implementation wouldn't change at runtime. This would be akin to the dispatcher in GCC's [function multi-versioning](https://lwn.net/Articles/691932/).
It is worth further investigating opportunities for optimization in this space. | Performance,NeedsInvestigation,compiler/runtime | low | Major |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.