id
int64 393k
2.82B
| repo
stringclasses 68
values | title
stringlengths 1
936
| body
stringlengths 0
256k
⌀ | labels
stringlengths 2
508
| priority
stringclasses 3
values | severity
stringclasses 3
values |
|---|---|---|---|---|---|---|
587,081,697 |
flutter
|
Use system setting for font smoothing
|
Moved from https://github.com/google/flutter-desktop-embedding/issues/524. See that issue for details.
Blocked on #39690
|
c: new feature,engine,platform-linux,a: desktop,P3,team-linux,triaged-linux
|
low
|
Minor
|
587,086,502 |
neovim
|
Display a diff for a buffer that was changed outside the current editing session
|
<!-- Before reporting: search existing issues and check the FAQ. -->
- `nvim --version`: 0.4.3
- Operating system/version: Ubuntu 19.04
- Terminal name/version: gnome-terminal 3.34
- `$TERM`: xterm-256color
### Steps to reproduce using `nvim -u NORC`
1. Open a file to edit in neovim.
2. Open the same file in another editor and do some changes to it and write to the file.
3. Run `:checktime` in neovim. If autoread is set, neovim will directly reload the buffer, else the user will be prompted for reloading the buffer.
### Actual behaviour
Every time the user is prompted for reloading the buffer, there is no way of knowing what changes occured in the buffer.
### Expected behaviour
After implementing this feature, a diff between the buffer opened in neovim and the changed file would be displayed along with a prompt for whether the user wants to reload a buffer or not.
I'd like to implement this feature as a part of the GSoC project for improving autoread.
|
enhancement
|
low
|
Major
|
587,168,612 |
nvm
|
$(nvm_sanitize_path "${NPM_CONFIG_PREFIX}");: command not found
|
Getting this in my shell when I open a new bash tab:
```
$(nvm_sanitize_path "${NPM_CONFIG_PREFIX}");: command not found
```
in my ~/.bashrc I have:
```
export NVM_DIR="$HOME/.nvm"
[ -s "$NVM_DIR/nvm.sh" ] && \. "$NVM_DIR/nvm.sh" # This loads nvm
[ -s "$NVM_DIR/bash_completion" ] && \. "$NVM_DIR/bash_completion" # This loads nvm bash_completion
```
i remember seeing the NPM_CONFIG_PREFIX stuff somewhere but don't see it now - I am on ubuntu 18.04
---------------------------
#### Operating system and version: ubuntu 18.04
#### `nvm debug` output:
<details>
<!-- do not delete the following blank line -->
```sh
nvm --version: v0.35.3
$SHELL: /bin/bash
$SHLVL: 1
${HOME}: /home/oleg
${NVM_DIR}: '${HOME}/.nvm'
${PATH}: ${HOME}/Android/Sdk/tools:${HOME}/Android/Sdk/tools/bin:${HOME}/.opam/system/bin:/usr/local/opt/bison/bin:/usr/local/opt/openldap/sbin:/usr/local/opt/openldap/bin:/opt/local/bin:/opt/local/sbin:/usr/local/opt/openldap/sbin:/usr/local/opt/openldap/bin:${HOME}/.oresoftware/bin:${HOME}/.local/bin:${HOME}/.oresoftware/bin:${HOME}/.local/bin:${NVM_DIR}/versions/node/v12.16.1/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:${HOME}/.dotnet/tools:/usr/local/go/bin:/usr/local/go/bin:${HOME}/.rvm/bin:${HOME}/.local/bin:${HOME}/.rvm/bin
$PREFIX: ''
${NPM_CONFIG_PREFIX}: ''
$NVM_NODEJS_ORG_MIRROR: ''
$NVM_IOJS_ORG_MIRROR: ''
shell version: 'GNU bash, version 4.4.20(1)-release (x86_64-pc-linux-gnu)'
uname -a: 'Linux 4.15.0-91-generic #92-Ubuntu SMP Fri Feb 28 11:09:48 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux'
OS version: Ubuntu 18.04.4 LTS
curl: /usr/bin/curl, curl 7.58.0 (x86_64-pc-linux-gnu) libcurl/7.58.0 OpenSSL/1.1.1 zlib/1.2.11 libidn2/2.0.4 libpsl/0.19.1 (+libidn2/2.0.4) nghttp2/1.30.0 librtmp/2.3
wget: /usr/bin/wget, GNU Wget 1.19.4 built on linux-gnu.
git: /usr/bin/git, git version 2.17.1
grep: /bin/grep (grep --color=auto), grep (GNU grep) 3.1
awk: /usr/bin/awk, awk: not an option: --version
sed: /bin/sed, sed (GNU sed) 4.4
cut: /usr/bin/cut, cut (GNU coreutils) 8.28
basename: /usr/bin/basename, basename (GNU coreutils) 8.28
rm: /bin/rm, rm (GNU coreutils) 8.28
mkdir: /bin/mkdir, mkdir (GNU coreutils) 8.28
xargs: /usr/bin/xargs, xargs (GNU findutils) 4.7.0-git
nvm current: v12.16.1
which node: ${NVM_DIR}/versions/node/v12.16.1/bin/node
which iojs:
which npm: ${NVM_DIR}/versions/node/v12.16.1/bin/npm
npm config get prefix: ${NVM_DIR}/versions/node/v12.16.1
npm root -g: ${NVM_DIR}/versions/node/v12.16.1/lib/node_modules
```
</details>
#### `nvm ls` output:
<details>
<!-- do not delete the following blank line -->
```sh
oleg@xps:~$ nvm ls
v6.17.1
v7.10.1
v8.15.1
v9.11.2
v10.7.0
v10.19.0
v11.3.0
v11.11.0
v11.15.0
v12.6.0
v12.9.0
-> v12.16.1
v13.9.0
v13.11.0
system
default -> 12 (-> v12.16.1)
node -> stable (-> v13.11.0) (default)
stable -> 13.11 (-> v13.11.0) (default)
iojs -> N/A (default)
unstable -> N/A (default)
lts/* -> lts/erbium (-> v12.16.1)
lts/argon -> v4.9.1 (-> N/A)
lts/boron -> v6.17.1
lts/carbon -> v8.17.0 (-> N/A)
lts/dubnium -> v10.19.0
lts/erbium -> v12.16.1
```
</details>
#### How did you install `nvm`?
The curl script
#### Is there anything in any of your profile files that modifies the `PATH`?
Yes definitely
|
needs followup
|
medium
|
Critical
|
587,199,554 |
pytorch
|
Method/constructor which takes as input angle and magnitude and returns a complex tensor
|
cc @ezyang @anjali411 @dylanbespalko
|
triaged,module: complex
|
low
|
Major
|
587,203,546 |
neovim
|
non-string Lua errors in vim.loop callbacks cause segfault
|
<!-- Before reporting: search existing issues and check the FAQ. -->
- `nvim --version`: NVIM v0.4.3
- Operating system/version: Linux 5.5.10
### Steps to reproduce using `nvim -u NORC`
```
nvim -u NORC -c 'lua vim.loop.new_idle():start(function() error() end)'
```
Using `new_idle` to get a function to run in loop callback but also happens in other kinds of callbacks like pipe reads.
### Actual behaviour
Neovim terminated by SIGSEGV.
### Expected behaviour
Error logged as an error message?
|
has:repro,bug-crash,lua
|
low
|
Critical
|
587,213,597 |
node
|
Retry on failed ICU data load, ignoring NODE_ICU_DATA or --icu-data-dir
|
Related to #30825 (I had the idea during its discussion) and, from the icebox, #3460
**Is your feature request related to a problem? Please describe.**
Currently, Node with an invalid/missing ICU data directory ( set with NODE_ICU_DATA or --icu-data-dir ) just fails.
```shell
$ nvm i --lts
Installing latest LTS version.
v12.16.1 is already installed.
$ node --icu-data-dir=/tmp
node: could not initialize ICU (check NODE_ICU_DATA or --icu-data-dir parameters)
```
**Describe the solution you'd like**
It might be possible to fall back *as if* the `--icu-data-dir` was not specified (or the default data dir in #30825 was not present.)
As early as that error message is generated, it might be possible to call `u_cleanup()` to unload ICU and then reinitialize it using the baked-in data or other defaults.
This could be done completely within `InitializeICUDirectory()`, although it might be advantageous to allow programmatic detection of the fact that this fallback happened. Ideas there:
1. expose `const char *u_getDataDirectory()` via the ICU process binding.
2. provide the eventual path string via some kind of variable
3. ~print out a warning~ (not a good idea)
4. some other flag
5. do nothing - this would probably be OK if there isn't a strong use case (other than perhaps test cases)
FYI @sam-github @sgallagher @nodejs/i18n-api
**Describe alternatives you've considered**
There's really no alternative to restarting the node process at this point.
|
i18n-api,cli
|
low
|
Critical
|
587,231,913 |
flutter
|
Consider converting system channels to standard codec
|
Currently many of the Flutter framework's system channels use the JSON codec, presumably purely for legacy reasons. This is not ideal for a number of reasons:
- Not all languages/platforms have a standard JSON library, so some embeddings (e.g., our Windows and Linux embeddings) need to use/link/ship a third-party JSON library just to talk to critical system channels.
- In fact, the Flutter engine is shipping RapidJSON purely for this reason, I believe.
- It essentially forces every new platform to implement two codecs: standard for compatibility with most (~all? I don't know if we have data on this), and JSON for compatibility with the framework.
- It's less efficient than the standard method codec (and the general advantages of JSON over binary don't really apply to ephemeral, internal messages)
This would be a breaking change for all embedding implementations. Some options:
- Just make it a breaking change. It should be very easy for people with custom embeddings to fix (since they will almost certainly have implemented the standard codec anyway)
- Add parallel channels with new names, and an opt-in method for an embeddings to say they want the new ones (set at engine startup).
/cc @Hixie for thoughts.
|
framework,c: proposal,P2,team-framework,triaged-framework
|
low
|
Major
|
587,291,021 |
youtube-dl
|
Support for Onlyfans
|
## Checklist
- [x] I'm reporting a new site support request
- [x] I've verified that I'm running youtube-dl version **2020.03.24**
- [x] I've checked that all provided URLs are alive and playable in a browser
- [x] I've checked that none of provided URLs violate any copyrights
- [x] I've searched the bugtracker for similar site support requests including closed ones
## Example URLs
-Free post https://onlyfans.com/16391013/killercleavage-free
-Free post https://onlyfans.com/15584274/hoodedhofree
-Free post https://onlyfans.com/13634727/lilianlacefree
-Paid post https://onlyfans.com/16352300/entina_cat
-Free post https://onlyfans.com/16384958/silvermoonfree
## Description
I think it would be very helpful to add this to the page list. Especially for mass backups of accounts disconnected from the internet. Always protecting content using credentials.
|
site-support-request
|
low
|
Critical
|
587,341,337 |
ant-design
|
Popconfirm does not work on Dropdown item
|
- [x] I have searched the [issues](https://github.com/ant-design/ant-design/issues) of this repository and believe that this is not a duplicate.
### Reproduction link
[](https://codesandbox.io/s/antd-bug-popconfirm-in-dropdown-tiw9l)
### Steps to reproduce
1. Put a Popconfirm around a Dropdown item.
2. Click the Dropdown item
### What is expected?
Popconfirm appears
### What is actually happening?
Popconfirm appears and immediately disappears, or never shows at all.
| Environment | Info |
|---|---|
| antd | 4.0.4 |
| React | 16.13.1 |
| System | MacOS |
| Browser | Chrome 80.0.3987.132 |
---
This worked in 3.x .
<!-- generated by ant-design-issue-helper. DO NOT REMOVE -->
|
🐛 Bug,Inactive,4.x,3.x
|
medium
|
Critical
|
587,367,651 |
vue
|
Object with prototype of `null` cannot be displayed using text interpolation
|
### Version
2.6.11
### Reproduction link
[https://codepen.io/hezedu/pen/GRJYXGR](https://codepen.io/hezedu/pen/GRJYXGR)
### Steps to reproduce
vue: 2.6.11
https://github.com/vuejs/vue/blob/a59e05c2ffe7d10dc55782baa41cb2c1cd605862/dist/vue.runtime.js#L96-L102
```js
toString(Object.create(null);
```
Will crash:
TypeError: Cannot convert object to primitive value
### What is expected?
TypeError
### What is actually happening?
error
<!-- generated by vue-issues. DO NOT REMOVE -->
|
improvement
|
medium
|
Critical
|
587,378,377 |
TypeScript
|
Module plugin template is misleading
|
**Expected behavior:**
I followed the template exactly and it can't do what it's claiming you can do. I would expect from the template that I could extend an existing module with a new function.
https://www.typescriptlang.org/docs/handbook/declaration-files/templates/module-plugin-d-ts.html
**Actual behavior:**
I am unable to extend the existing module with a new function.
https://github.com/MCKRUZ/TS-Plugin-HW
|
Docs
|
low
|
Minor
|
587,413,838 |
PowerToys
|
Source code line number and highlighter
|
Having written a number of articles and books, making code listings stand out and be able to reference in text takes some work. The following website: "[http://www.planetb.ca/syntax-highlight-word](http://www.planetb.ca/syntax-highlight-word)" takes a code snippet copied from VS, and after selecting the language, will add line numbers and alternate gray and write bars for each line of code. The result can be copied and pasted into Word, but further work has to be performed to address formatting errors for longer code snippets and better font selection.
A separate tool or a tool in Visual Studio/VSCode that lets one take code snippets and outputs:
1. Puts in the line numbers with alternate gray and write bars for each line of code.
2. Maintain the code text color and style as it is displayed in Visual Studio.
3. Maintain all indents and code formatting
4. The user can also select the font or change text to one color.
5. The result can be copied into Microsoft Word.
|
Idea-New PowerToy
|
low
|
Critical
|
587,423,016 |
electron
|
Cookie 'before-added' event
|
<!-- As an open source project with a dedicated but small maintainer team, it can sometimes take a long time for issues to be addressed so please be patient and we will get back to you as soon as we can.
-->
### Preflight Checklist
<!-- Please ensure you've completed the following steps by replacing [ ] with [x]-->
* [x] I have read the [Contributing Guidelines](https://github.com/electron/electron/blob/master/CONTRIBUTING.md) for this project.
* [x] I agree to follow the [Code of Conduct](https://github.com/electron/electron/blob/master/CODE_OF_CONDUCT.md) that this project adheres to.
* [x] I have searched the issue tracker for a feature request that matches the one I want to file, without success.
### Problem Description
<!-- Is your feature request related to a problem? Please add a clear and concise description of what the problem is. -->
I'm working on an node module to add much more privacy for electron applications and I wanted to ask about adding a new event to the Cookies API. For web privacy and tracking-protection, one of the best measures that can be taken is blocking third party cookies. The current ['changed'](https://www.electronjs.org/docs/api/cookies?q=before#event-changed) event only gets called ***after*** the cookie gets added. Ideally, there would be an event being called when a website is trying to add a cookie and the Electron event would include a callback deciding whether to go through with the cookie addition or to cancel the request and not add the cookie.
### Proposed Solution
<!-- Describe the solution you'd like in a clear and concise manner -->
Event for the Electron [Cookies API](https://www.electronjs.org/docs/api/cookies) called **before** a cookie is added:
```js
cookies.on('before-added', (cookie, callback) => {
callback(true); // Keep cookie
callback(false); // Remove cookie
});
```
The callback should take in one boolean parameter determining if the cookie should be added as intended or removed.
### Alternatives Considered
<!-- A clear and concise description of any alternative solutions or features you've considered. -->
One possible, although not as neat, alternative would be adding a callback to the existing ['changed'](https://www.electronjs.org/docs/api/cookies?q=before#event-changed) event so that the change could be cancelled. Yet this method would apply for all forms of cookie modification not just when a cookie is added.
### Additional Information
<!-- Add any other context about the problem here. -->
N/A
|
enhancement :sparkles:
|
medium
|
Major
|
587,439,499 |
vue-element-admin
|
edge 访问浏览器报错
|
<!--
注意：为更好的解决你的问题，请参考模板提供完整信息，准确描述问题，信息不全的 issue 将被关闭。
Note: In order to better solve your problem, please refer to the template to provide complete information, accurately describe the problem, and the incomplete information issue will be closed.
-->
## Bug report 用edge浏览器打开页面，点击刷新，侧边栏不能点击
#### Steps to reproduce（问题复现步骤）
<!--
1. https://panjiachen.github.io/vue-element-admin/#/permission/role（登陆就有问题，如果没有，第二步）
2. 点击浏览器刷新按钮
3. 出现问题，侧边栏不能点击报错
4，打开控制台，出现错误提示
TypeError: Unable to get property 'parent' of undefined or null reference render
-->
####
####
初步排查肯能是SiderBarItem.vue组件问题
<!--
Please only use Codepen, JSFiddle, CodeSandbox or a github repo
-->
#### Other relevant information（格外信息）
只有在edge不打开控制台的情况下出错，打开控制台不报错
|
polyfill
|
low
|
Critical
|
587,444,200 |
pytorch
|
NCCL version upgrade for PyTorch
|
## 🐛 Bug
<!-- A clear and concise description of what the bug is. -->
Build PyTorch from source code would fail on some old Linux release.
This issue has been described here: https://github.com/NVIDIA/nccl/issues/244, and Nvidia folks had it fixed (https://github.com/NVIDIA/nccl/commit/7f2b337e703d73ed369937c9996e1f3d5f664ad0). However, that change has not been included in the PyTorch source code, so if people need to build PyTorch from source code on some old linux system, it would still fail for the same problem.
The current nccl version in the latest PyTorch release is still 2.4.8, i think it can be upgraded to 2.5.6 now.
I can see that there is another open ticket for the same issue: https://github.com/pytorch/pytorch/issues/29093
## To Reproduce
Steps to reproduce the behavior:
1. Checkout PyTorch source code.
2. Build on a linux system with kernel that is below 3.9 (This is when SO_REUSEPORT was introduced).
## Expected behavior
Build should succeed on Linux system that is below 3.9.
```
- PyTorch Version (e.g., 1.0): 1.4 master
- OS (e.g., Linux): Linux (Below 3.9)
- How you installed PyTorch (`conda`, `pip`, source): source
- Build command you used (if compiling from source): python setup.py develop
- Python version: 3.6
- CUDA/cuDNN version: irrelevant
- GPU models and configuration: irrelevant
- Any other relevant information: No
```
## Additional context
No
|
module: build,triaged
|
low
|
Critical
|
587,504,599 |
flutter
|
[Proposal] Icon support Linux desktop application
|
## Use case
The default Linux application icon looks ugly...🥴 Flutter is way worthy than that... 😎
## Proposal
This looks much better (photo-shopped)
|
c: new feature,tool,a: quality,platform-linux,c: proposal,a: desktop,a: build,P3,team-linux,triaged-linux
|
high
|
Critical
|
587,524,896 |
TypeScript
|
Incorrectly typed argument variant
|
**TypeScript Version:** 3.8.3
**Search Terms:**
Parameters, Array, Overloading, Type inference
**Code**
```ts
type VariantA = [(err: Error, data: string) => any]
type VariantB = [(data: string) => any, (err: Error) => any]
type Args = VariantA | VariantB
function fn (...rest: Args): void { /* do something */ }
fn((err, data) => {
err.stack // Correct: Typed as Error
data.charAt(0) // Correct: Typed as string
})
fn((data) => {
data // --> Error: Mistyped as any, should be string <--
}, (err) => {
err.stack // Correct: Typed as Error
})
```
**Expected behavior:**
If two functions are passed in (_VariantB_), the first paramter should be typed as `data: string`.
**Actual behavior:**
If two functions are passed in, the first function's parameter is typed as `any`.
|
Bug
|
low
|
Critical
|
587,539,005 |
pytorch
|
How to support single-process-multiple-devices in DistributedDataParallel other than CUDA device
|
Hi,
I am investigating to extend the DistributedDataParallel to other accelerator devices than CUDA devices.
Not only to support single-process-single-device but also to support the single-process-multiple-devices and multple-processes-multiple-devices.
There are a lot of CUDA dependency in the DistributedDataParallel.
My question is:
1. How to override CUDA logical dependency and dispatch the gather and scatter (and other APIs used) to the c10d backend without modifying the distributed.py ? [https://github.com/pytorch/pytorch/blob/master/torch/nn/parallel/distributed.py](url)
cc @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @rohan-varma @gqchen @aazzolini @osalpekar @jiayisuse @agolynski @SciPioneer @H-Huang @mrzzd @cbalioglu @gcramer23
|
oncall: distributed
|
low
|
Minor
|
587,566,986 |
angular
|
With ivy, injecting a parent directive with the @host annotation fails in nested structures
|
# 🐞 bug report
### Is this a regression?
Yes, in angular 8 this works fine. But after an update to angular 9 it stops working. If I opt-out ivy in the tsconfig.app.json ("enableIvy": false) it is working again.
### Description
The problem occurs if a parent directive is injected with the @Host annotation in nested structures. So the first injection works fine. But all injections in deeper nested structers fails and null is injected.
In the root component a ui property is defined which contains nested child elements.
```
@Component({
selector: 'app-root',
template: `
<app-container [element]="ui"></app-container>
`
})
export class AppComponent {
ui: ElementDef = {
name: 'container1', child: {
name: 'container2', child: {
name: 'container3', child: {
name: 'container4'
}
}
}
};
}
```
The elements are rendered by the container component. Each rendered element is embedded in a div with a directive "appContainerWrapper" attached.
The parent directive of a container is injected in the constructor with the @Host annotation.
```
@Component({
selector: 'app-container',
template: `
<div appContainerWrapper *ngIf="element">
<p>Container: {{element.name}}</p>
<app-container [element]="element.child"></app-container>
</div>
`
})
export class ContainerComponent implements OnInit {
@Input() element: ElementDef;
constructor(@Optional() @Host() private directive: ContainerWrapperDirective) {}
[...]
ngOnInit(): void {
if (this.directive !== null) {
this.directive.update(color);
} else {
// in nested containers this.directive is null if ivy is enabled :(
}
}
}
```
The directive should set the background color of the parent element if `update `is called from the child element.
```
@Directive({
selector: '[appContainerWrapper]'
})
export class ContainerWrapperDirective {
constructor(private containerWrapperRef: ElementRef, private renderer: Renderer2) {}
update(color: string): void {
const nativeElement = this.containerWrapperRef.nativeElement;
this.renderer.setStyle(nativeElement, 'background-color', color);
}
}
```
## 🔬 Minimal Reproduction
The problem cannot be reproduced in stackblitz. I therefor created a sample application on [Github](https://github.com/martinbu/ivy-directive-issue)
The repository contains two branches:
1. **master:** Angular 9 with ivy enabled --> the parent directive is only injected in the first container where a directive is available. In all further containers the injected directive is null. Therefor only one "background is set".
2. **opt-out-ivy:** same as the master branch but set "enableIvy" in tsconfig.app.json to false --> everything works as expected as with angular 8. In each container the parent directive is injected.
## 🌍 Your Environment
**Angular Version:**
<pre><code>
Angular CLI: 9.0.7
Node: 12.16.1
OS: win32 x64
Angular: 9.0.7
... animations, cli, common, compiler, compiler-cli, core, forms
... language-service, platform-browser, platform-browser-dynamic
... router
Ivy Workspace: Yes
Package Version
-----------------------------------------------------------
@angular-devkit/architect 0.900.7
@angular-devkit/build-angular 0.900.7
@angular-devkit/build-optimizer 0.900.7
@angular-devkit/build-webpack 0.900.7
@angular-devkit/core 9.0.7
@angular-devkit/schematics 9.0.7
@ngtools/webpack 9.0.7
@schematics/angular 9.0.7
@schematics/update 0.900.7
rxjs 6.5.4
typescript 3.7.5
webpack 4.41.2
</code></pre>
|
type: bug/fix,freq2: medium,area: core,state: confirmed,core: di,P3
|
low
|
Critical
|
587,575,696 |
godot
|
Image stretch when Window Resize is enabled
|
**Godot version:**
3.2.1
**OS/device including version:**
macOS Catalina 10.15.4 / MacBookPro11,1
**Issue description:**
If Window Resize is enabled, the image gets stretched horizontally.
It doesn't matter is Fullscreen is enabled or not. The problem occurs in both situations.
**Note:**
- I set the Environment color to white, but when the image is stretched the borders are black, so it is not showing the environment color. Something else must be happening.
- I am running this on an external monitor that has 1920x1080 resolution (if this has any relevance)
- Here is a complete video that shows the problem: https://drive.google.com/open?id=1rfRweWs5Yo0MgLJRYsfnOl7Uj8REgdKI
**Steps to reproduce:**
Run the attached Godot project with and without Window Resize enabled.
**Minimal reproduction project:**
[Godot_Window_Bug.zip](https://github.com/godotengine/godot/files/4380118/Godot_Window_Bug.zip)
|
platform:macos,topic:porting
|
medium
|
Critical
|
587,616,315 |
go
|
proxy.golang.org: long cache TTL causes confusing UX for 'go get <module>@<branchname>'
|
<!--
Please answer these questions before submitting your issue. Thanks!
For questions please use one of our forums: https://github.com/golang/go/wiki/Questions
-->
### What version of Go are you using (`go version`)?
<pre>
$ go version
go version go1.14.1 windows/amd64
</pre>
### Does this issue reproduce with the latest release?
Yes
### What operating system and processor architecture are you using (`go env`)?
<details><summary><code>go env</code> Output</summary><br><pre>
$ go env
set GO111MODULE=
set GOARCH=amd64
set GOBIN=
set GOCACHE=C:\Users\josh\AppData\Local\go-build
set GOENV=C:\Users\josh\AppData\Roaming\go\env
set GOEXE=.exe
set GOFLAGS=
set GOHOSTARCH=amd64
set GOHOSTOS=windows
set GOINSECURE=
set GONOPROXY=
set GONOSUMDB=
set GOOS=windows
set GOPATH=E:\go
set GOPRIVATE=
set GOPROXY=https://proxy.golang.org,direct
set GOROOT=c:\go
set GOSUMDB=sum.golang.org
set GOTMPDIR=
set GOTOOLDIR=c:\go\pkg\tool\windows_amd64
set GCCGO=gccgo
set AR=ar
set CC=gcc
set CXX=g++
set CGO_ENABLED=1
set GOMOD=E:\src\rainway-metapod\go.mod
set CGO_CFLAGS=-g -O2
set CGO_CPPFLAGS=
set CGO_CXXFLAGS=-g -O2
set CGO_FFLAGS=-g -O2
set CGO_LDFLAGS=-g -O2
set PKG_CONFIG=pkg-config
set GOGCCFLAGS=-m64 -mthreads -fmessage-length=0 -fdebug-prefix-map=C:\Users\josh\AppData\Local\Temp\go-build763362010=/
tmp/go-build -gno-record-gcc-switches
</pre></details>
### What did you do?
`go get github.com/RainwayApp/metapod@modulize` and then attempt to import and use metapod in a program
<!--
If possible, provide a recipe for reproducing the error.
A complete runnable program is good.
A link on play.golang.org is best.
-->
```go
package main
import "github.com/RainwayApp/metapod"
func main() {
metapod.Create()
}
```
### What did you expect to see?
No errors! woo!
### What did you see instead?
```module github.com/RainwayApp/metapod@latest found (v0.0.0-20190515153323-337878358c65), but does not contain package github.com/RainwayApp/metapod```
Whats interesting is that the downloaded module is devoid of go files which I dont really get (this is even after trying to go get the module again after deleting it from here).
I got around this by creating a new branch (modulize2) which has all the files - however im not sure how its possible for it to end up in this state in the first place so just thought id bring it up...
|
NeedsInvestigation,modules,proxy.golang.org
|
low
|
Critical
|
587,639,522 |
TypeScript
|
Set as executable files that are intended to be binaries
|
<!-- 🚨 STOP 🚨 𝗦𝗧𝗢𝗣 🚨 𝑺𝑻𝑶𝑷 🚨
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 -->
- `mode`
- `binary`
- `executable`
- `package.json bin`
## Suggestion
<!-- A summary of what you'd like to see added or changed -->
Currently, the TypeScript compiler always generates files with the `644` mode. But some files are meant to be executable: they may start with `#!/usr/bin/env node`; they may be marked as `bin` in `package.json`; they may have an executable mode set on the `.ts` source; and so forth.
It’d be great if the TypeScript compiler would take any of the hints I mentioned above, figure that the file should be executable, and create the `.js` as so.
At least the TypeScript compiler doesn’t **change** the mode of existing `.js` files when overwriting them, so I can `chmod` once, and as long as never delete the file, it’ll have the right mode.
## Use Cases
<!--
What do you want to use this for?
What shortcomings exist with current approaches?
-->
This is mostly for development and running the binary from the command-line directly. Once you install the package, npm puts the binary in the `node_modules/.bin` folder, `npx` picks it up, and all is good.
## Examples
<!-- Show how this would be used and what the behavior would be -->
You’re working on a project that should provide a binary: `src/my-binary.ts`. You do one of the things I mentioned to indicate that the file should be executable. You run `tsc`, and the generated file at `lib/my-binary.js` is marked as `+x`. Then you can call the binary with `lib/my-binary.js` from the command-line and it works.
## 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
|
medium
|
Critical
|
587,659,114 |
godot
|
OS.execute trouble with shell evaluation ("export PATH=test && echo $PATH" doesn't work)
|
**Godot version:**
3.2.1
**OS/device including version:**
macos
**Issue description:**
OS.execute("/bin/bash", ['-c',"export PATH=test && echo $PATH"], true, output,true)
print(output)
print result
[/usr/bin:/bin:/usr/sbin:/sbin
]
don't work
**Steps to reproduce:**
**Minimal reproduction project:**
|
bug,platform:linuxbsd,platform:macos,topic:porting,confirmed
|
low
|
Major
|
587,659,752 |
godot
|
[TextureRect] rect_scale when changing texture
|
**Godot version:**
3.2.1.stable
**OS/device including version:**
Windows 10 Pro
**Issue description:**
(Atleast when using animations) When changing a texture (from "null") on the TextureRect it briefly resets the rect_scale to 1,1.
**Steps to reproduce:**
- Create a control and put a texture rect inside it
- Create an animation state machine
- Create a show (start animation), base and hide animation
- At the start of the show animation "update" the texture (set to some texture) and set rect_scale to (0, 0) and at the end move it to (1,1)
- At the end of the hide animation "update" the texture (make it null) and set rect_scale to (1,1) and at the end move it to (0,0)
- Make controls so you can start the show animation and travel to the hide animation
- Start the show animation, travel to the hide (wait for it to finish) and then finally start the show animation, you can see that the rect_scale changes to (1,1) when changing the texture and then after a brief moment goes into it's rect_scale on which it should be (by the animation)
**Minimal reproduction project:**
https://drive.google.com/file/d/113X-mEWHL5ad1khAM6G4VG7CpapT7aOP/view?usp=sharing
|
bug,topic:rendering
|
low
|
Minor
|
587,677,110 |
go
|
cmd/compile: consider extending '-spectre' option to other architectures
|
@rsc recently added a `-spectre` flag to the compiler and implemented a couple of speculative execution exploit mitigations on amd64. These involved using conditional moves to prevent out of bounds indexed memory accesses (`-spectre=index` added in https://golang.org/cl/222660) and making use of call stack optimizations to prevent indirect function calls being speculatively executed (`-spectre=ret` added in https://golang.org/cl/222661).
My understanding is that these mitigations could be important when the Go program being compiled is something like a hypervisor or operating system that has access to data or functionality that guest programs should be restricted from accessing. I'm still in the process of getting my head round these changes and how they might apply to other CPU architectures but my understanding so far is that, to be totally safe, we should apply similar mitigations on other platforms unless there are hardware mitigations in place.
The new flag also has potential for cross-platform incompatibilities since currently, for example, `-spectre=index` will cause the compiler to fail on non-amd64 platforms. It would be nice if these flags 'just worked' on other platforms.
I personally am responsible for maintaining the s390x port. I think it would be straightforward to implement equivalent mitigations on s390x:
1) index: use conditional moves in exactly the same way as the amd64 backend
2) ret: use EX{,RL} (execute) instructions to indirectly call branch instructions (more info here:
https://www.phoronix.com/scan.php?page=news_item&px=S390-Expoline-Linux-4.16)
Does anyone have any thoughts on what we should do for other architectures, if anything?
|
Security,NeedsInvestigation,compiler/runtime
|
low
|
Major
|
587,719,279 |
flutter
|
SliverPersistentHeader initial state
|
## Proposal
`SliverPersistentHeader` / `SliverPersistentHeaderDelegate`
This header has two main status - minExtent (collapsed) and maxExtent(expanded)
Initially, this header is in expanded status.
I think this header can be collapsed initally
Just like SearchBar on android.
It can be expanded or iconified (collapsed) initally.
Thank you for your great support.
Looking forward hearing from you.
|
c: new feature,framework,f: scrolling,would be a good package,P3,team-framework,triaged-framework
|
low
|
Minor
|
587,749,382 |
node
|
Describe alternatives to require.extensions
|
https://nodejs.org/api/modules.html#modules_require_extensions says `require.extensions` is deprecated.
Per https://github.com/TypeStrong/ts-node/issues/641 and https://github.com/TypeStrong/ts-node/issues/989 there are no alternatives. They reference further the documentation of @babel/node, @babel/register, and coffeescript.
• Either re-introduce require.extensions or describe more precisely the alternatiles that should be used.
|
module,deprecations
|
low
|
Major
|
587,756,776 |
material-ui
|
[docs] Allow providing a custom theme and preview it live
|
<!-- Provide a general summary of the feature in the Title above -->
We're going to make a lot of modifications to our theme soon, and I was thinking that it'd be really cool if I could enter a custom theme (JS and all) in the material-ui docs, through an inline editor, and navigate through all the docs with that theme.
Benefits:
* Previewing and testing custom theming options is a breeze
* Remove all "customization" demos from the components, and add a button that loads a pre-defined global theme, which then modifies all demos in real time: "Try Button with different themes"
* Potentially improve the interactivity of the customization docs, making theming more aproachable for every one
This is obviously a huge effort, but as you can already change the theme colors across the website, maybe there is some infrastructure in place already to do this.
|
priority: important,design,scope: docs-infra
|
low
|
Major
|
587,766,574 |
kubernetes
|
Stop applying the beta.kubernetes.io/os and beta.kubernetes.io/arch
|
It seems that according to plan, kubenrnetes 1.18 will remove support for the `beta.kubernetes.io/os` and` beta.kubernetes.io/arch` labels. Will we still do this work?
---
https://kubernetes.io/docs/reference/kubernetes-api/labels-annotations-taints/
|
area/kubelet,sig/node,kind/feature,priority/important-longterm,lifecycle/frozen,triage/accepted
|
medium
|
Critical
|
587,783,734 |
pytorch
|
runtime error when loding heavy dataset
|
## 🐛 Bug
i used torch.save to save two very big tensors one being the training dataset and the other one being the test dataset (separated by torch.utils.data.random_split)
the original tensor was of size (4803, 354, 4000)
when i do that it’s very slow to save (wich is understandable), both files have the same size (wich is weird since the testing set is only 20% of the original while the training set is 80% of the original)
and when i try to load the dataset using torch.load it gives me this error (note : i am on a gpu kernel but i get the same error on a cpu kernel)
RuntimeError Traceback (most recent call last)
in
----> 1 training_set = torch.load(‘trainaction.pt’)
2 testing_set = torch.load(‘testaction.pt’)
3
4 batch_size = 200
5 DATA_PATH = ‘./data’
~.conda\envs\py36\lib\site-packages\torch\serialization.py in load(f, map_location, pickle_module, **pickle_load_args)
527 with _open_zipfile_reader(f) as opened_zipfile:
528 return _load(opened_zipfile, map_location, pickle_module, **pickle_load_args)
–> 529 return _legacy_load(opened_file, map_location, pickle_module, **pickle_load_args)
530
531
~.conda\envs\py36\lib\site-packages\torch\serialization.py in _legacy_load(f, map_location, pickle_module, **pickle_load_args)
700 unpickler = pickle_module.Unpickler(f, **pickle_load_args)
701 unpickler.persistent_load = persistent_load
–> 702 result = unpickler.load()
703
704 deserialized_storage_keys = pickle_module.load(f, **pickle_load_args)
~.conda\envs\py36\lib\site-packages\torch\serialization.py in persistent_load(saved_id)
661 location = _maybe_decode_ascii(location)
662 if root_key not in deserialized_objects:
–> 663 obj = data_type(size)
664 obj._torch_load_uninitialized = True
665 deserialized_objects[root_key] = restore_location(obj, location)
RuntimeError: [enforce fail at …\c10\core\CPUAllocator.cpp:47] ((ptrdiff_t)nbytes) >= 0. alloc_cpu() seems to have been called with negative number: 18446744066554005248
## Environment
PyTorch version: 1.4.0
Is debug build: No
CUDA used to build PyTorch: 10.1
OS: Microsoft Windows 10 Home
GCC version: Could not collect
CMake version: Could not collect
Python version: 3.6
Is CUDA available: Yes
CUDA runtime version: 10.2.89
GPU models and configuration: GPU 0: GeForce GTX 1070
Nvidia driver version: 441.22
cuDNN version: Could not collect
Versions of relevant libraries:
[pip] numpy==1.18.1
[pip] numpydoc==0.9.2
[pip] torch==1.4.0
[pip] torchvision==0.5.0
[conda] _pytorch_select 1.1.0 cpu
[conda] blas 1.0 mkl
[conda] mkl 2020.0 166
[conda] mkl-service 2.3.0 py36hb782905_0
[conda] mkl_fft 1.0.15 py36h14836fe_0
[conda] mkl_random 1.1.0 py36h675688f_0
[conda] pytorch 1.4.0 py3.6_cuda101_cudnn7_0 pytorch
[conda] torchvision 0.5.0 py36_cu101 pytorch
|
needs reproduction,triaged
|
low
|
Critical
|
587,840,615 |
godot
|
Extra suffix is not appended at the end of a binary name in some cases
|
**Godot version:**
3.0, 4.0
**OS/device including version:**
Windows 10, likely any other OS.
**Steps to reproduce:**
## With mono module enabled:
```
scons platform=windows target=release_debug bits=64 module_mono_enabled=yes mono_glue=no extra_suffix=mygame
```
Results in binary name such as:
`godot.windows.opt.tools.64.mygame.mono.exe`
Should be like:
`godot.windows.opt.tools.64.mono.mygame.exe`
## With sanitizers:
```
scons platform=linuxbsd target=debug tools=yes use_asan=yes use_lsan=yes extra_suffix=mygame
```
Results in binary name such as:
`godot.linuxbsd.tools.64.mygames` (notice appended `s`)
Should be like:
`godot.linuxbsd.tools.64.s.mygame`
---
Should likely be handled by refactor as suggested in #24030.
|
bug,topic:buildsystem
|
low
|
Critical
|
587,869,569 |
TypeScript
|
New error: Type 'string' is not assignable to type 'string & ...'
|
https://github.com/material-components/material-components-web-components/blob/b1862998d80f4b0b917c817e6efde15a2504d8da/packages/linear-progress/src/mwc-linear-progress-base.ts#L109
```
packages/linear-progress/src/mwc-linear-progress-base.ts:109:9 - error TS2322: Type 'string' is not assignable to type 'string & ((property: string) => string) & ((property: string) => string) & ((index: number) => string) & ((property: string) => string) & ((property: string, value: string | null, priority?: string | undefined) => void)'.
Type 'string' is not assignable to type '(property: string) => string'.
109 this.bufferElement
~~~~~~~~~~~~~~~~~~
110 .style[property as Exclude<keyof CSSStyleDeclaration, 'length'|'parentRule'>] =
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
```
https://github.com/material-components/material-components-web-components/blob/b1862998d80f4b0b917c817e6efde15a2504d8da/packages/linear-progress/src/mwc-linear-progress-base.ts#L119
```
packages/linear-progress/src/mwc-linear-progress-base.ts:119:9 - error TS2322: Type 'string' is not assignable to type 'string & ((property: string) => string) & ((property: string) => string) & ((index: number) => string) & ((property: string) => string) & ((property: string, value: string | null, priority?: string | undefined) => void)'.
Type 'string' is not assignable to type '(property: string) => string'.
119 this.primaryBar
~~~~~~~~~~~~~~~
120 .style[property as Exclude<keyof CSSStyleDeclaration, 'length'|'parentRule'>] =
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
```
To repro:
1) `npm ci`
2) `npm run build:styling`
3) `tsc -b -f`
|
Docs
|
low
|
Critical
|
587,874,956 |
rust
|
Layout for enums does not match MIR (and types) very well
|
Enum layout computation entirely ignores ZST uninhabited ("absent") variants, which can lead to the layout being quite different in structure from the user-visible enum:
* if there are more than 1 non-absent variants, `Variants::Multiple` is used.
* if there is exactly 1 non-absent variant, `Variants::Single` is used.
* if all variants are absent, the layout of `!` is used, so the entire thing ends up looking like a union with 0 fields.
When actually asking for the layout of such an absent variant, it has to be made up "on the spot".
In terms of testing things like Miri (or anything else that needs to work with layout variants information), this leads to the "interesting" situation that we need an enum with at least 3 variants where 1 is "absent" to make sure that `Variants::Multiple` with absent variants is handled correctly.
This could be improved by generalizing `Variants::Multiple` to also support "discriminants" of "type" `()` and `!`, thus uniformly handling all enums.
Also see the second half of https://github.com/rust-lang/rust/issues/69763#issuecomment-595699745, and discussion below https://github.com/rust-lang/rust/issues/69763#issuecomment-603657246.
Cc @eddyb
|
C-cleanup,T-compiler,A-MIR,A-layout
|
low
|
Minor
|
587,877,388 |
flutter
|
Reliance on iana.org for localizations is flakey
|
`dev/tools/localization/localizations_utils.dart` references iana.org which is causing CI to fail sometimes. We should have local mirrors.
example from cirrus:
```
▌22:06:09▐ ELAPSED TIME: 5.306s for bin/cache/dart-sdk/bin/dart dev/tools/localization/bin/gen_localizations.dart --material in .
Unhandled exception:
HttpException: , uri = https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
ERROR: Last command exited with 255.
Command: bin/cache/dart-sdk/bin/dart dev/tools/localization/bin/gen_localizations.dart --material
Relative working directory: .
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
```
Example luci failure: https://logs.chromium.org/logs/flutter/buildbucket/cr-buildbucket.appspot.com/8884911422563139120/+/steps/run_test.dart_for_framework_tests_shard/0/stdout
```
00:22 +43 -2: test/localization/gen_l10n_test.dart: generateCode plural messages should throw attempting to generate a plural message with no resource attributes [E]
HttpException: , uri = https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry
dart:_http _HttpClient.getUrl
localization/localizations_utils.dart 302:50 precacheLanguageAndRegionTags
test/localization/gen_l10n_test.dart 54:11 main.<fn>
```
|
team,tool,a: internationalization,c: proposal,P2,team-tool,triaged-tool
|
low
|
Critical
|
587,887,630 |
rust
|
Tracking Issue for `-Z src-hash-algorithm`
|
This is the tracking issue for the unstable option `-Z src-hash-algorithm`
### Steps
- [x] Implementation PR #69718
- [ ] Adjust documentation
- [ ] Stabilization PR
### Unresolved Questions
- Should we have a separate option in the target specification to specify the preferred hash algorithm, or continue to use `is_like_msvc`?
- Should continue to have a command line option to override the preferred option? Or is it acceptable to require users to create a custom target specification to override the hash algorithm?
|
A-debuginfo,T-compiler,B-unstable,C-tracking-issue,S-tracking-needs-summary,A-CLI
|
low
|
Major
|
587,891,985 |
flutter
|
if (kRelease) return null; does not treeshake
|
## Use case
Flutter framework uses the following pattern:
```
DiagnosticPropertiesBuilder get builder {
if (kReleaseMode)
return null;
if (_cachedBuilder == null) {
_cachedBuilder = DiagnosticPropertiesBuilder();
value?.debugFillProperties(_cachedBuilder);
}
return _cachedBuilder;
}
```
value?.debugFillProperties causes all debugFillProperties not to be treeshaken. Which on the web is > %2 code size increase (tested using flutter build web --profile).
Wrapping this code in assert on the other hand does tree shake offending methods.
## Proposal
Options:
1- Consistently handle the above across dart2js and AOT and treeshake code below without assert.
2- Introduce asserts into framework code explicitely. Check all current usages of if (kRelease) across the code base.
|
c: new feature,framework,c: performance,dependency: dart,perf: app size,P3,team-framework,triaged-framework
|
low
|
Critical
|
587,953,480 |
pytorch
|
torch.jit.script works for x*sigmoid(x) but not for x*sin(x)
|
cc: @ptrblck
## 🐛 Bug
In the following example, I observe that `x * sigmoid(x)` is JITted as expected. In the first call to `foo`, I see 1 GPU kernel for sigmoid and 1 for mul. In the second call, I see 1 call to `kernel_0` which is the JITted kernel. However, if I replace `sigmoid` wtih `sin`, seemingly another simple pointwise operation, then the JIT does not seem to be working and I observe 4 kernels (2 for sin and 2 for mul). However, if I use the flags
```python
torch._C._jit_set_profiling_mode(False)
torch._C._jit_set_profiling_executor(False)
```
then I see `kernel_0` being invoked twice as expected.
## To Reproduce
```python
import torch
@torch.jit.script
def foo(x):
return x * torch.sigmoid(x)
#return x * torch.sin(x)
x = torch.rand(2048,2048).cuda()
y = foo(x)
y = foo(x)
```
## Steps to reproduce the behavior:
Run the above code as `nvprof python foo.py` to see the list of kernels.
## Expected behavior
I would expect `x * sin(x)` to be JITted as well.
## Environment
```vim
PyTorch version: 1.4.0a0+a5b4d78
Is debug build: No
CUDA used to build PyTorch: 10.2
OS: Ubuntu 18.04.4 LTS
GCC version: (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0
CMake version: version 3.14.0
Python version: 3.6
Is CUDA available: Yes
CUDA runtime version: 10.2.89
GPU models and configuration:
GPU 0: Tesla V100-SXM2-16GB
GPU 1: Tesla V100-SXM2-16GB
GPU 2: Tesla V100-SXM2-16GB
GPU 3: Tesla V100-SXM2-16GB
GPU 4: Tesla V100-SXM2-16GB
GPU 5: Tesla V100-SXM2-16GB
GPU 6: Tesla V100-SXM2-16GB
GPU 7: Tesla V100-SXM2-16GB
Nvidia driver version: 440.33.01
cuDNN version: /usr/lib/x86_64-linux-gnu/libcudnn.so.7.6.5
Versions of relevant libraries:
[pip] msgpack-numpy==0.4.3.2
[pip] numpy==1.17.2
[pip] pytorch-transformers==1.1.0
[pip] torch==1.4.0a0+a5b4d78
[pip] torch-dct==0.1.5
[pip] torchio==0.13.18
[pip] torchtext==0.4.0
[pip] torchvision==0.5.0a0
[conda] magma-cuda101 2.5.2 1 local
[conda] mkl 2020.0 166
[conda] mkl-include 2020.0 166
[conda] nomkl 3.0 0
[conda] pytorch-transformers 1.1.0 pypi_0 pypi
[conda] torch 1.4.0a0+a5b4d78 pypi_0 pypi
[conda] torchtext 0.4.0 pypi_0 pypi
[conda] torchvision 0.5.0a0 pypi_0 pypi
```
cc @suo
|
oncall: jit,triaged
|
low
|
Critical
|
587,967,148 |
flutter
|
Include Gradle version in tool crash GitHub template
|
Include in the **Flutter Application Metadata** section.
https://github.com/flutter/flutter/blob/2dda1297bb499b84de89321c1b19e40516038f5d/packages/flutter_tools/lib/src/reporting/github_template.dart#L133-L143
|
tool,t: gradle,a: triage improvements,P3,team-tool,triaged-tool
|
low
|
Critical
|
587,967,309 |
TypeScript
|
New error: Type of property 'defaultProps' circularly references itself in mapped type
|
https://github.com/nteract/nteract/blob/af734c46893146c617308f4ae1e40bf267e8875f/packages/connected-components/src/header-editor/styled.ts#L24
```
packages/connected-components/src/header-editor/styled.ts:24:34 - error TS2615: Type of property 'defaultProps' circularly references itself in mapped type 'Pick<ForwardRefExoticComponent<Pick<Pick<any, Exclude<keyof ReactDefaultizedProps<StyledComponentInnerComponent<WithC>, ComponentPropsWithRef<StyledComponentInnerComponent<WithC>>>, StyledComponentInnerAttrs<...> | ... 1 more ... | StyledComponentInnerAttrs<...>> | Exclude<...> | Exclude<...> | Exclude<...>> & Parti...'.
24 export const EditableAuthorTag = styled(AuthorTag)`
~~~~~~~~~~~~~~~~~
```
https://github.com/nteract/nteract/blob/af734c46893146c617308f4ae1e40bf267e8875f/packages/presentational-components/src/components/prompt.tsx#L85
Also, I'm seeing each of these errors printed twice.
To repro:
1) `yarn`
2) `tsc -b -f`
Note that it fails with a different (apparently unrelated) error in 3.8.
|
Needs Investigation,Rescheduled
|
medium
|
Critical
|
587,967,332 |
flutter
|
Create way to automate updating of dev/tools/localization/language_subtag_registry.dart
|
`dev/tools/localization/language_subtag_registry.dart` is cached from iana.org, `dev/tools/localization/gen_subtag_registry.dart` generates it but there is nothing that is automatically generating it regularly.
related pr: https://github.com/flutter/flutter/pull/53286
|
tool,framework,a: internationalization,P2,team-framework,triaged-framework
|
low
|
Minor
|
587,977,260 |
pytorch
|
MacOS Error: subprocess.CalledProcessError: Command '['cmake', '--build', '.', '--target', 'install', '--config', 'Release', '--', '-j', '12']' returned non-zero exit status 1.
|
When trying to install Pytorch from source with CUDA enabled, I get the following error after running the last command in the set of instructions for installation (https://github.com/pytorch/pytorch#from-source).
ninja: error: '/usr/local/cuda/lib/libnvToolsExt.dylib', needed by 'lib/libtorch_cuda.dylib', missing and no known rule to make it
Traceback (most recent call last):
File "setup.py", line 745, in <module>
build_deps()
File "setup.py", line 316, in build_deps
cmake=cmake)
File "/Users/jonsaadfalcon/Documents/PoloClub Research/AdversarialProject/pytorch/tools/build_pytorch_libs.py", line 62, in build_caffe2
cmake.build(my_env)
File "/Users/jonsaadfalcon/Documents/PoloClub Research/AdversarialProject/pytorch/tools/setup_helpers/cmake.py", line 339, in build
self.run(build_args, my_env)
File "/Users/jonsaadfalcon/Documents/PoloClub Research/AdversarialProject/pytorch/tools/setup_helpers/cmake.py", line 141, in run
check_call(command, cwd=self.build_dir, env=env)
File "/Users/jonsaadfalcon/anaconda3/lib/python3.7/subprocess.py", line 363, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['cmake', '--build', '.', '--target', 'install', '--config', 'Release', '--', '-j', '12']' returned non-zero exit status 1.
The previously solved issues for this problem were on Windows so idk how to solve it on MacOS. Thank you for the help.
cc @ngimel
|
module: build,module: cuda,triaged,module: macos
|
low
|
Critical
|
587,978,972 |
create-react-app
|
include support for jest-runner-groups
|
### Is your proposal related to a problem?
<!--
Provide a clear and concise description of what the problem is.
For example, "I'm always frustrated when..."
-->
I'd like to categorize my jest tests so that I can run only a certain subset of them.
### Describe the solution you'd like
<!--
Provide a clear and concise description of what you want to happen.
-->
There is already a package to facilitate this called jest-runner-groups, but I cannot use this in conjunction with create-react-app since it needs me to set the "runner" option which is not currently supported.
### Describe alternatives you've considered
<!--
Let us know about other solutions you've tried or researched.
-->
the other alternative is to use `--testNamePattern` which does work but is more cumbersome to use as I need to include a unique word in the title of every test, or surround them in a `describe` block
### Additional context
<!--
Is there anything else you can add about the proposal?
You might want to link to related issues here, if you haven't already.
-->
(Write your answer here.)
|
issue: proposal,needs triage
|
low
|
Minor
|
587,981,534 |
pytorch
|
PyTorch / libtorch executables fail when built against libcuda stub library
|
## 🐛 Bug
PyTorch's cmake setup for CUDA includes a couple `stubs/` directories in the search path for `libcuda.so`:
https://github.com/pytorch/pytorch/blob/v1.4.0/cmake/public/cuda.cmake#L179
Searching for libcuda in the stubs/ directories is needed in various cases, for example when:
- the build host is GPU-less and has CUDA Toolkit but not the GPU driver installed, or
- the build toolchain doesn't include normal system search paths that would normally house libcuda. This is likely if the Anaconda toolchain is being used to build PyTorch.
However, PyTorch's build setup tells cmake to add any non-PyTorch-local libraries to built objects' `RPATH`:
https://github.com/pytorch/pytorch/blob/v1.4.0/cmake/Dependencies.cmake#L14
So the few objects that link directly to libcuda (e.g. `lib/libcaffe2_nvrtc.so` and `bin/test_dist_autograd`) may end up with the libcuda stubs/ directory included in RPATH.
During loading, RPATH has priority over normal system search paths and even `LD_LIBRARY_PATH`. So these objects will prefer the libcuda stub, even on hosts where libcuda is present in the normal location.
The same cmake setup is re-used for libtorch, so the same will occur for libtorch apps as well:
https://discuss.pytorch.org/t/torch-is-available-returns-false/73753
The libcuda stub is not a functional CUDA runtime, so PyTorch and libtorch objects affected by this won't be able to use CUDA even if libcuda is installed in the normal system search path.
This change would prevent a libcuda stubs directory from being added to objects' RPATH, but doesn't seem like a great solution. It abuses a cmake internal mechanism to trick it into avoiding adding the stubs directory.
```
diff --git a/cmake/public/cuda.cmake b/cmake/public/cuda.cmake
index a5c50b90df..3a89c9e8ad 100644
--- a/cmake/public/cuda.cmake
+++ b/cmake/public/cuda.cmake
@@ -187,6 +187,20 @@ find_library(CUDA_NVRTC_LIB nvrtc
PATHS ${CUDA_TOOLKIT_ROOT_DIR}
PATH_SUFFIXES lib lib64 lib/x64)
+# Configuration elsewhere (CMAKE_INSTALL_RPATH_USE_LINK_PATH) arranges that any
+# library used during linking will be added to the objects' RPATHs. That's
+# never correct for a libcuda stub library (the stub library is suitable for
+# linking, but not for runtime, and if present in RPATH will be prioritized
+# ahead of system search path). If libcuda was found in the "stubs" directory,
+# abuse CMAKE_PLATFORM_IMPLICIT_LINK_DIRECTORIES to instruct cmake NOT to
+# include it in RPATH This leaves us without a known path to libcuda in the
+# objects, but there's no help for that if the only libcuda visible at build
+# time is the stub.
+if(CUDA_CUDA_LIB MATCHES "stubs")
+ get_filename_component(LIBCUDA_STUB_DIR ${CUDA_CUDA_LIB} DIRECTORY)
+ list(APPEND CMAKE_PLATFORM_IMPLICIT_LINK_DIRECTORIES ${LIBCUDA_STUB_DIR})
+endif()
+
# Create new style imported libraries.
# Several of these libraries have a hardcoded path if CAFFE2_STATIC_LINK_CUDA
# is set. This path is where sane CUDA installations have their static
```
I hope someone more familiar with cmake and the general pytorch build setup might have a better idea.
## To Reproduce
Steps to reproduce the behavior:
1. Built pytorch on a system with CUDA Toolkit but not GPU driver installed
1. `objdump -p <install_dir>/lib/libcaffe2_nvrtc.so | grep RPATH`
## Expected behavior
PyTorch and libtorch objects should build successfully against libcuda stub, and then run successfully against "real" libcuda (at least when libcuda is present in a default system search path).
## Environment
- PyTorch Version (e.g., 1.0): 1.4.0
- OS (e.g., Linux): Linux
- How you installed PyTorch (`conda`, `pip`, source): source
- Build command you used (if compiling from source): `python setup.py install`
- Python version: `Python 3.6.9 :: Anaconda, Inc.`
- CUDA/cuDNN version: 10.2 / 7.6.5
- GPU models and configuration: V100
- Any other relevant information: Using Anaconda toolchain to build
cc @ezyang @gchanan @zou3519 @seemethere
|
high priority,module: binaries,module: build,triaged
|
low
|
Critical
|
588,003,454 |
flutter
|
Text on desktop Windows 10 looks rough(no antialiasing)
|
Flutter channel: master
Flutter version: 1.16.3-pre.40
Windows 10 build: 19592
Monitor native resolution: 3840x2160
System resolution: 2560x1440
Scaling: 100%
[](https://postimg.cc/qtHZnbrb)
Left is flutter, right is UWP(expected text smoothing)
A more closed up screenshot comparison
[](https://postimg.cc/D8FZjFgn)
flutter
[](https://postimg.cc/0MYvDMhm)
uwp
Reproduction steps:
In powershell, run
`$env:ENABLE_FLUTTER_DESKTOP="true"`
In terminal, run
`flutter config --enable-windows-desktop`
`flutter create <your-app-name>`
`flutter run -d windows`
See the text on the screen for result
|
engine,platform-windows,a: typography,a: desktop,P3,team-windows,triaged-windows
|
medium
|
Critical
|
588,017,406 |
flutter
|
[animations] Allow different containers for source of open and target of closing animation
|
## Use case
I'm using OpenContainer with a FAB to open a new screen and add a new item to a list of OpenContainer Cards. The problem is once you add a new item, the OpenContainer animates back into the FAB which is confusing visually.
My implementation is very similar to:
https://github.com/flutter/packages/blob/master/packages/animations/example/lib/container_transition.dart
## Proposal
Add ability to change the route on the fly and animate back to the item that was just added, similar to what can be achieved with Hero animations.
|
c: new feature,package,c: proposal,p: animations,team-ecosystem,P3,triaged-ecosystem
|
low
|
Major
|
588,020,866 |
react
|
Bug: Event handlers on custom elements work on the client but not on the server
|
<!--
Please provide a clear and concise description of what the bug is. Include
screenshots if needed. Please test using the latest version of the relevant
React packages to make sure your issue has not already been fixed.
-->
React version: 16.13.1
## Steps To Reproduce
```js
ReactDOMServer.renderToString(<div-x onClick={() => console.log('clicked')} />)
```
vs
```js
ReactDOM.render(<div-x onClick={() => { console.log('clicked'); }} />, root)
```
<!--
Your bug will get fixed much faster if we can run your code and it doesn't
have dependencies other than React. Issues without reproduction steps or
code examples may be immediately closed as not actionable.
-->
Link to code example: https://jsfiddle.net/hsug65x0/2/
<!--
Please provide a CodeSandbox (https://codesandbox.io/s/new), a link to a
repository on GitHub, or provide a minimal code example that reproduces the
problem. You may provide a screenshot of the application if you think it is
relevant to your bug report. Here are some tips for providing a minimal
example: https://stackoverflow.com/help/mcve.
-->
## The current behavior
Custom element event handlers only work client side. In SSR the code of the event handler becomes an attribute value.
## The expected behavior
Custom element event handlers work in SSR.
|
Component: Server Rendering,Type: Needs Investigation
|
low
|
Critical
|
588,032,350 |
flutter
|
Channels are ~2x slower on Android versus iOS
|
Microbenchmarks from https://github.com/gaaclarke/MessagingTiming
Things are going to be a bit slower because of JNI, 1ms seems excessive.
**Math:**
(0.4µs / 2.5µs) * (1000µs / 80µs) = 2
**Android:**
**iOS:**
|
platform-android,engine,c: performance,perf: speed,P3,team-android,triaged-android
|
low
|
Major
|
588,032,524 |
electron
|
Closing a child window brings a previously activate application to the front and the parent window is sent to the back
|
### Preflight Checklist
* [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 an issue that matches the one I want to file, without success.
### Issue Details
* **Electron Version:** 8.2.0
* **Operating System:** Windows 10
* **Last Known Working Electron version:** not sure
### Expected Behavior
When my electron app is visible and I have a child window opened, I expect that closing the child window will keep the main electron window on the front.
### Actual Behavior
In a specific case (see below), an application that was previously over the electron app gets brought to the front upon closing the window. Electron app is sent to the back.
### To Reproduce
1. `npm start` the below script.
2. In devtools, open a child window (e.g. `window.open("https://google.com")`).
3. Select another application from the task bar to bring it to the front of the electron app.
4. Click the electron app icon in the taskbar to bring it to the front.
5. Close the window.
6. Observe that the electron app is sent to the back and replaced with the other app that was previously brought to the front.
main.js:
```
const { app, BrowserWindow } = require('electron')
async function createWindow() {
const mainWindow = new BrowserWindow({ webPreferences: { nativeWindowOpen: true } });
await mainWindow.loadURL("https://google.com");
mainWindow.webContents.openDevTools();
mainWindow.webContents.on("new-window", (e, url, framename, disp, options, additionalFeatures, referrer) => {
options.parent = mainWindow;
});
}
app.on('ready', createWindow)
```
|
platform/windows,bug :beetle:,8-x-y,10-x-y
|
medium
|
Major
|
588,034,686 |
flutter
|
Review flutter attach performance
|
On a simple test, it takes ~8s to connect to an already running flutter observatory via `flutter attach`. Though ~3s was from cleaning up the devfs and closing the process.
Log gist at https://gist.github.com/xster/5becc4937c2cef8d5e84660d0601abd1
There's a 2397ms apparently pure Dart part in the middle of the process too.
We should optimize.
|
tool,c: performance,a: existing-apps,P2,team-tool,triaged-tool
|
low
|
Major
|
588,053,097 |
TypeScript
|
New error: Type instantiation is excessively deep and possibly infinite.
|
This one's actually new in 3.8. There are some additional errors in 3.9 - I'll post them below.
https://github.com/davidkpiano/xstate/blob/2e0bd5dcb11c8ffc5bfc9fba1d3c66fb7546c4bd/packages/core/src/stateUtils.ts#L157
```
packages/core/src/stateUtils.ts:157:36 - error TS2589: Type instantiation is excessively deep and possibly infinite.
157 return getValueFromAdj(rootNode, getAdjList(config));
~~~~~~~~~~~~~~~~~~
packages/core/src/State.ts:181:12 - error TS2589: Type instantiation is excessively deep and possibly infinite.
181 return new State(config);
~~~~~~~~~~~~~~~~~
packages/core/src/interpreter.ts:208:5 - error TS2589: Type instantiation is excessively deep and possibly infinite.
208 this.parent = parent;
~~~~~~~~~~~~~~~~~~~~
packages/core/src/interpreter.ts:783:13 - error TS2345: Argument of type '{ type: string; data: any; }' is not assignable to parameter of type 'string | EventObject | Event<EventObject>[] | Event<EventObject>'.
Object literal may only specify known properties, and 'data' does not exist in type 'EventObject'.
783 data: err
~~~~~~~~~
```
To repro:
1) `yarn`
2) `tsc -b -f tsconfig.monorepo.json`
|
Needs Investigation
|
low
|
Critical
|
588,076,178 |
flutter
|
Don't destroy engine by default in the activity if the engine was provided by a subclass
|
https://cs.opensource.google/flutter/engine/+/master:shell/platform/android/io/flutter/embedding/android/FlutterFragmentActivity.java;l=552
is conceptually at odds with
https://cs.opensource.google/flutter/engine/+/master:shell/platform/android/io/flutter/embedding/android/FlutterFragmentActivity.java;l=150
|
platform-android,engine,P2,team-android,triaged-android
|
low
|
Minor
|
588,081,805 |
pytorch
|
torch::normal only supports (double, double), but at::normal supports (double, double) / (double, Tensor) / (Tensor, double) / (Tensor, Tensor)
|
`torch::normal` only supports (double, double) as input, but `at::normal` supports (double, double) / (double, Tensor) / (Tensor, double) / (Tensor, Tensor) as input.
Python `torch.normal` also supports all four types of inputs (https://pytorch.org/docs/stable/torch.html#torch.normal), and we should do the same for `torch::normal`.
Some notes:
- `torch::normal` is currently defined in`torch/csrc/autograd/generated/variable_factories.h`
- Per conversation with Ed, we should put the tracing logic in VariableType and just dispatch `torch::normal` to VariableType
(This issue was originally reported by @ailzhang)
cc @yf225
cc @ailzhang @ezyang
|
module: cpp,triaged
|
low
|
Minor
|
588,101,099 |
pytorch
|
[JIT] dropout fails on legacy mode
|
## 🐛 Bug
## To Reproduce
Steps to reproduce the behavior:
On pytorch master build (1.5.0a0+efbd6b8), run
```python
python test/test_jit_legacy.py -k test_dropout_cuda
```
The test fails.
## Expected behavior
However,
```python
python test/test_jit.py -k test_dropout_cuda
```
passes.
## Environment
- PyTorch Version (e.g., 1.0): 1.5.0a0+efbd6b8
- OS (e.g., Linux): Fedora 29
- Python version: 3.7.6
- CUDA/cuDNN version: 10.1
## Additional context
This may be related to #34361. The test failure may also be related to legacy profiling executer. By commenting out the following line, the legacy test will also pass.
https://github.com/pytorch/pytorch/blob/efbd6b8533bfa42f9e57d11b453b8a8bf4cf18ea/test/test_jit.py#L168
I'm also not sure if `test/test_jit_legacy.py` runs on pytorch CI?
cc @suo @ptrblck @jjsjann123
|
oncall: jit,triaged
|
low
|
Critical
|
588,122,896 |
opencv
|
Support the property CAP_PROP_SETTINGS under linux
|
##### System information (version)
<!-- Example
- OpenCV => 4.2
- Operating System / Platform => Windows 64 Bit
- Compiler => Visual Studio 2017
-->
- OpenCV => 4.2
- Operating System / Platform =>Ubuntu 18.04:
- Compiler => Qt Creator 4.11.0
##### Detailed description
when I use Videocapture , I want to use the property CAP_PROP_SETTINGS , but it was only support windows direct show , Do you consider support this property under linux .
<!-- 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
-->
##### Issue submission checklist
- [x] I report the issue, it's not a question
<!--
OpenCV team works with answers.opencv.org, Stack Overflow and other communities
to discuss problems. Tickets with question without real issue statement will be
closed.
-->
- [x] I checked the problem with documentation, FAQ, open issues,
answers.opencv.org, Stack Overflow, etc and have not found solution
<!--
Places to check:
* OpenCV documentation: https://docs.opencv.org
* FAQ page: https://github.com/opencv/opencv/wiki/FAQ
* OpenCV forum: https://answers.opencv.org
* OpenCV issue tracker: https://github.com/opencv/opencv/issues?q=is%3Aissue
* Stack Overflow branch: https://stackoverflow.com/questions/tagged/opencv
-->
- [x] I updated to latest OpenCV version and the issue is still there
<!--
master branch for OpenCV 4.x and 3.4 branch for OpenCV 3.x releases.
OpenCV team supports only latest release for each branch.
The ticket is closed, if the problem is not reproduced with modern version.
-->
- [ ] There is reproducer code and related data files: videos, images, onnx, etc
<!--
The best reproducer -- test case for OpenCV that we can add to the library.
Recommendations for media files and binary files:
* Try to reproduce the issue with images and videos in opencv_extra repository
to reduce attachment size
* Use PNG for images, if you report some CV related bug, but not image reader
issue
* Attach the image as archite to the ticket, if you report some reader issue.
Image hosting services compress images and it breaks the repro code.
* Provide ONNX file for some public model or ONNX file with with random weights,
if you report ONNX parsing or handling issue. Architecture details diagram
from netron tool can be very useful too. See https://lutzroeder.github.io/netron/
-->
|
feature,priority: low,category: videoio(camera)
|
low
|
Critical
|
588,134,871 |
create-react-app
|
Add new template for TypeSafe GraphQL (Apollo + GraphQL codegen) like Typescript template.
|
### Is your proposal related to a problem?
<!--
Provide a clear and concise description of what the problem is.
For example, "I'm always frustrated when..."
-->
No
### Describe the solution you'd like
<!--
Provide a clear and concise description of what you want to happen.
-->
As GraphQL (Apollo) users are getting increased a lot, we have to setup all Apollo, graphQL codegen, their plugins and other such tools in CRA to bootstrap a React+Apollo app. It will be great to create a template.
### Describe alternatives you've considered
<!--
Let us know about other solutions you've tried or researched.
-->
No
### Additional context
<!--
Is there anything else you can add about the proposal?
You might want to link to related issues here, if you haven't already.
-->
I can work on this issue, if approved
|
issue: proposal,needs triage
|
low
|
Minor
|
588,183,524 |
pytorch
|
Shared-QK transformer for the transformer (nn.activation.MultiheadAttention) module in PyTorch?
|
## 🚀 Feature
<!-- A clear and concise description of the feature proposal -->
We can reduce the number of parameters by sharing the linear projection layers for queries and keys in the nn.Transformer and nn.activation.MultiheadAttention modules.
I think we edit the init and forward function in nn.activation.MultiheadAttention.
https://github.com/pytorch/pytorch/blob/be0cdf5d157a84dd28bd95ca8f5f82306c0340f5/torch/nn/modules/activation.py#L757
https://github.com/pytorch/pytorch/blob/be0cdf5d157a84dd28bd95ca8f5f82306c0340f5/torch/nn/modules/activation.py#L853
## Motivation
The recent paper, Reformer, proposed the shared-QK Transformer that shares the linear projection layer for keys and queries. It reduces the number of parameters and memory space while not affecting the model performance.
https://arxiv.org/abs/2001.04451
I think it is a very simple but efficient technique. If PyTorch provides this technique, many researchers can efficiently manage resources such as GPU memory.
<!-- Add any other context or screenshots about the feature request here. -->
cc @albanD @mruberry @jbschlosser @walterddr @mikaylagawarecki
|
module: nn,triaged,needs research
|
low
|
Major
|
588,208,950 |
godot
|
RigidBody sometimes misses reporting a contact
|
**Godot version:**
3.2.1
**Issue description:**
The body_entered signal is sometimes not triggered on contact. Not entirely random, but depending on phsyics settings.
**Steps to reproduce:**
Download minimal project and watch the bounce counter in the console.
**Minimal reproduction project:**
https://www.dropbox.com/s/a3n3kvam5cd813x/MinimalProject.zip?dl=0
|
bug,topic:physics,needs testing
|
low
|
Major
|
588,220,014 |
go
|
cmd/go: go get -u with GIT_TERMINAL_PROMPT=1 combines separate git username/password prompts, which can lead to password being exposed
|
<!--
Please answer these questions before submitting your issue. Thanks!
For questions please use one of our forums: https://github.com/golang/go/wiki/Questions
-->
### What version of Go are you using (`go version`)?
<pre>
$ go version go1.14.1 linux/amd64
</pre>
### Does this issue reproduce with the latest release?
Not Tested
### What operating system and processor architecture are you using (`go env`)?
<details><summary><code>go env</code> Output</summary><br><pre>
$ go env
GO111MODULE=""
GOARCH="amd64"
GOBIN=""
GOCACHE="/home/nibin/.cache/go-build"
GOENV="/home/nibin/.config/go/env"
GOEXE=""
GOFLAGS=""
GOHOSTARCH="amd64"
GOHOSTOS="linux"
GOINSECURE=""
GONOPROXY=""
GONOSUMDB=""
GOOS="linux"
GOPATH="/home/nibin/go"
GOPRIVATE=""
GOPROXY="https://proxy.golang.org,direct"
GOROOT="/snap/go/5569"
GOSUMDB="sum.golang.org"
GOTMPDIR=""
GOTOOLDIR="/snap/go/5569/pkg/tool/linux_amd64"
GCCGO="gccgo"
AR="ar"
CC="gcc"
CXX="g++"
CGO_ENABLED="1"
GOMOD="/home/nibin/golang/mod-exp/sample-mods/mysample/go.mod"
CGO_CFLAGS="-g -O2"
CGO_CPPFLAGS=""
CGO_CXXFLAGS="-g -O2"
CGO_FFLAGS="-g -O2"
CGO_LDFLAGS="-g -O2"
PKG_CONFIG="pkg-config"
GOGCCFLAGS="-fPIC -m64 -pthread -fmessage-length=0 -fdebug-prefix-map=/tmp/go-build009035298=/tmp/go-build -gno-record-gcc-switches"
</pre></details>
### What did you do?
I tried to update go module dependency ( 2 separate go modules residing in 2 repos) ( the dependency(s) resides in a private repo ) using <code>GIT_TERMINAL_PROMPT=1 GOSUMDB=off go get -u</code>
<!--
If possible, provide a recipe for reproducing the error.
A complete runnable program is good.
A link on play.golang.org is best.
-->
### What did you expect to see?
#### Assume Password Entered is : <code>mypassword</code>
```
Username for 'https://gitlab.com': myusername
Password for 'https://[email protected]':
Username for 'https://gitlab.com': myusername
Password for 'https://[email protected]':
```
### What did you see instead?
#### Assume Password Entered is : <code>mypassword</code>
```
Username for 'https://gitlab.com':
Username for 'https://gitlab.com': myusername
Password for 'https://[email protected]':
Password for 'https://[email protected]':
```
|
NeedsDecision,GoCommand
|
low
|
Critical
|
588,231,398 |
godot
|
WebGL warning about TEXTURE_CUBE_MAP in Firefox
|
**Godot version:**
3.2.1-stable
**OS/device including version:**
Mageia 8 x86_64
Firefox 73.0.1, 74.0 (albeit with other bug #37245), Beta 75.0b8 and Nightly 76.0a1.
**Issue description:**
When running the simple MRP from #37245 in Firefox, the following warnings are displayed in the Web Console:
```
Error: WebGL warning: generateMipmap: Tex image TEXTURE_CUBE_MAP_POSITIVE_X level 0 is incurring lazy initialization. index.js:7:260712
Error: WebGL warning: generateMipmap: Tex image TEXTURE_CUBE_MAP_NEGATIVE_X level 0 is incurring lazy initialization. index.js:7:260712
Error: WebGL warning: generateMipmap: Tex image TEXTURE_CUBE_MAP_POSITIVE_Y level 0 is incurring lazy initialization. index.js:7:260712
Error: WebGL warning: generateMipmap: Tex image TEXTURE_CUBE_MAP_NEGATIVE_Y level 0 is incurring lazy initialization. index.js:7:260712
Error: WebGL warning: generateMipmap: Tex image TEXTURE_CUBE_MAP_POSITIVE_Z level 0 is incurring lazy initialization. index.js:7:260712
Error: WebGL warning: generateMipmap: Tex image TEXTURE_CUBE_MAP_NEGATIVE_Z level 0 is incurring lazy initialization. index.js:7:260712
```
CC @clayjohn @Faless
**Steps to reproduce:**
- Download MRP "Export"
- Run in Firefox (`python3 -m http.server` to start web server in the project folder)
**Minimal reproduction project:**
[Minimal Project Crash.zip](https://github.com/godotengine/godot/files/4368764/Minimal.Project.Crash.zip)
[Minimal Project Crash Export.zip](https://github.com/godotengine/godot/files/4368765/Minimal.Project.Crash.Export.zip)
|
bug,platform:web,topic:rendering,confirmed
|
low
|
Critical
|
588,257,048 |
vscode
|
Event for onWill|DidRename|Create|Delete should indicate what caused the event (explorer vs workspace edit)
|
Backstory: https://github.com/redhat-developer/vscode-java/issues/1222#issuecomment-603750889
We have events that fore before and after files are being created, deleted, or renamed. They fire when users manipulate files, e.g. in the explorer, or when the workspace-edits-api is used. There are cases in which it is valuable to know what the source of said events are.
|
feature-request,api,file-io
|
low
|
Minor
|
588,279,068 |
ant-design
|
Carousel RTL sliding direction
|
- [x] I have searched the [issues](https://github.com/ant-design/ant-design/issues) of this repository and believe that this is not a duplicate.
### Reproduction link
[https://ant.design/components/carousel-cn/#components-carousel-demo-basic](https://ant.design/components/carousel-cn/#components-carousel-demo-basic)
### Steps to reproduce
- Open link
- Select RTL
- Tap 4 sliders to switch
### What is expected?
currently 1, click 2
- The slider should slide in from the left
### What is actually happening?
- The slider should slide in from the right
| Environment | Info |
|---|---|
| antd | 4.0.4 |
| React | The latest |
| System | mac |
| Browser | chrome |
<!-- generated by ant-design-issue-helper. DO NOT REMOVE -->
|
help wanted,Inactive,rtl
|
low
|
Major
|
588,308,620 |
youtube-dl
|
nzgeo.com videos
|
<!--
######################################################################
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 2020.03.24. 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 **2020.03.24**
- [x] I've checked that all provided URLs are alive and playable in a browser
- [x] I've checked that none of provided URLs violate any copyrights
- [x] I've searched the bugtracker for similar site support requests including closed ones
## Example URLs
<!--
Provide all kinds of example URLs support for which should be included. Replace following example URLs by yours.
-->
- Single video: https://www.nzgeo.com/video/stewart-island/
## Description
<!--
Provide any additional information.
If work on your issue requires account credentials please provide them or explain how one can obtain them.
-->
Videos from https://www.nzgeo.com/ play OK in Firefox. From the page's HTML source, the following URL seems to be the one to use:
<source src="//d24vm4mjvxzrpc.cloudfront.net/AZLTN52l/hls/AZLTN52l.m3u8" type="application/x-mpegurl">
|
site-support-request
|
low
|
Critical
|
588,312,928 |
flutter
|
Smoke-test that localhost servers are allowed by developer's computer
|
## Steps to Reproduce
<!-- Please tell us exactly how to reproduce the problem you are running into. -->
1. dev channel v1.16.2 , run web Exception
## Logs
Launching lib\main.dart on Chrome in debug mode...
Building application for the web...
Finished with error: SocketException: Failed to create server socket (OS Error: Failed to start accept), address = localhost, port = 52113
|
tool,customer: crowd,platform-web,P2,team-web,triaged-web
|
low
|
Critical
|
588,366,127 |
godot
|
Signal 'files_dropped' not emitting when engine is run as administrator
|
**Godot version:** 3.2.1.stable.official
**OS/device including version:** Windows 10 x64
**Issue description:** Engine doesn't react when a file or multiple files are dropped over game window. Sample project still works with 3.2 build.
**Steps to reproduce:** connect _SceneTree_ with _files_dropped_ signal.
**Minimal reproduction project:** [FileDropTest.zip](https://github.com/godotengine/godot/files/4386951/FileDropTest.zip)
|
bug,platform:windows,topic:editor,confirmed
|
low
|
Major
|
588,381,633 |
TypeScript
|
"export class" does not compile with outfile=foo (module=none, es6)
|
**TypeScript Version:** 3.8.3
**Search Terms:**
**Code**
```jsonc
{
"$schema": "http://json.schemastore.org/tsconfig",
"compilerOptions": {
"module": "none",
"target": "es6",
"types": [],
"outFile": "newFile.js" // <- important
},
"files": [
"main.ts"
]
}
```
```ts
// main.ts
declare module Common {
abstract class absCom {
constructor();
}
}
export /* <- important */ class Com extends Common.absCom {
constructor() {
super();
}
}
```
**Expected behavior:**
This should compile without problems.
**Actual behavior:**
When I configure an `outfile` **no file** is generated (`js` and no `.d.ts`, `.js.map`).
But we get **no error message** about this!
This is the same with `--verbose`:
```
[13:11:21] Project 'foo/tsconfig.json' is out of date because output file 'foo/newFile.js' does not exist
[13:11:21] Building project 'foo/tsconfig.json'...
[13:11:23] Updating unchanged output timestamps of project 'foo/tsconfig.json'...
```
Test 1:
Removing the `outfile` generates the following block in `main.js`:
````js
"use strict";
Object.defineProperty(exports, "__esModule", { value: true });
class Com extends Common.absCom {
constructor() {
super();
}
}
exports.Com = Com;
````
Test 2:
Removing `export` in `main.ts` we get in `newFile.js`:
````js
class Com extends Common.absCom {
constructor() {
super();
}
}
````
Thats why I think merging `Object.defineProperty(exports, "__esModule", { value: true });` into one `outfile` causes the abort.
**Playground Link:**
Not possible, as we need `outfile` to trigger the bug.
**Related Issues:**
none found
|
Bug
|
low
|
Critical
|
588,453,197 |
rust
|
Macro Invocations Behave Questionable Regarding Semicolons
|
### Calling Statement Macros with “return values”?
Macros can be used both for expressions and for sequences of statements. The latter being useful to create new definitions (e.g. one or even multiple `let` clause(s)).
I’m observing that the general rule that _expressions ending in a “Block” of type `()` don’t need a semicolon when used as a statement_ translates in a weird way to even macros with multiple statements. For example
```rust
macro_rules! s {
($x:ident) => {
let $x = 1;
"hello world"
}
}
```
can’t be used as `s!{x}`. However `s!{x};` does work (as does `s!(x);`). I’m questioning how sensible it is to support these kind of macros since they have some kind of “return value”, whose type even is relevant for if they need a trailing `;`, while the “return value” cannot actually in any way (that I can think of) be retrieved.
A way nicer approach at resolving the non-working `s!{x}` is of course to change the macro:
```rust
macro_rules! s {
($x:ident) => {
let $x = 1;
"hello world"; // ⟵ added semicolon
}
}
```
### Ignored Semicolon
One (or at least I) would expect the macro `m1` below to produce just a statement (judging from the semicolon), i.e. be illegal in expression context. However it somehow _does_ pass as an expression, totally ignoring the semicolon. In the return position of a function it does produce `()` though [when called like `m1!{}` – on the other hand `m1!()` would still produce the integer].
```rust
macro_rules! m1 {
() => {
1;
}
}
macro_rules! m2 {
() => {
1
}
}
fn main() {
let x1 = m1!{};
println!("{}", x1);
let x2 = m1!{};
println!("{}", x2);
m1!();
m1!{}
m2!();
// m2!{} // illegal
m2!{};
m1!{} // apparently even returns ()
}
```
([Playground](https://play.rust-lang.org/?version=stable&mode=debug&edition=2018&gist=a36e69afb99a3d85eb5614baf5e850ae))
Output:
```
1
1
```
Errors:
```
Compiling playground v0.0.1 (/playground)
Finished dev [unoptimized + debuginfo] target(s) in 0.40s
Running `target/debug/playground`
```
|
A-frontend,A-parser,A-macros,T-lang,C-bug
|
low
|
Critical
|
588,458,523 |
youtube-dl
|
Ucha.se -- Teaching - Lessons
|
## Checklist
- [x ] I'm reporting a new site support request
- [x ] I've verified that I'm running youtube-dl version **2020.03.24**
- [x ] I've checked that all provided URLs are alive and playable in a browser
- [x ] I've checked that none of provided URLs violate any copyrights
- [x ] I've searched the bugtracker for similar site support requests including closed ones
## Example URLs
- Single video: https://ucha.se/watch/711/Sasedni-agli-Protivopolozhni-agli/urok/1676
- Playlist: --
## Description
I'am not sure how this website is streaming the videos, but i guess that are using AES and .ts and some key in addition.
https://proxy1.ucha.se/uchase/_definist_/smil:videos/07112017-Popraveno-6-Sasedni-agli/stream.smil/chunklist_b410806.m3u8?s=9cb2e85bc47e85b36232c7333b65e750
https://ucha.se/streaming/key/?s=9cb2e85bc47e85b36232c7333b65e750
|
site-support-request
|
low
|
Critical
|
588,461,264 |
scrcpy
|
how to install scrcpy correctly on deepin.
|
I had problems with the installation on deepin. I looked at similar problems but none solved them.
stormtrooper@stormtrooper-C4T7:~$ scrcpy
INFO: scrcpy 1.12.1 <https://github.com/Genymobile/scrcpy>
ERROR: '/usr/local/share/scrcpy/scrcpy-server' does not exist or is not a regular file (manually)
stormtrooper@stormtrooper-C4T7:~$ snap list
Name Version Rev Tracking Publisher Notes
core 16-2.43.3 8689 stable canonical✓ core,devmode (by snap)
|
snap
|
low
|
Critical
|
588,476,662 |
TypeScript
|
Private members have implicit any type in generated declaration files
|
It seems the bug that was fixed in #36216 is happening again:
When generating delcarations for private class members, no type information is appended which causes the declaration file to generate errors when imported into a project with noImplicitAny enabled.
**TypeScript Version:** 3.8.3
**Code**
```ts
// Test.ts
export default class Test {
private _number: number = 0;
public str: string = '';
private get num() { return this._number; }
}
```
**Expected behavior:**
```ts
// Test.d.ts
export default class Test {
private num: any;
str: string;
private get num(): any;
}
```
**Actual behavior:**
```ts
// Test.d.ts
export default class Test {
private _number;
str: string;
private get num();
}
```
**Related Issues:** #36216, #36639
|
Needs Investigation
|
low
|
Critical
|
588,491,505 |
vue
|
"refInFor" searches upwards of inline-template boundary
|
### Version
2.6.11
### Reproduction link
[https://jsfiddle.net/b71adnym/2/](https://jsfiddle.net/b71adnym/2/)
### Steps to reproduce
Nest a component with an inline-template that has a ref in a v-for.
### What is expected?
component.$ref.ref should point to the respective element.
### What is actually happening?
component.$ref.ref is a singleton array with the element reference.
---
checkInFor loops over the element ancestors up to the document root, so it believes that the ref is in a loop, but it actually isn't.
The check should stop at the component root.
<!-- generated by vue-issues. DO NOT REMOVE -->
|
bug
|
low
|
Minor
|
588,494,678 |
TypeScript
|
`import './file.json'`: This module is declared with using 'export ='
|
In short, JSON files's implied type definition is always defined via `export =`, even if I'm using `module: es2015`, causing it to error unless I use an unnecessary flag (`allowSyntheticDefaultImports`)
**TypeScript Version:** 3.9.0-dev.20200326
**Search Terms:**
import require json module commonjs es2015
**Code**
```ts
import list from "./list.json";
for (const item of list) {
console.log(item)
}
```
```json
{
"compilerOptions": {
"outDir": "distribution",
"target": "es2017",
"module": "es2015",
"moduleResolution": "node",
"resolveJsonModule": true,
"strict": true
},
"files": [
"index.ts"
]
}
```
**Expected behavior:**
`list` is of type `string[]`
**Actual behavior:**
```rb
index.ts:1:8 - error TS1259: Module '"./list"' can only be default-imported using the 'allowSyntheticDefaultImports' flag
1 import list from "./list.json";
~~~~
list.json:1:1
1 [
~
This module is declared with using 'export =', and can only be used with a default import when using the 'allowSyntheticDefaultImports' flag.
Found 1 error.
```
**Workaround:**
1. `allowSyntheticDefaultImports: true` works as expected, but that flag is unsafe and I don't want (need) to use it.
2.
```ts
declare module './list.json' {
const list: string[];
export default list;
}
```
**Playground Link:** [repro.zip](https://github.com/microsoft/TypeScript/files/4387999/repro.zip)
**Related Issues:** https://github.com/microsoft/TypeScript/issues/15146
|
Suggestion,In Discussion
|
low
|
Critical
|
588,526,944 |
excalidraw
|
Allow loading a drawing from URL
|
I saw [this tweet](https://twitter.com/excalidraw/status/1243185992531677184) and like the idea that Excalidraw can actually be used as an open "drawing" format by other apps.
It would be nice if apps could "export to Excalidraw" using a simple link, something like:
```
https://excalidraw.com/?load=[url-encoded json payload here]
```
A big drawing might quickly reach the max standard length of an URL, but if it's only loaded on the client side, it might work fine?
|
enhancement
|
medium
|
Major
|
588,570,991 |
rust
|
Tracking issue for `Write::write_all_vectored`
|
This is a tracking issue for `io::Write::write_all_vectored`.
Feature gate: `#![feature(write_all_vectored)]`.
**Steps**:
- [x] Implement the RFC (#70612).
- [ ] Stabilization PR.
**Unresolved questions**:
- [ ] Can we improve the API? Currently the method takes the `IoSlice`s as mutable slice and modifies them. This is a pretty unusual and potentially error-prone API. We could either find a way to not mutate the argument or to enforce (via the type system) the argument can't be used by the user afterwards. Or we can decide that such an unusual API is fine for this rather low level method.
---
Original issue:
<details>
In the `io::Write` trait we've got the helpful `write_all` method, which calls `write` in a loop to write all bytes. However there is no such function for `write_vectored`. I would suggest adding a function called `write_all_vectored` to performance such a task.
A possible implementation. Note that `bufs` is a mutable slice, also see the discussion in https://github.com/rust-lang/futures-rs/pull/1741/files. On the playground: https://play.rust-lang.org/?version=nightly&mode=debug&edition=2018&gist=872e9d973bd8101e7724292f87a82869.
```rust
pub trait Write {
// ...
fn write_all_vectored(&mut self, mut bufs: &mut [IoSlice<'_>]) -> io::Result<()> {
while !bufs.is_empty() {
match self.write_vectored(bufs) {
Ok(0) => {
return Err(Error::new(
ErrorKind::WriteZero,
"failed to write whole buffer",
));
}
Ok(n) => bufs = IoSlice::advance(mem::replace(&mut bufs, &mut []), n),
Err(ref e) if e.kind() == ErrorKind::Interrupted => {}
Err(e) => return Err(e),
}
}
Ok(())
}
}
```
Related: https://github.com/rust-lang/futures-rs/pull/1741
/cc @cramertj
</details>
|
T-libs-api,B-unstable,C-tracking-issue,A-io,Libs-Tracked
|
medium
|
Critical
|
588,589,044 |
godot
|
Windows always appear on first screen
|
**Godot version:**
5f1107aa3
**OS/device including version:**
Windows 10
**Issue description:**
Any new Window will appear on the first screen (monitor). IMO they should appear on the same screen as parent Window.
**Steps to reproduce:**
1. Open Godot editor
2. Move it to second screen
3. Add node
4. Dialog appears on another screen :/
EDIT:
I just noticed that any popup (including sub-menus) have this issue ;_;
|
bug,topic:editor,confirmed,usability
|
low
|
Major
|
588,591,481 |
pytorch
|
Multiprocessing map gets stuck if doing inference on loaded model
|
## 🐛 Bug
When doing inference on a loaded model through the torch.multiprocessing.map function the code gets stuck. The same does not apply if I use a model that is not loaded (e.g. I just instantiate one with random weights) or if I do not use multiprocessing but use the loaded model.
I guess is somewhat related to: [this issue 16943](https://github.com/pytorch/pytorch/issues/16943)
## To Reproduce
A simple code snippet to show the structure of my code is:
```python
# This is evaluator.py
class Evaluator(nn.Module):
def __init__(self):
self.model = nn.Linear(10, 5, bias=False)
ckpt = torch.load(filepath, map_location=self.device)
self.load_state_dict(ckpt)
def forward(self, x):
return self.model(x)
```
```python
# This is the main file main.py
import torch.multiprocessing as mp
from evaluator import Evaluator
pool = None
class Main():
def __init__(self):
global pool
eval = Evaluator()
pool = mp.Pool(processes=4)
def execute(self, data):
global pool
results = pool.map(eval.forward, data)
return results
```
```python
# This is the script
from main import Main
if __name__ == "__main__":
main = Main()
data = [list of tensors]
results = main.execute(data)
```
## Expected behavior
What I expect is that both when I load the model or when I don't load the model the map function would return. What happens instead is that if I load the model the map function remains stuck.
This does not happens if I do not use the map, but a normal for loop, without any multiprocess.
## Environment
PyTorch version: 1.4.0
Is debug build: No
CUDA used to build PyTorch: 10.1
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: No
CUDA runtime version: No CUDA
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
Versions of relevant libraries:
[pip3] numpy==1.18.1
[pip3] torch==1.4.0
[pip3] torchvision==0.5.0
[conda] Could not collect
|
module: multiprocessing,triaged
|
medium
|
Critical
|
588,678,350 |
create-react-app
|
[Docs] Add details regarding custom environment file usage to the "Environment Variables" section.
|
### Is your proposal related to a problem?
<!--
Provide a clear and concise description of what the problem is.
For example, "I'm always frustrated when..."
-->
We currently found it a little confusing to discover how arbitrary `dotenv` files could be used.
The information is there on the [Deployment page](https://create-react-app.dev/docs/deployment/#customizing-environment-variables-for-arbitrary-build-environments), but it was not readily discoverable from the [Environment Variables page](https://create-react-app.dev/docs/adding-custom-environment-variables).
### Describe the solution you'd like
<!--
Provide a clear and concise description of what you want to happen.
-->
Move the ["Customizing Environment Variables for Arbitrary Build Environments"](https://create-react-app.dev/docs/deployment/#customizing-environment-variables-for-arbitrary-build-environments) section from the Deployment page to the Environment Variables page, as a subsection to ["Adding Development Environment Variables In .env"](https://create-react-app.dev/docs/adding-custom-environment-variables#adding-development-environment-variables-in-env).
The old section on the Deployment page will be replaced by a link pointing to the Environment Variables page.
### Describe alternatives you've considered
<!--
Let us know about other solutions you've tried or researched.
-->
Instead of moving the content from Deployment, we could just make a reference to it in Environment Variables.
### Additional context
<!--
Is there anything else you can add about the proposal?
You might want to link to related issues here, if you haven't already.
-->
I'm more than happy to submit a PR for this! Wanted to create an issue first though.
|
issue: proposal,needs triage
|
low
|
Minor
|
588,726,316 |
pytorch
|
Support tensor.cumsum() for 1-dim tensors
|
```
>>> torch.rand(3).cumsum()
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: cumsum() received an invalid combination of arguments - got (), but expected one of:
* (name dim, torch.dtype dtype)
* (int dim, torch.dtype dtype)
```
Currently one must pass `dim=0` or `dim=-1`, which makes code too verbose (and even a bit confusing) if one works only with 1-dim tensors
My proposed solution: makes `dim=-1` default choice, even for multidim tensors
|
triaged,module: numpy
|
low
|
Critical
|
588,729,041 |
godot
|
HeightMapShape collision does not update properly when the body is translated
|
Godot 3.2.1
Godot 3.1
Windows 10 64 bits
Following this issue (have a read for insights): https://github.com/Zylann/godot_heightmap_plugin/issues/129
Adding a `HeightMapShape` to a static body in the editor (through the use of tool scripts) works fine, until the body is moved to a non-origin position.
*All raycasts that are not going through the AABB it occupied at origin will always miss*.
This causes other plugins relying on raycasts to fail, and the Snap-to-Floor functionality also fails.
_This does not reproduce in a running game_, and it only happens with this type of shape.
It happens regardless of the body being moved or being loaded not at origin.
### Reproduction steps
Investigating what is actually happenning wasn't obvious at first, so I created a new plugin which allows to "see" the result of raycasts going from the camera to every visible pixel, allowing to visualise collisions directly:
https://github.com/Zylann/godot_collision_scanner
(see the issue linked above for more screenshots).
Here is a simplified project:
[HeightmapTranslated.zip](https://github.com/godotengine/godot/files/4389901/HeightmapTranslated.zip)
1) Open main.tscn: it contains a tool script which setups a heightmap with collider.
2) Go to ProjectSettings, and enable the `Collision scanner` plugin. The view will darken, and after a few seconds you should see raycast results, matching the terrain.
3) Now click on the `Terrain` node, and translate it by about 50 units along X or Z.
4) Move the view a bit to force an update of the collision viewer: the view will turn dark again, and after a few seconds you'll see updated raycast results: only those that went through "the AABB the terrain would cover if it was at origin" will show up, everything else will remain dark. This is wrong, it should still match the visuals.
Note: the body *did* call `PhysicsServer::body_set_state()`, and you can see collision actually *matches* the translated body, just not outside of this AABB. Something else must be missing.
|
bug,topic:physics
|
low
|
Minor
|
588,732,293 |
terminal
|
While in fullscreen mode, terminal does not resize properly when changing screen resolution
|
<!--
🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
I ACKNOWLEDGE THE FOLLOWING BEFORE PROCEEDING:
1. If I delete this entire template and go my own path, the core team may close my issue without further explanation or engagement.
2. If I list multiple bugs/concerns in this one issue, the core team may close my issue without further explanation or engagement.
3. If I write an issue that has many duplicates, the core team may close my issue without further explanation or engagement (and without necessarily spending time to find the exact duplicate ID number).
4. If I leave the title incomplete when filing the issue, the core team may close my issue without further explanation or engagement.
5. If I file something completely blank in the body, the core team may close my issue without further explanation or engagement.
All good? Then proceed!
-->
<!--
This bug tracker is monitored by Windows Terminal development team and other technical folks.
**Important: When reporting BSODs or security issues, DO NOT attach memory dumps, logs, or traces to Github issues**.
Instead, send dumps/traces to [email protected], referencing this GitHub issue.
If this is an application crash, please also provide a Feedback Hub submission link so we can find your diagnostic data on the backend. Use the category "Apps > Windows Terminal (Preview)" and choose "Share My Feedback" after submission to get the link.
Please use this form and describe your issue, concisely but precisely, with as much detail as possible.
-->
# Environment
```none
Windows build number: Microsoft Windows [Version 10.0.18363.752]
Windows Terminal version: v0.10.781.0
```
# Steps to reproduce
1. Enter full-screen mode with the `toggleFullscreen` comand
2. Change the screen resolution
# Expected behavior
The terminal resizes fill the entire screen with the new screen resolution dimensions (this may require adjusting the size in both dimensions).
# Actual behavior
The terminal is resized such that it fills the screen in one dimension and the black bars are introduced around the terminal in the other dimension to fill up the remaining space.
|
Issue-Bug,Area-UserInterface,Product-Terminal,Priority-3
|
low
|
Critical
|
588,793,229 |
flutter
|
_RenderButtonBarRow issues
|
_RenderButtonBarRow has a bunch of problems in its design:
1. It lays out its children multiple times, which is highly dubious. It violates our "single-pass" layout concept, and is thus potentially O(N^2). It's also dangerous in the case of descendants that do non-idempotent layout (e.g. that trigger behaviour when they are laid out).
2. It doesn't implement the intrinsic dimension logic (computeMinIntrinsicWidth et al) so it interacts poorly with widgets like IntrinsicWidth, which are commonly seen where ButtonBars are seen (e.g. in dialogs).
3. It overrides `constraints`, which is probably turning a monomorphic getter into a polymorphic getter and costing us performance across the board.
|
framework,f: material design,P2,team-design,triaged-design
|
low
|
Major
|
588,809,567 |
flutter
|
Support UIAutomator tests of Flutter module in existing android app
|
Hello.
I've been on investigation on flutter and made a decision that flutter is wonderful to cover legacy applications.
However, when it comes to integration testing (or acceptance testing) with legacy applications, I've been failing to figure out how.
Here is my situation.
I'd like to test "a new feature by flutter". But like I mentioned before, I failed to find a way.
Is it possible to run "flutter drive" on an existing android app that contains flutter module in it?
|
a: tests,tool,d: examples,a: existing-apps,P3,team-tool,triaged-tool
|
low
|
Critical
|
588,837,641 |
pytorch
|
TorchScript can't use lists in conditionals
|
## 🐛 Bug
This is a reasonably common pattern in Python for testing if a list is empty. Seems like TorchScript could support it by just teaching the frontend to translate it to `if len(ls) != 0:`
## To Reproduce
Steps to reproduce the behavior:
```python
import torch.utils.collect_env
torch.utils.collect_env.main()
class VerySimpleModel(torch.nn.Module):
def forward(self):
if [1]:
return 1
return 0
em = VerySimpleModel()
print(em())
sm = torch.jit.script(em)
print(sm())
```
```
~/work/venv-pt3-nightly/lib/python3.7/site-packages/torch/jit/_recursive.py in create_methods_from_stubs(concrete_type, stubs)
277 rcbs = [m.resolution_callback for m in stubs]
278 defaults = [get_default_args(m.original_method) for m in stubs]
--> 279 concrete_type._create_methods(defs, rcbs, defaults)
280
281 def create_script_module(nn_module, stubs_fn, share_types=True):
RuntimeError:
Could not cast value of type List[int] to bool:
```
(Note: it also doesn't show where in the TorchScript code this is happening.)
## Expected behavior
Prints
```
1
1
```
## Environment
Collecting environment information...
PyTorch version: 1.5.0+cpu
Is debug build: No
CUDA used to build PyTorch: None
OS: Fedora release 30 (Thirty)
GCC version: (GCC) 9.1.1 20190503 (Red Hat 9.1.1-1)
CMake version: version 3.14.4
Python version: 3.7
Is CUDA available: No
CUDA runtime version: No CUDA
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
Versions of relevant libraries:
[pip3] numpy==1.17.0
[pip3] pytorch-pretrained-bert==0.6.2
[pip3] torch==1.2.0
[pip3] torchtext==0.4.0
[conda] Could not collect
## Additional context
The JIT team are the coolest people I know.
cc @suo
|
oncall: jit,triaged,enhancement
|
low
|
Critical
|
588,846,667 |
TypeScript
|
Add Command-Line Syntax-Only Validation
|
## Search Terms
Keywords I searched:
- Syntax
- Validation
## Suggestion
I would like to see a command line option that makes `tsc` only validate the syntax of TypeScript (and possibly JavaScript) files within the project, but does not perform type checking. It would not emit any files.
## Use Cases
- (My current use case) TypeScript generated by code generation programs that skips generation of certain symbols, or needs to import symbols from other modules where those modules may not be available yet, so their types cannot be known. The generated files would almost always have type-related errors that cannot reliably be automatically fixed, so just validation of the syntax alone would be helpful.
- Certain CI/CD workflows could benefit from the lower overhead of just checking syntax only in large projects for every commit, and only performing type checking for commits that are expected to result in a build.
## Examples
The command
```bash
tsc --noTypeCheck index.ts
```
could return the normal errors and exit code that it would if the syntax were incorrect, but never check typing.
## 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
|
588,894,064 |
deno
|
Consume source maps from JavaScript files
|
Current Deno will remap stack traces on files that have been compiled by Deno, as Deno caches both the original file and the emitted map, and loads those from the cache. When consuming JavaScript with source maps from other sources (like already compiled code) Deno does not handle the source map and ignore source map pragma included in files.
Deno should support both external source maps and inline source maps when fetching modules and store any source maps in the cache to be used when remapping stack traces.
This will have the side-effect of supporting the generation of source maps for bundles generated by `deno bundle`.
One consideration is that source maps optionally can include the original sources. If they do, then when emitting an error, the original source line is available, but if they don't, we would not be able to log-out/resolve that original source line.
|
feat,cli
|
medium
|
Critical
|
588,955,311 |
TypeScript
|
Syntax error in a union type which contains a callback type
|
<!-- 🚨 STOP 🚨 𝗦𝗧𝗢𝗣 🚨 𝑺𝑻𝑶𝑷 🚨
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.
-->
**TypeScript Version:** 3.8.3
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:**
- algebraic data types
**Expected behavior:**
Both `b` and `c` definitions are correct
**Actual behavior:**
Syntax errors:
Texts of errors (from left to right)
```
Exported variable 'c' has or is using private name ''.(4025)
Type expected.(1110)
```
```
';' expected.(1005)
```
```
'(' expected.(1005)
```
<!-- Did you find other bugs that looked similar? -->
**Related Issues:**
**Code**
```ts
type ArrFunc = () => Array<number>;
let b: Array<number> | number | ArrFunc;
let c: Array<number> | number | () => Array<number>;
```
<details><summary><b>Output</b></summary>
```ts
"use strict";
let b;
let c;
Array();
```
</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/#)
|
Suggestion,Experience Enhancement
|
low
|
Critical
|
588,962,752 |
rust
|
Implement conversion traits for usize/isize together with a portability lint
|
[This PR](https://github.com/rust-lang/rust/pull/37423) was closed 3.5 years ago just prior to the development of the [portability lint RFC](https://github.com/rust-lang/rfcs/blob/master/text/1868-portability-lint.md). In light of the portability lint RFC having been accepted, we should be able to resurrect that PR.
This issue tracks that work. I've submitted it to TWiR's CFP.
|
T-lang,T-libs-api
|
medium
|
Critical
|
588,997,696 |
TypeScript
|
Clarify class property initializer behavior
|
## The problem
Briefly, if a class property isn't initialized with the declaration, it will be initialized as `undefined`, no matter how you initialize it in constructor. This will lead to v8 performance issue because of property type transition in many cases.
So I think it's very important to mention the behavior difference in documentation.
## Documentation difference
In [Class field declarations](https://github.com/tc39/proposal-class-fields#fields-without-initializers-are-set-to-undefined) proposal from TC39, it's very clearly mentioned that "Fields without initializers are set to `undefined`."
Thus in TS handbook, there is nowhere mentioned how the field declaration without initializer will be handled. Actually the documentation lead us to believe it's the same result to initialize a property in declaration or in constructor. E.G., in the [Classes default example](https://www.typescriptlang.org/docs/handbook/classes.html#classes), it says:
> In the last line we construct an instance of the Greeter class using new. This calls into the constructor we defined earlier, creating a new object with the Greeter shape, and running the constructor to initialize it.
But actually it's like the TC39 proposal, **if you don't assign a default value in declaration, the property will be initialized as `undefined` in constructor (before your actual initialize code).**
Take a look at a simple example
```
class Test {
x: number;
constructor(x: number) {
this.x = x;
}
}
```
With tsc, this will be generated as the following js
```
"use strict";
var Test = /** @class */ (function () {
function Test(x) {
Object.defineProperty(this, "x", {
enumerable: true,
configurable: true,
writable: true,
value: void 0
});
this.x = x;
}
return Test;
}());
```
As you can see, property `this.x` is firstly defined with `void 0` (undefined) and then assigned with `x` parameter.
## Why should we care
In the previous example, there is a type transition for the property `x` which is sadly not a good idea in v8. I assume it cause the Test instances to be polymorphic or it cause the Test instances generate a new hidden class. Anyway, **the property access performance drops significantly after the construction.**
Test case:
[index.js.zip](https://github.com/microsoft/TypeScript/files/4392502/index.js.zip)
From the simple test case with a loop of Vec3 calculations, we can see the performance difference
1. Initialize in declaration(Vec3_104): 4-5ms
2. Initialize in constructor(Vec3_110): 28-32ms
From more complexe use case in our engine's particle system simulation, the fps drops from 60fps to 30fps.
## Conclusion
I'd like to suggest a very clear description of different initializer behaviors in TypeScript documentation, so that developers know that they really need to initialize in declaration.
If possible, we'd like TSC to improve the logic:
If developers are using constructor to initialize properties, then use that value for `defineProperty` directly, or use a type compatible value for `defineProperty`.
We have also submit [an issue](https://github.com/babel/babel/issues/11342) to babel which have similar behavior.
|
Docs
|
low
|
Major
|
589,042,781 |
flutter
|
Add support for Ascender, Descender, XHeight alignment widgets.
|
Reopening #25018 since it was closed without resolution.
When laying out Text widgets, it is sometimes necessary to align them by the top of the capital letter or the bottom of lowercase letters like 'y'. There is a Baseline widget, but no way to align text by its ascender, descender, or x-height.
For a helpful diagram, see:
https://en.wikipedia.org/wiki/Ascender_(typography)
It is insufficient to simply have a getter for font metrics. For example, you can't have negative padding in a widget, so you couldn't align a Text object to the top of its container by simply knowing the ascent. You could wrap it in another widget, but that would be a lot of boilerplate, pollute the tree, and make it unnecessarily difficult to read the code for what should be a simple property as follows:
`Text(
'Sample Text',
upperConstraint: TextConstraint.CapHeight, // Ascent, XHeight, etc
lowerConstraint: TextConstraint.Baseline, // Descent, Bottom, etc
)`
|
c: new feature,framework,a: typography,c: proposal,P3,team-framework,triaged-framework
|
low
|
Minor
|
589,056,567 |
flutter
|
build apk --release crash. Execution failed for task ':app:minifyReleaseWithProguard'
|
cant build/run app and flutter doctor shows everything is ok:
-----------------------------
flutter doctor -v
[✓] Flutter (Channel stable, v1.12.13+hotfix.8, on Linux, locale en_US.UTF-8)
• Flutter version 1.12.13+hotfix.8 at /home/USER/Android/flutter
• Framework revision 0b8abb4724 (6 weeks ago), 2020-02-11 11:44:36 -0800
• Engine revision e1e6ced81d
• Dart version 2.7.0
[✓] Android toolchain - develop for Android devices (Android SDK version 29.0.3)
• Android SDK at /home/USER/Android/Sdk
• Android NDK location not configured (optional; useful for native profiling
support)
• Platform android-29, build-tools 29.0.3
• Java binary at: /opt/android-studio/jre/bin/java
• Java version OpenJDK Runtime Environment (build
1.8.0_212-release-1586-b4-5784211)
• All Android licenses accepted.
[✓] Android Studio (version 3.6)
• Android Studio at /opt/android-studio
• Flutter plugin version 44.0.2
• Dart plugin version 192.7761
• Java version OpenJDK Runtime Environment (build
1.8.0_212-release-1586-b4-5784211)
|
c: crash,platform-android,tool,t: gradle,a: release,P2,team-android,triaged-android
|
low
|
Critical
|
589,059,162 |
react
|
Batching makes it difficult to perform imperative actions like focus
|
React version: 16.9.5
## Steps To Reproduce
1. Overwrite batched updates with the following code: `ReactDOM.unstable_batchedUpdates = callback => callback()`
2. Batched updates aren't actually overwritten.
## Reasoning
I recognize that this may not be classified as bug because it isn't a documented feature but I have tried to search for a different solution but to no avail. Fixing this behavior can open a new way of using React. I tried writing on [Stack Overflow](https://stackoverflow.com/questions/60852219/disable-unstable-batchedupdates-on-event-handlers-in-react) and writing to [@gaearon](https://twitter.com/antoniostoilkov/status/1242789762307588097).
I have a number of arguments which support the disabling of batched updates in event handlers and in effects initialization. If anybody is willing to read a document and consider this scenario I am willing to write an RFC.
|
Type: Discussion
|
medium
|
Critical
|
589,063,227 |
pytorch
|
RuntimeError during converting Reformer model to TorchScript
|
## 🐛 Bug
<!-- A clear and concise description of what the bug is. -->
## To Reproduce
Steps to reproduce the behavior:
1. install reformer-pytorch from https://github.com/lucidrains/reformer-pytorch
2. install transformers
3. Run script
```python
import torch
import torch.nn as nn
import torch.nn.functional as F
import numpy as np
from reformer_pytorch import Reformer, ReformerLM
from transformers import BertTokenizer, AdamW
tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
tokenizer.max_len = 40960
model = ReformerLM(
num_tokens= tokenizer.vocab_size,
dim = 768,
depth = 12,
max_seq_len = tokenizer.max_len,
heads = 16,
lsh_dropout = 0.1,
emb_dim = 256, # embedding factorization for further memory savings
causal = False, # auto-regressive or not
bucket_size = 64, # average size of qk per bucket, 64 was recommended in paper
n_hashes = 4, # 4 is permissible per author, 8 is the best but slower
ff_chunks = 200, # number of chunks for feedforward layer, make higher if there are memory issues
weight_tie = False, # tie parameters of each layer for no memory per additional depth
attn_chunks = 8, # process lsh attention in chunks, only way for memory to fit when scaling to 16k tokens
num_mem_kv = 128, # persistent learned memory key values, from all-attention paper
twin_attention = False, # both branches of the reversible network will be attention
use_full_attn = False, # use full self attention, for comparison
full_attn_thres = 128, # use full attention if context length is less than set value
use_scale_norm = True, # use scale norm from 'Transformers without tears' paper
#fixed_position_emb = True
axial_position_emb = True,
axial_position_shape = (640, 64),
axial_position_dims = (128, 128)
)
x = torch.randint(0, 30522, (2, 40960))
model.eval()
traced_cell = torch.jit.trace(model, x)
```
```
RuntimeError Traceback (most recent call last)
<ipython-input-1-2cc80d9cd643> in <module>
46 #print(datetime.datetime.now())
47 #my_module = torch.jit.script(model)
---> 48 traced_cell = torch.jit.trace(model, x)
49 print("ok")
/opt/anaconda/envs/torch-nigtly-reformer/lib/python3.6/site-packages/torch/jit/__init__.py in trace(func, example_inputs, optimize, check_trace, check_inputs, check_tolerance, _force_outplace, _module_class, _compilation_unit)
879 return trace_module(func, {'forward': example_inputs}, None,
880 check_trace, wrap_check_inputs(check_inputs),
--> 881 check_tolerance, _force_outplace, _module_class)
882
883 if (hasattr(func, '__self__') and isinstance(func.__self__, torch.nn.Module) and
/opt/anaconda/envs/torch-nigtly-reformer/lib/python3.6/site-packages/torch/jit/__init__.py in trace_module(mod, inputs, optimize, check_trace, check_inputs, check_tolerance, _force_outplace, _module_class, _compilation_unit)
1041 else:
1042 _check_trace([inputs], func, check_trace_method,
-> 1043 check_tolerance, _force_outplace, True, _module_class)
1044 finally:
1045 torch.jit._trace_module_map = old_module_map
/opt/anaconda/envs/torch-nigtly-reformer/lib/python3.6/site-packages/torch/autograd/grad_mode.py in decorate_context(*args, **kwargs)
13 def decorate_context(*args, **kwargs):
14 with self:
---> 15 return func(*args, **kwargs)
16 return decorate_context
17
/opt/anaconda/envs/torch-nigtly-reformer/lib/python3.6/site-packages/torch/jit/__init__.py in _check_trace(check_inputs, func, traced_func, check_tolerance, force_outplace, is_trace_module, _module_class)
673 traced_outs = run_mod_and_filter_tensor_outputs(traced_func, inputs, 'trace')
674 fn_outs = run_mod_and_filter_tensor_outputs(func, inputs, 'Python function')
--> 675 if compare_outputs(traced_outs, fn_outs, 'Python function'):
676 check_outs = run_mod_and_filter_tensor_outputs(check_mod_func, inputs, 'repeated trace')
677 compare_outputs(traced_outs, check_outs, 'repeated trace')
/opt/anaconda/envs/torch-nigtly-reformer/lib/python3.6/site-packages/torch/jit/__init__.py in compare_outputs(original, reference, match_what)
661 ref = ref.dequantize()
662 torch.testing.assert_allclose(orig.double(), ref.double(), rtol=check_tolerance,
--> 663 atol=torch.testing._get_default_tolerance(orig, ref)[1])
664 except AssertionError as e:
665 maybe_warn_nondeterministic()
/opt/anaconda/envs/torch-nigtly-reformer/lib/python3.6/site-packages/torch/testing/__init__.py in assert_allclose(actual, expected, rtol, atol, equal_nan)
28 rtol, atol = _get_default_tolerance(actual, expected)
29
---> 30 close = torch.isclose(actual, expected, rtol, atol, equal_nan)
31 if close.all():
32 return
RuntimeError: [enforce fail at CPUAllocator.cpp:64] . DefaultCPUAllocator: can't allocate memory: you tried to allocate 20002897920 bytes. Error code 12 (Cannot allocate memory)
```
## Expected behavior
Get model, converted to TorchScript
## Environment
PyTorch version: 1.5.0.dev20200305
Is debug build: No
CUDA used to build PyTorch: 10.1
OS: Ubuntu 18.04 LTS
GCC version: (Ubuntu 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:
GPU 0: Quadro RTX 8000
GPU 1: Quadro RTX 8000
Nvidia driver version: 440.33.01
cuDNN version: Could not collect
Versions of relevant libraries:
[pip] numpy==1.18.1
[pip] reformer-pytorch==0.19.2
[pip] torch==1.5.0.dev20200305
[pip] torchvision==0.6.0a0+15aa7e2
[conda] blas 1.0 mkl
[conda] mkl 2019.5 281 conda-forge
[conda] mkl-service 2.3.0 py36h516909a_0 conda-forge
[conda] mkl_fft 1.1.0 py36hc1659b7_1 conda-forge
[conda] mkl_random 1.1.0 py36hb3f55d8_0 conda-forge
[conda] pytorch 1.5.0.dev20200305 py3.6_cuda10.1.243_cudnn7.6.3_0 pytorch-nightly
[conda] reformer-pytorch 0.19.2 pypi_0 pypi
[conda] torchvision 0.6.0.dev20200305 py36_cu101 pytorch-nightly
- PyTorch Version: 1.5.0.dev20200305
- OS: Ubuntu 18.04 LTS
- How you installed PyTorch: Conda
- Python version: 3.6
- CUDA/cuDNN version: 1.5.0.dev20200305/not collect
- GPU models and configuration: RTX 8000 x 2
## Additional context
I trying to convert Reformer model, witch storing on cpu Into TorchScript. But, during converting model allocate all RAM(128 GB) and fail with RuntimeError.
If I trying to infer model in torch, all going ok, and model allocate only 7GB RAM with batch size 2.
cc @suo
|
triage review,oncall: jit,triaged
|
low
|
Critical
|
589,071,324 |
pytorch
|
`torch.bernoulli` and `torch.rand` to support `dtype=torch.bool` kwarg (and sypport operation of `tensor.uniform_` on bool tensors)
|
```
----> 1 torch.bernoulli(torch.rand(10), dtype = torch.bool)
TypeError: bernoulli() received an invalid combination of arguments - got (Tensor, dtype=torch.dtype), but expected one of:
* (Tensor input, torch.Generator generator, Tensor out)
* (Tensor input, float p, torch.Generator generator)
```
Related: https://github.com/pytorch/pytorch/issues/31779
cc @vincentqb @fritzo @neerajprad @alicanb @vishwakftw
|
module: distributions,triaged
|
low
|
Critical
|
589,071,493 |
TypeScript
|
[Feature] Utility type for T[keyof T] or new keyword valueof
|
## Search Terms
typeof keyof utlity type
## Suggestion
**Alternative 1:**
Create a new Utility type expressing `T[keyof T]`
Example `Value<T> = T[keyof T]`.
**Alternative 2:**
Create a new keyword equivalent to `T[keyof T]`
Example `valueof T = T[keyof T]`
## Use Cases
While `T[keyof T]` is short and concise, from my experience a lot of developers new to TypeScript usually have no idea what this means or why it's written like that when they first see it.
I also feel that it's not as easy to Google a the answer to that, in comparison to established utility and advanced types like Pick, Omit, Record, etc. which all have their own entries in the TS documentation.
Additionally Object.keys and Object.values are similar concepts.
The presence of a `keyof` keyword would suggest a `valueof` should also exist, yet it's missing and one must employ a "special" approach to replicate.
A simple solution is to write my own custom type for the project, but that just introduces additional abstraction and the knowledge/practice isn't transferable to other projects, without implementing the same custom type somewhere in the code first.
And lastly, and this is _not_ a big issue, but I feel it's redundant having to write the type for T twice in on the same line just a few symbols apart.
With short names such as T it's okay, but with longer type names the code becomes more cumbersome to read.
## Examples
```
type T = {
a: string;
b: number
c: number[]
...etc
}
// Value<T> = string | number | number[]...
// valueof T = string | number | number[]...
```
## 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).
Perhaps some conflict on point 8...
|
Suggestion,Awaiting More Feedback
|
medium
|
Major
|
589,073,665 |
pytorch
|
Support bool input tensors for argmax / argmin / sort / topk and other functions
|
argmax and other functions make sense for finding position of True element. Functions working with uint8 should likely be able to handle torch.bool as well.
```
In [16]: torch.__version__
Out[16]: '1.5.0.dev20200310'
```
```
In [11]: torch.ones(3, dtype = torch.bool).argmax()
---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
~/convasr/ot/vis.py in <module>
----> 1 torch.ones(3, dtype = torch.bool).argmax()
RuntimeError: "argmax_cpu" not implemented for 'Bool'
In [12]: torch.ones(3, dtype = torch.bool).argmin()
---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
~/convasr/ot/vis.py in <module>
----> 1 torch.ones(3, dtype = torch.bool).argmin()
RuntimeError: "argmin_cpu" not implemented for 'Bool'
In [13]: torch.ones(3, dtype = torch.bool).argsort()
---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
~/convasr/ot/vis.py in <module>
----> 1 torch.ones(3, dtype = torch.bool).argsort()
RuntimeError: _th_sort not supported on CPUType for Bool
In [14]: torch.ones(3, dtype = torch.bool).topk(1)
---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
~/convasr/ot/vis.py in <module>
----> 1 torch.ones(3, dtype = torch.bool).topk(1)
RuntimeError: "topk_cpu" not implemented for 'Bool'
In [15]: torch.ones(3, dtype = torch.bool).sort()
---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
~/convasr/ot/vis.py in <module>
----> 1 torch.ones(3, dtype = torch.bool).sort()
RuntimeError: _th_sort not supported on CPUType for Bool
```
|
triaged,enhancement,module: reductions
|
low
|
Critical
|
589,100,135 |
rust
|
`ninja` is only used in Windows builds on CI.
|
AFAICT this is the only thing that adds `--enable-ninja`, and it's Windows-only:
https://github.com/rust-lang/rust/blob/b9ac29169ac541a64471b83d8019cd0b442f8065/src/ci/scripts/install-ninja.sh#L9-L16
We also autodetect `ninja` and use it, but only for MSVC specifically (#43767 / #44084).
On top of being (supposedly) faster than relying on `make`, IME `ninja` also has less log spew when building LLVM (but it's unclear if it would help on CI, see #70468).
cc @rust-lang/infra
|
A-LLVM,C-enhancement,T-infra
|
low
|
Minor
|
589,120,669 |
pytorch
|
Fail to load torch script model
|
## 🐛 Bug
<!-- A clear and concise description of what the bug is. -->
When following this tutorial: https://pytorch.org/tutorials/advanced/cpp_export.html
Step 3, executing this command with my model:
./example-app <path_to_model>/traced_resnet_model.pt
produces an error when trying to load model.
## To Reproduce
Steps to reproduce the behavior:
1. Follow tutorial steps exactly, except
2. Change load model line to: std::shared_ptr<torch::jit::script::Module> module = torch::jit::load(argv[1]); (found in another issue because this tutorial is outdated)
3. Save mobilenet model from pytorch as traced model using torch.jit.trace
4. comment out try catch block in script to see what exactly the error is when loading model
fails with this error:
libc++abi.dylib: terminating with uncaught exception of type c10::Error: istream reader failed: reading file. (validate at /Users/ishaghodgaonkar/pytorch/caffe2/serialize/istream_adapter.cc:32)
frame #0: c10::Error::Error(c10::SourceLocation, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) + 64 (0x10dcf24e0 in libc10.dylib)
frame #1: caffe2::serialize::IStreamAdapter::validate(char const*) const + 166 (0x10b7f8256 in libcaffe2.dylib)
frame #2: caffe2::serialize::IStreamAdapter::read(unsigned long long, void*, unsigned long, char const*) const + 159 (0x10b7f833f in libcaffe2.dylib)
frame #3: caffe2::serialize::PyTorchStreamReader::init() + 222 (0x10b7f43be in libcaffe2.dylib)
frame #4: caffe2::serialize::PyTorchStreamReader::PyTorchStreamReader(std::__1::unique_ptr<caffe2::serialize::ReadAdapterInterface, std::__1::default_delete<caffe2::serialize::ReadAdapterInterface> >) + 109 (0x10b7f51cd in libcaffe2.dylib)
frame #5: torch::jit::load(std::__1::unique_ptr<caffe2::serialize::ReadAdapterInterface, std::__1::default_delete<caffe2::serialize::ReadAdapterInterface> >, c10::optional<c10::Device>) + 522 (0x1098ad9da in libtorch.1.dylib)
frame #6: torch::jit::load(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, c10::optional<c10::Device>) + 93 (0x1098adbdd in libtorch.1.dylib)
frame #7: main + 127 (0x109123f4f in example-app)
frame #8: start + 1 (0x7fff765b73d5 in libdyld.dylib)
frame #9: 0x0 + 2 (0x2 in ???)
Abort trap: 6
## Expected behavior
Load the model without runtime error, print 'ok'
## Environment
PyTorch version: N/A
Is debug build: N/A
CUDA used to build PyTorch: N/A
OS: Mac OSX 10.14.6
GCC version: Could not collect
CMake version: version 3.13.4
Python version: 2.7
Is CUDA available: N/A
CUDA runtime version: Could not collect
GPU models and configuration: Could not collect
Nvidia driver version: Could not collect
cuDNN version: Could not collect
Versions of relevant libraries:
[pip] numpy==1.17.2
[pip] torch==1.4.0
[pip] torch-extension==0.1
[pip] torchsummary==1.5.1
[pip] torchvision==0.5.0
[conda] blas 1.0 mkl
[conda] mkl 2019.1 144
[conda] mkl-include 2019.4 233
[conda] mkl-service 1.1.2 py37hfbe908c_5
[conda] mkl_fft 1.0.6 py37h27c97d8_0
[conda] mkl_random 1.0.2 py37h27c97d8_0
[conda] pytorch 1.4.0 py3.7_0 pytorch
[conda] torch 1.1.0a0+866c4e3 pypi_0 pypi
[conda] torchvision 0.1.7 pypi_0 pypi
## Additional context
I also tried using torch::jit::script instead of torch::jit::trace as mentioned in another issue. Same error.
Code:
#include <torch/script.h> // One-stop header.
#include <iostream>
#include <memory>
int main(int argc, const char* argv[]) {
if (argc != 2) {
std::cerr << "usage: example-app <path-to-exported-script-module>\n";
return -1;
}
// Deserialize the ScriptModule from a file using torch::jit::load().
std::shared_ptr<torch::jit::script::Module> module = torch::jit::load(argv[1]);
std::cout << "ok\n";
}
cc @suo @SsnL
|
oncall: jit,triaged
|
low
|
Critical
|
589,122,408 |
pytorch
|
Apple Review Rejected. ITMS-90338: Non-public API usage In Pytorch For iOS
|
I use LibTorch(1.4) in pod. But When I upload the ipa binary file, Apple Review Rejected. The Email content is as below:
ITMS-90338: Non-public API usage - The app references non-public symbols in Easy Math: _daxpy_, _dcopy_, _ddot_, _dgemm_, _dgemv_, _dger_, _dscal_, _saxpy_, _scopy_, _sdot_, _sgemm_, _sgemv_, _sger_, _sscal_. If method names in your source code match the private Apple APIs listed above, altering your method names will help prevent this app from being flagged in future submissions. In addition, note that one or more of the above APIs may be located in a static library that was included with your app. If so, they must be removed. For further information, visit the Technical Support Information at http://developer.apple.com/support/technical/
And I found these api in THBlas.cpp. How can I solve this problem. Hope receive your reply.
|
triaged,module: ios
|
low
|
Minor
|
589,129,268 |
rust
|
CI logs should be split by build steps, or stages at the very least.
|
Right now most of the build is under a single "run the build" CI step, which results in massive logs (e.g. #70468), even if for common tasks like looking at a failure, only a small part is relevant.
At the very least we could do something like this to split it up:
* `./x.py build --stage 0`
* `./x.py build --stage 1`
* `./x.py build`
* `./x.py test`
This would be trickier with `distcheck`, I suppose, not sure how relevant that is.
It might also be slightly slower, so we should measure it.
cc @rust-lang/infra
|
C-enhancement,T-infra
|
low
|
Critical
|
589,134,736 |
rust
|
#[repr(transparent)] for str and String
|
Similarly to https://github.com/rust-lang/rust/issues/52976, I'd suggest that `str` and `String` should have a guaranteed transparent representation of `[u8]` and `Vec<u8>` correspondingly.
They are a newtype wrappers of these types under the hood anyway, and I don't see any reasons why they would need to behave differently in the future, and making them different probably would become a breaking change anyway as it would break things like `String::from_raw_parts` constructed from raw parts of the `Vec`.
While `String::from_raw_parts` already covers most common cases, there are still other cases where it would be desired to have guaranteed representation.
One such example is constructing `Rc<String>` from a `Rc<Vec<u8>>` where a library might want to check UTF-8 validity before construction and then `Rc::from_raw` + `Rc::into_raw` one type into another without actually copying the data.
Currently this is not possible to do safely because it's not guaranteed that the `*const Vec<u8>` returned from the first `into_raw` is compatible with `*const String` accepted by the second.
While this particular case could be alleviated by adding a special method, there are lots of other containers where such compatibility between these types might be desirable, and I don't see any reasons why it couldn't be guaranteed at a lower level.
|
T-libs-api,A-repr
|
low
|
Major
|
589,184,038 |
go
|
cmd/asm: update where NOFRAME is valid
|
The current comment for `NOFRAME` in `runtime/textflag.h` says:
```
// TODO(mwhudson): only implemented for ppc64x at present.
```
It's been implemented for lots (all?) architectures since.
Check our coverage and update (delete) this comment.
@mwhudson @zhangfannie
|
Documentation,NeedsFix,compiler/runtime
|
low
|
Minor
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.