id
int64 393k
2.82B
| repo
stringclasses 68
values | title
stringlengths 1
936
| body
stringlengths 0
256k
β | labels
stringlengths 2
508
| priority
stringclasses 3
values | severity
stringclasses 3
values |
|---|---|---|---|---|---|---|
627,940,534 |
godot
|
Quick Open Script search should show exact matches on top
|
**Godot version:**
3.2.1-stable_win64
**OS/device including version:**
<!-- Specify GPU model, drivers, and the backend (GLES2, GLES3, Vulkan) if graphics-related. -->
Windows 10 64
**Issue description:**
Searching in the Quick Open Script window is not sorting the scripts properly. It should prioritize exact (and complete) match of script names before not exact matches.
**Steps to reproduce:**
use the search option in the Quick Open Script window
|
topic:editor
|
low
|
Minor
|
627,947,474 |
rust
|
Confusing Error for writeln! and write!
|
```rust
let test = 0;
writeln!("{}", test);
```
Gives the error
```
error: format argument must be a string literal
--> src/main.rs:3:20
|
3 | writeln!("{}", test);
| ^^^^
|
help: you might be missing a string literal to format with
|
3 | writeln!("{}", "{}", test);
| ^^^^^
error[E0599]: no method named `write_fmt` found for reference `&'static str` in the current scope
--> src/main.rs:3:5
|
3 | writeln!("{}", test);
| ^^^^^^^^^^^^^^^^^^^^^ method not found in `&'static str`
|
= note: this error originates in a macro (in Nightly builds, run with -Z macro-backtrace for more info)
```
The first error should be removed because it's misleading.
[Playground](https://play.rust-lang.org/?version=nightly&mode=debug&edition=2018&gist=87254cc2a9e95c9a471526270f19f9e1)
|
A-diagnostics,P-low,T-compiler,D-papercut
|
low
|
Critical
|
627,949,334 |
electron
|
setIgnoreMouseEvents for BrowserView
|
### 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
My browser is using BrowserView for displaying external content and it's currently impossible to draw DOM Elements over BrowserView. So I decided to use also BrowserView for browser popups. But it also has some drawbacks like higher resources usage, since there are multiple BrowserView instances for displaying some popups.
Screenshot:
Pink - outline of BrowserView
BrowserView per popup:
### Proposed Solution
Add `setIgnoreMouseEvents` method just like in BrowserWindow (I'm not using BrowserWindows for popups, as it still has problems with transparency on Linux and weird showing animations on Windows). That way I could make one transparent BrowserView with popups and I would be able to track mouse position and toggle `setIgnoreMouseEvents` respectively.
Pink - outline of BrowserView
Cyan - the global overlay based on BrowserView
Screenshot:
Global overlay based on BrowserView:
I hope my paint skills will help you understand the problem better :)
### Alternatives Considered
Separate BrowserViews for each popup.
I think that could also help solving problems with migrating from `<webview>` to BrowserView
|
enhancement :sparkles:
|
medium
|
Major
|
627,960,322 |
pytorch
|
[JIT] nn.ModuleList loses None objects inside it after scripting
|
## π Bug
<!-- A clear and concise description of what the bug is. -->
## To Reproduce
Executing torch.jit.script over my model is working however it returns a model that fails at runtime.
Looking deeply the nn.ModuleList is loosing None elements from the Modulelist.
Here, above I attach a code for reproducing the error:
```
import os
import sys
import torch.nn as nn
import torch.nn.functional as F
import torch
from torchvision import transforms
from PIL import Image
class TestBlock(nn.Module):
def __init__(self):
super(TestBlock, self).__init__()
layers = []
layers.append(None)
layers.append(None)
layers.append(nn.Conv2d(3, 64, kernel_size=3, stride=2, padding=1,
bias=False))
self.layer = nn.ModuleList(layers)
def forward(self,x):
for aux in self.layer:
print("ENTER")
if aux is not None:
x = aux(x)
print("Not None")
return x
```
Creating model and tracing it:
```
model=TestBlock()
traced_cell=torch.jit.script(model)
```
Testing model with an image:
```
img = Image.open("test.png")
my_transforms = transforms.Compose([transforms.Resize((1002,1002)),
transforms.ToTensor(),
transforms.Normalize(
[0.485, 0.456, 0.406],
[0.229, 0.224, 0.225])])
img_input= my_transforms(img).unsqueeze(0).cpu()
res=model(img_input)
```
This outputs the next:
```
ENTER
ENTER
ENTER
Not None
```
Traced version output:
```
res=traced_cell(img_input)
```
```
ENTER
Not None
```
## Expected behavior
Get same output as original model
cc @suo
|
triage review,oncall: jit
|
low
|
Critical
|
627,975,492 |
pytorch
|
Make torch.cross `dim` parameter work intuitively
|
Hi PyTorch team,
Thanks a lot for maintaining such a great library.
**Problem**
`torch.cross` has `dim` arg, and by default `dim=-1`. However, it turns out, that if `dim` is not set, that dim is chosen to be the first dimension of size 3. This is very weird for me as a user and I spent half a day looking for a bug because I did not read the doc in full.
**Solution**
Either to set default `dim` to `None` pushing the reader to go further in docs to understand what it means. A better way would be to remove "first dimension of size 3" thing and just use `dim` parameter.
cc @ezyang @gchanan @mruberry @rgommers @SsnL
|
module: bc-breaking,triaged,module: numpy,topic: bc breaking
|
low
|
Critical
|
627,976,545 |
node
|
Inconsistent http stream behavior
|
Hey there,
* **Version**: 14.3.0
* **Platform**: MacOS Catalina
* **Subsystem**:
### What steps will reproduce the bug?
We have web-server and we want to limit the amount of data that can be sent by the client.
In the code below, I've tried to implement such a server (it counts the number of chunks) and client that sends a stream of infinity size to the server.
I assume that after the 10th chunk server should send the response to the client which will lead to 2 things on the client-side: `error` event on `request` object (because of premature stream close) and `response` callback been called (because the server sends a response to the client).
However, it turns out that this code prints `EPIPE` error in 100% of cases which is correct, but it prints response only in ~50% of cases.
If you change `res.end()` to `res.end("something")` (so that response will contain a body) then it will always print error and response, so the client will always be able to get a response from the server.
```
const http = require('http');
const { Readable } = require('stream');
const server = new http.Server();
server.on('request', (req, res) => {
let i = 0;
req.on('data', chunk => {
i++;
if (i === 10) {
res.statusCode = 413;
res.end(); // try to change it to res.end("too big")
}
});
});
server.listen(3000, makeRequest);
function makeRequest() {
const request = http.request(
'http://localhost:3000',
{method: 'POST'},
response => {
console.log("response finished", response.statusCode);
server.close();
}
);
request.on('error', err => {
console.log("request error", err);
});
Readable.from(generate()).pipe(request);
}
function* generate() {
while(true) { yield "value"; }
}
```
### What is the expected behavior?
I expect to see consistent behavior, so the client should be able to always get a response.
|
http,macos
|
low
|
Critical
|
627,986,092 |
ant-design
|
Antd 4.x Tree When virtual scroll true and height prop set, when updated with new data brings scrollbar to the bottom
|
- [ ] I have searched the [issues](https://github.com/ant-design/ant-design/issues) of this repository and believe that this is not a duplicate.
### What problem does this feature solve?
The selected Treenode should stay as is i.e in selected state while the other nodes not in view port get added and user doesn't feel about the new data. This works fine without virtual scroll.
### What does the proposed API look like?
I dont have much knowledge on the code yet. I am rasing this as a feature request which is a must with virtual scroll as without this the virtual scroll may be half useful
CodeSandbox to reproduce:
https://codesandbox.io/s/infallible-hofstadter-vrnyw?file=/MyTree.jsx
<!-- generated by ant-design-issue-helper. DO NOT REMOVE -->
|
π Bug,Inactive
|
low
|
Major
|
627,986,457 |
godot
|
[TRACKER] C++ modules related issues
|
**Godot version:**
3.2, 4.0
**Issue description:**
There are some existing issues related to module dependencies which may cause run-time and compile-time errors, or need better user-friendly interface to ensure that certain modules cannot be disabled, or whether the usage of modules may reveal hidden built-in bugs (which can be either workaround on the module level or fixed on the core level, so these are not necessarily the responsibility of module authors, but which nonetheless worth to link).
## Bugs
```[tasklist]
- [x] #36091
- [x] #32217
- [x] #32216
- [x] #39186
- [x] #38759
- [ ] #37301
- [ ] #37234
- [x] #36582
- [ ] #35788
- [x] #35740
- [ ] #33659
- [x] #33641
- [x] #33336
- [x] #33279
- [ ] #32265
- [ ] #31123
- [x] #30333
- [x] #30629
- [x] #28949
- [x] #28650
- [ ] #25448
- [x] #13272
- [x] #39219
- [x] #41710
- [x] #45462
- [x] #45342
- [x] #44536
- [x] #50462
```
## Documentation
```[tasklist]
- [ ] #15675
- [ ] #11766
- [ ] #9949
```
## Proposals
Some issues remain in Limbo due to indecision regarding the philosophical principles of module development (what's deemed as a module now should be core, for instance see attempted fixes #32272), so it may not worth fixing some of the bugs at the moment (raising awareness for contributors).
- Godot's C++ modules should be self-contained godotengine/godot-proposals#569.
---
### List of modules which cannot be disabled without...
### ... compile-time errors
##### with `target=debug tools=yes`:
- `regex`: rename dialog: https://github.com/godotengine/godot/pull/31290#issuecomment-520227369)
- `freetype`: font rendering
- `svg`: editor icons scaling
- `gdscript` #41710 (3.2)
##### with `platform=windows tools=yes bits=32`:
- ~~`denoise`: #39186, #38759.~~
### ... run-time errors
##### with `target=debug tools=yes`:
- `gdnavigation`: #36091
- `bullet`: #32217, #32216
- `mbedtls`: asset library (requires SSL), crash on project startup.
- `glslang`: shader compilation (?) numerous errors spam on startup.
|
bug,topic:core,documentation,tracker
|
low
|
Critical
|
627,989,222 |
godot
|
instances_cull_ray not reliable when scene translated, rotated and/or scaled.
|
Windows 10 Pro (10.0.18362)
Godot v3.2 stable
GPU: NVidia GTX 1660 Super
<!-- Specify GPU model, drivers, and the backend (GLES2, GLES3, Vulkan) if graphics-related. -->
The problem is that instances_cull_ray does not return expected results consistently when scene has been translated, rotated and/or scaled.
My hack is to use my own ray to intersect with the translated AABB of each node inside a scene which produces better results.
|
bug,topic:rendering,topic:3d
|
low
|
Minor
|
627,994,366 |
PowerToys
|
[FZ Editor] Hotkey to quickly edit current layout
|
# Summary of the new feature/enhancement
I often find myself doing just small adjustments to a layout, i.e. changing the ratio between two zones. Because #1193 is not implemented yet, this can be achieved right now by using a custom grid layout and editing it whenever I need to make an adjustment along with the "windows assigned to a zone will match new size/position" option. However, if I want to edit the layout I have to open the "choose layout" window, click the "Edit selected layout" button and (most of the time) move the "Custom layout creator" window out of the way before I can actually make the adjustment. After that I also have to click on an apply button twice. Having a hotkey to instantly open the "Edit selected layout" screen would make this process much faster.
From a user's perspective this could look like this:
1. Configure a hotkey in the PowerToys settings window
2. Select a layout and position windows within it
3. Whenever a quick adjustment is required, hit the configured hotkey
4. In the editor, make the layout adjustment
5. Hit the escape/enter key to apply the changes and close the editor
6. The adjustment is applied to the windows in the layout
|
Idea-Enhancement,Help Wanted,FancyZones-Editor,FancyZones-Hotkeys,Product-FancyZones
|
low
|
Major
|
628,006,049 |
rust
|
Tracking Issue for `string_remove_matches`
|
<!--
Thank you for creating a tracking issue! π Tracking issues are for tracking a
feature from implementation to stabilisation. Make sure to include the relevant
RFC for the feature if it has one. Otherwise provide a short summary of the
feature and link any relevant PRs or issues, and remove any sections that are
not relevant to the feature.
Remember to add team labels to the tracking issue.
For a language team feature, this would e.g., be `T-lang`.
Such a feature should also be labeled with e.g., `F-my_feature`.
This label is used to associate issues (e.g., bugs and design questions) to the feature.
-->
This is a tracking issue for `string_remove_matches` implemented in #71780
The feature gate for the issue is `#![feature(string_remove_matches)]`.
- [ ] Stabilization PR ([see instructions on rustc-dev-guide][stabilization-guide])
- [ ] Remove allocations (see discussion on #71780)
[stabilization-guide]: https://rustc-dev-guide.rust-lang.org/stabilization_guide.html#stabilization-pr
[doc-guide]: https://rustc-dev-guide.rust-lang.org/stabilization_guide.html#documentation-prs
### Implementation history
<!--
Include a list of all the PRs that were involved in implementing the feature.
-->
#50015 first attempt at this by @frewsxcv.
|
T-libs-api,B-unstable,C-tracking-issue,A-str,Libs-Tracked
|
low
|
Critical
|
628,011,283 |
PowerToys
|
[Run] Return recent items in results
|
# Summary of the new feature/enhancement
When opening files or other items in applications, those recent items are right-clickable from the taskbar from the applications' icons. For example, for PuTTY the recently used saved hosts are listed.
**PowerToys Run** should be able to search from those as well when applications are pinned (or running) in the taskbar.
That would make it very fast to start new sessions in PuTTY or open favourite files from other applications.
<!--
A clear and concise description of what the problem is that the new feature would solve.
Describe why and how a user would use this new functionality (if applicable).
-->
# Proposed technical implementation details (optional)
(unable to say)
<!--
A clear and concise description of what you want to happen.
-->
|
Idea-Enhancement,Product-PowerToys Run
|
low
|
Minor
|
628,048,711 |
PowerToys
|
[Run] Add common used programs above search bar
|
It would be nice if one could add program icons or common used program above the search bar for power toys ?
|
Idea-Enhancement,Product-PowerToys Run
|
low
|
Minor
|
628,060,395 |
flutter
|
Flutter tool crashes when you type Ctrl-\
|
Steps to reproduce:
* `flutter analyze --watch` ~~or `flutter run`~~
* Press ^\ (Ctrl Backslash).
Tool crashes.
|
c: crash,tool,a: quality,has reproducible steps,P2,found in release: 3.7,found in release: 3.11,team-tool,triaged-tool
|
low
|
Critical
|
628,074,515 |
godot
|
Strange error messages for AnimationPlayer that works completely fine
|
Godot Version: 3.2.1
I have an AnimationPlayer node for my player character. It works just fine, but every time I launch the game, I get this spam in the output:
https://imgur.com/a/R8iv5rH
scene/animation/animation_player.cpp:887 - Failed setting key at time 0 in Animation '' at Node '/root/EditorNode/@@585/@@586/@@594/@@596/@@600/@@604/@@605/@@606/@@622/@@623/@@632/@@633/@@6278/@@6117/@@6118/@@6119/@@6120/@@6121/Player/Body/AnimationPlayer', Track 'Model/TrailingCloth/Skeleton:polygon'. Check if property exists or the type of key is right for the property
There is no problem with the animations. The AnimationPlayer plays everything fine. It's just these weird error messages that don't give an animation name.
The error messages didn't use to be there - I believe they started when I reorganized the nodes in the player scene. I added a new root Node2D to the scene, and moved the player Body, it's Model and the AnimationPlayer of the model under that root. Don't remember the exact steps I used there - but it was my first time with that, so I didn't do it in the cleanest way. After that, my game continued to work fine, but these error messages started showing up.
|
bug,topic:core,confirmed,needs testing
|
low
|
Critical
|
628,120,516 |
create-react-app
|
Extract jest-preset-react
|
CRA separately released eslint-config-react-app, babel-preset-react-app.
It would be great if cra could release a jest-preset-react package separately for everyone to reuse.
|
issue: proposal,needs triage
|
low
|
Minor
|
628,122,779 |
bitcoin
|
Implement ASN-based banning
|
It would be useful if it were possible to ban IP ranges not just by netmask, but also by ASN number (as provided by asmap).
Suggested by @gmaxwell.
|
Feature,P2P
|
low
|
Minor
|
628,149,766 |
pytorch
|
send a Tensor to Cuda very slow
|
## π Bug
<!-- A clear and concise description of what the bug is. -->
if I call cuda() func after init cuda immediately, the time 'send a tensor to cuda' is a very small value
but If I call cuda()func after sleep(30s), the time 'send a tensor to cuda' taken is very large.
## To Reproduce
Steps to reproduce the behavior:
1γ image_test = np.array(np.random.randint(0, 256, size=(200, 200, 3), dtype='uint8'))
torch.cuda.synchronize()
start = time.time()
torch.from_numpy(image_test).cuda()
torch.cuda.synchronize()
end = time.time()
logging.info('image_trate_process_init_1: {}'.format(end - start))
2γ time.sleep(30)
3γ image_test_2 = np.array(np.random.randint(0, 256, size=(200, 200, 3), dtype='uint8'))
torch.cuda.synchronize()
start = time.time()
torch.from_numpy(image_test_2).cuda()
torch.cuda.synchronize()
end = time.time()
logging.info('image_trate_process_init_2: {}'.format(end - start))
4γ image_test_3 = np.array(np.random.randint(0, 256, size=(200, 200, 3), dtype='uint8'))
torch.cuda.synchronize()
start = time.time()
torch.from_numpy(image_test_3).cuda()
torch.cuda.synchronize()
end = time.time()
logging.info('image_trate_process_init_3: {}'.format(end - start))
1 time: 7.843971252441406e-05 s
3 time:1.4345457553863525 s
4 time:0.0001685619354248047 s
<!-- If you have a code sample, error messages, stack traces, please provide it here as well -->
## Expected behavior
<!-- A clear and concise description of what you expected to happen. -->
## Environment
Please copy and paste the output from our
[environment collection script](https://raw.githubusercontent.com/pytorch/pytorch/master/torch/utils/collect_env.py)
(or fill out the checklist below manually).
You can get the script and run it with:
```
wget https://raw.githubusercontent.com/pytorch/pytorch/master/torch/utils/collect_env.py
# For security purposes, please check the contents of collect_env.py before running it.
python collect_env.py
```
- PyTorch Version (e.g., 1.0):1.3
- OS (e.g., Linux): ubuntu
- How you installed PyTorch (`conda`, `pip`, source): pip
- Build command you used (if compiling from source):
- Python version: 3.7
- CUDA/cuDNN version: 10.1
- GPU models and configuration:
- Any other relevant information:
## Additional context
<!-- Add any other context about the problem here. -->
cc @ngimel @VitalyFedyunin
|
module: performance,module: cuda,triaged
|
low
|
Critical
|
628,181,182 |
excalidraw
|
Exporting a rectangle has artifacts
|
Not really sure what's going on. It only showed up like this for a single rectangle of my scene.
https://excalidraw.com/#json=5657069742981120,_yDCE1snJ6GYCVSk-wpLWA
|
bug
|
low
|
Minor
|
628,182,419 |
electron
|
Support inAppPurchase.restoreCompletedPurchases() in Electron 8 and 7
|
<!-- 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. -->
`inAppPurchase.restoreCompletedPurchases` function had been added to Electron 9 in https://github.com/electron/electron/pull/21461, and it would be really helpful if it can also be added to Electron 8.
### Proposed Solution
<!-- Describe the solution you'd like in a clear and concise manner -->
Also merge above pull request to Electron 8
### Alternatives Considered
<!-- A clear and concise description of any alternative solutions or features you've considered. -->
N/A
### Additional Information
<!-- Add any other context about the problem here. -->
I would also hope that Electron can support refresh receipt function. From [Apple documentation](https://developer.apple.com/documentation/storekit/), `In most cases, all your app needs to do is refresh its receipt and deliver the products in its receipt. The refreshed receipt contains a record of the userβs purchases in this app, on this device or any other device.`, so a function to refresh receipt would be better than restore transactions because `restoreCompletedPurchases` would add extra transaction to the receipt rather than simply retrieving the latest content of the receipt.
|
enhancement :sparkles:
|
low
|
Minor
|
628,306,262 |
godot
|
WARNING: cleanup: ObjectDB Instances still exist! when using GetWorld2d().DirectSpaceState
|
**Godot version:**
v3.2.2 beta3 mono
Also tried it on v3.2.1 stable mono with the same result.
**OS/device including version:**
Windows10
**Issue description:**
When using GetWorld2d().DirectSpaceState this error is shown when quitting the application.
`WARNING: cleanup: ObjectDB Instances still exist! At: core/object.cpp:2135`
**Steps to reproduce:**
1. Generate new mono project.
2. Add a node and attach a C# script to it.
3. Save this as the main scene.
4. In the script in _PhysicsProcess store the result of GetWorld2d().DirectSpaceState to a variable.
5. Play the project and then quit.
|
bug,platform:windows,confirmed,topic:physics,topic:dotnet
|
low
|
Critical
|
628,335,145 |
PowerToys
|
[FancyZones] Investigate GUID change for applied template layout
|
## Summary
Any template modification including changing space around zones changes layout id. We should investigate if it would make sense to keep the same guid in an **applied** template . We need to make sure it won't affect other desktops as well.
|
FancyZones-Editor,Product-FancyZones,Area-Quality
|
low
|
Minor
|
628,342,260 |
go
|
os/exec: LookPath() doesn't consider chroot
|
<!--
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 go1.14.3 linux/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
GO111MODULE=""
GOARCH="amd64"
GOBIN=""
GOCACHE="/home/scottclarke/.cache/go-build"
GOENV="/home/scottclarke/.config/go/env"
GOEXE=""
GOFLAGS=""
GOHOSTARCH="amd64"
GOHOSTOS="linux"
GOINSECURE=""
GONOPROXY=""
GONOSUMDB=""
GOOS="linux"
GOPATH="/home/scottclarke/go"
GOPRIVATE=""
GOPROXY="https://proxy.golang.org,direct"
GOROOT="/usr/local/go"
GOSUMDB="sum.golang.org"
GOTMPDIR=""
GOTOOLDIR="/usr/local/go/pkg/tool/linux_amd64"
GCCGO="gccgo"
AR="ar"
CC="gcc"
CXX="g++"
CGO_ENABLED="1"
GOMOD=""
CGO_CFLAGS="-g -O2"
CGO_CPPFLAGS=""
CGO_CXXFLAGS="-g -O2"
CGO_FFLAGS="-g -O2"
CGO_LDFLAGS="-g -O2"
PKG_CONFIG="pkg-config"
GOGCCFLAGS="-fPIC -m64 -pthread -fmessage-length=0 -fdebug-prefix-map=/tmp/go-build189733440=/tmp/go-build -gno-record-gcc-switches"
</pre></details>
### What did you do?
<!--
If possible, provide a recipe for reproducing the error.
A complete runnable program is good.
A link on play.golang.org is best.
-->
```
package main
import (
"log"
"os"
"os/exec"
"syscall"
)
func main() {
cmd := exec.Command("hello")
cmd.Stdout = os.Stdout
cmd.Stderr = os.Stderr
cmd.Dir = "/"
cmd.Env = []string{"PATH=/bin"}
cmd.SysProcAttr = &syscall.SysProcAttr{Chroot: "/home/scottclarke/chroot"}
if err := cmd.Start(); err != nil {
log.Fatalf("%#v", err.Error())
}
cmd.Wait()
}
```
Where `/home/scottclarke/chroot` contains a chroot which has an executable file `/bin/hello`, and there is no executable `hello` in the PATH of the system prior to the chroot.
### What did you expect to see?
The `hello` executable being run, producing the output `hello world`.
### What did you see instead?
The following error:
`2020/06/01 11:36:32 "exec: \"hello\": executable file not found in $PATH"`
This happens because the [`exec.Command()`](https://golang.org/src/os/exec/exec.go?s=6278:6323#L159) function calls `LookPath()` and sets the `lookPathErr` if it fails, which is checked in `Start()` and causes a failure. However this doesn't take into account the fact that the command will be run in a chroot.
|
help wanted,NeedsInvestigation
|
low
|
Critical
|
628,346,892 |
TypeScript
|
JSDoc missing syntax for `new Map<string, string>()`
|
**TypeScript Version:** 3.7.x-dev.201xxxxx
**Search Terms:** JSDoc Map generics assignment any unsafe
**Code**
```js
class MyServer {
constructor () {
// new Map is Map<any, any>; cannot call Map constructor with new Map<string, string[]>
/** @type {Map<string, string[]>} */
this.functions = new Map()
}
}
const myServer = new MyServer()
myServer.functions.get('foo')
```
**Expected behavior:**
Expected to be able to say what the Map contains when creating a new map. In typescript you can do `new Map<string, string[]>()` but theres no explicit generic parameter function invocation syntax for JSDoc
**Actual behavior:**
The type `Map<any, any>` exists. This fails the eslint rule of `no-unsafe-assignment` and this fails the `no-any` rule of tslint etc.
Note that doing something similar for arrays works because there's a special case inference for `[]` and it also works for `Set` because the constructor is defined differently and it has generic inference in the initialization statement.
**Playground Link:** https://www.typescriptlang.org/play?useJavaScript=true#code/MYGwhgzhAECyCeBlApgJwG5ugbwFDWmAHsA7CAF1QFdhyjVoAKASh3wOgHoAqb6AAXLwADshywwwgDwVUASxIBzADTRZCxQG0AugD4AvtG6d2BcgAs5EAHQAzKiVpzSMALzQSyAO5xJLdvq4gbjEZOTQALZIaJgM7p4+CCgYaP5RybF2Dk4u1orI5IwA5LZEREXMuCHgUHDRKQx4BKGyNHQMLGwcXLwCQqI4AIKoqGDwMpQaBkYm3RZWWY7kzmTQ7joBQVUt4ekxWPHedRmplXsNizlkeQXFpeXMQA
**Related Issues:** Yes, in eslint https://github.com/typescript-eslint/typescript-eslint/issues/2109
|
Suggestion,In Discussion
|
low
|
Minor
|
628,347,789 |
PowerToys
|
[Image Resizer] Explorer shell context submenu
|
I have the idea to turn the "Resize with Image Resizer" _item_ (in the shell context menu) into a _menu_ of its own. Also, maybe it would be an extra idea to be able to individually show/hide the sizes. Thinking of a "Show in context menu" check box or a toggle switch in the Settings window.
If none of the presets are selected to be shown, then the menu item should be as it is now.
Here's an example of the context menu menu:
Notice the "[action] with [app]" and the ellipsis...?
|
Idea-Enhancement,Product-Image Resizer
|
low
|
Minor
|
628,358,551 |
PowerToys
|
Powertoys Menu Bar Idea
|
Hi ,
I have an idea for a new powertoy / feature, something like a menu bar on top,
It will look something like this,
It can have themes like this image,
Please suggest what you think about this.
|
Idea-New PowerToy
|
low
|
Major
|
628,386,533 |
nvm
|
Installed using the script instructions on SailfishOS, but "env: can't execute `node`: No such file or directory"
|
<!-- Thank you for being interested in nvm! Please help us by filling out the following form if youβre having trouble. If you have a feature request, or some other question, please feel free to clear out the form. Thanks! -->
#### Operating system and version:
Sailfish OS 3.3.0.16
#### `nvm debug` output:
<details>
<!-- do not delete the following blank line -->
```sh
[nemo@Sailfish Templates]$ nvm debug
nvm --version: v0.35.3
$SHELL: /bin/bash
$SHLVL: 1
${HOME}: /home/nemo
${NVM_DIR}: '${HOME}/.nvm'
${PATH}: /usr/local/bin:/bin:/usr/bin:/usr/local/sbin:/usr/sbin
$PREFIX: ''
${NPM_CONFIG_PREFIX}: ''
$NVM_NODEJS_ORG_MIRROR: ''
$NVM_IOJS_ORG_MIRROR: ''
shell version: 'GNU bash, version 3.2.57(1)-release (armv7l-unknown-linux-gnueabi)'
uname -a: 'Linux 4.4.153-perf #6 SMP PREEMPT Mon Mar 23 15:23:16 UTC 2020 aarch64 GNU/Linux'
OS version: Sailfish OS 3.3.0.16 (Rokua)
curl: /usr/bin/curl, curl 7.64.0-DEV (armv7l-unknown-linux-gnueabi) libcurl/7.64.0-DEV OpenSSL/1.0.2o zlib/1.2.11
wget: /usr/bin/wget, GNU Wget 1.20.1 built on linux-gnueabihf.
git: /usr/bin/git, git version 2.25.0
grep: /bin/grep, BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
awk: /bin/awk, GNU Awk 3.1.5
sed: /bin/sed, GNU sed version 4.1.5
cut: /bin/cut, BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
basename: /bin/basename, BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
rm: /bin/rm, BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
mkdir: /bin/mkdir, BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
xargs: /usr/bin/xargs, BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
nvm current: none
which node:
which iojs:
which npm:
npm config get prefix: -bash: npm: command not found
npm root -g: -bash: npm: command not found
```
</details>
#### `nvm ls` output:
<details>
<!-- do not delete the following blank line -->
```sh
[nemo@Sailfish Templates]$ nvm ls
ls: invalid option -- 'q'
BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
Usage: ls [-1AaCxdLHRFplinshrSXvctukZ] [-w WIDTH] [FILE]...
List directory contents
-1 One column output
-a Include entries which start with .
-A Like -a, but exclude . and ..
-x List by lines
-d List directory entries instead of contents
-L Follow symlinks
-H Follow symlinks on command line
-R Recurse
-p Append / to dir entries
-F Append indicator (one of */=@|) to entries
-l Long listing format
-i List inode numbers
-n List numeric UIDs and GIDs instead of names
-s List allocated blocks
-lc List ctime
-lu List atime
--full-time List full date and time
-h Human readable sizes (1K 243M 2G)
--group-directories-first
-S Sort by size
-X Sort by extension
-v Sort by version
-t Sort by mtime
-tc Sort by ctime
-tu Sort by atime
-r Reverse sort order
-Z List security context and permission
-w N Format N columns wide
--color[={always,never,auto}] Control coloring
ls: invalid option -- 'q'
BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
Usage: ls [-1AaCxdLHRFplinshrSXvctukZ] [-w WIDTH] [FILE]...
List directory contents
-1 One column output
-a Include entries which start with .
-A Like -a, but exclude . and ..
-x List by lines
-d List directory entries instead of contents
-L Follow symlinks
-H Follow symlinks on command line
-R Recurse
-p Append / to dir entries
-F Append indicator (one of */=@|) to entries
-l Long listing format
-i List inode numbers
-n List numeric UIDs and GIDs instead of names
-s List allocated blocks
-lc List ctime
-lu List atime
--full-time List full date and time
-h Human readable sizes (1K 243M 2G)
--group-directories-first
-S Sort by size
-X Sort by extension
-v Sort by version
-t Sort by mtime
-tc Sort by ctime
-tu Sort by atime
-r Reverse sort order
-Z List security context and permission
-w N Format N columns wide
--color[={always,never,auto}] Control coloring
N/A
ls: invalid option -- 'q'
BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
Usage: ls [-1AaCxdLHRFplinshrSXvctukZ] [-w WIDTH] [FILE]...
List directory contents
-1 One column output
-a Include entries which start with .
-A Like -a, but exclude . and ..
-x List by lines
-d List directory entries instead of contents
-L Follow symlinks
-H Follow symlinks on command line
-R Recurse
-p Append / to dir entries
-F Append indicator (one of */=@|) to entries
-l Long listing format
-i List inode numbers
-n List numeric UIDs and GIDs instead of names
-s List allocated blocks
-lc List ctime
-lu List atime
--full-time List full date and time
-h Human readable sizes (1K 243M 2G)
--group-directories-first
-S Sort by size
-X Sort by extension
-v Sort by version
-t Sort by mtime
-tc Sort by ctime
-tu Sort by atime
-r Reverse sort order
-Z List security context and permission
-w N Format N columns wide
--color[={always,never,auto}] Control coloring
ls: invalid option -- 'q'
BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
Usage: ls [-1AaCxdLHRFplinshrSXvctukZ] [-w WIDTH] [FILE]...
List directory contents
-1 One column output
-a Include entries which start with .
-A Like -a, but exclude . and ..
-x List by lines
-d List directory entries instead of contents
-L Follow symlinks
-H Follow symlinks on command line
-R Recurse
-p Append / to dir entries
-F Append indicator (one of */=@|) to entries
-l Long listing format
-i List inode numbers
-n List numeric UIDs and GIDs instead of names
-s List allocated blocks
-lc List ctime
-lu List atime
--full-time List full date and time
-h Human readable sizes (1K 243M 2G)
--group-directories-first
-S Sort by size
-X Sort by extension
-v Sort by version
-t Sort by mtime
-tc Sort by ctime
-tu Sort by atime
-r Reverse sort order
-Z List security context and permission
-w N Format N columns wide
--color[={always,never,auto}] Control coloring
ls: invalid option -- 'q'
BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
Usage: ls [-1AaCxdLHRFplinshrSXvctukZ] [-w WIDTH] [FILE]...
List directory contents
-1 One column output
-a Include entries which start with .
-A Like -a, but exclude . and ..
-x List by lines
-d List directory entries instead of contents
-L Follow symlinks
-H Follow symlinks on command line
-R Recurse
-p Append / to dir entries
-F Append indicator (one of */=@|) to entries
-l Long listing format
-i List inode numbers
-n List numeric UIDs and GIDs instead of names
-s List allocated blocks
-lc List ctime
-lu List atime
--full-time List full date and time
-h Human readable sizes (1K 243M 2G)
--group-directories-first
-S Sort by size
-X Sort by extension
-v Sort by version
-t Sort by mtime
-tc Sort by ctime
-tu Sort by atime
-r Reverse sort order
-Z List security context and permission
-w N Format N columns wide
--color[={always,never,auto}] Control coloring
ls: invalid option -- 'q'
BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
Usage: ls [-1AaCxdLHRFplinshrSXvctukZ] [-w WIDTH] [FILE]...
List directory contents
-1 One column output
-a Include entries which start with .
-A Like -a, but exclude . and ..
-x List by lines
-d List directory entries instead of contents
-L Follow symlinks
-H Follow symlinks on command line
-R Recurse
-p Append / to dir entries
-F Append indicator (one of */=@|) to entries
-l Long listing format
-i List inode numbers
-n List numeric UIDs and GIDs instead of names
-s List allocated blocks
-lc List ctime
-lu List atime
--full-time List full date and time
-h Human readable sizes (1K 243M 2G)
--group-directories-first
-S Sort by size
-X Sort by extension
-v Sort by version
-t Sort by mtime
-tc Sort by ctime
-tu Sort by atime
-r Reverse sort order
-Z List security context and permission
-w N Format N columns wide
--color[={always,never,auto}] Control coloring
ls: invalid option -- 'q'
BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
Usage: ls [-1AaCxdLHRFplinshrSXvctukZ] [-w WIDTH] [FILE]...
List directory contents
-1 One column output
-a Include entries which start with .
-A Like -a, but exclude . and ..
-x List by lines
-d List directory entries instead of contents
-L Follow symlinks
-H Follow symlinks on command line
-R Recurse
-p Append / to dir entries
-F Append indicator (one of */=@|) to entries
-l Long listing format
-i List inode numbers
-n List numeric UIDs and GIDs instead of names
-s List allocated blocks
-lc List ctime
-lu List atime
--full-time List full date and time
-h Human readable sizes (1K 243M 2G)
--group-directories-first
-S Sort by size
-X Sort by extension
-v Sort by version
-t Sort by mtime
-tc Sort by ctime
-tu Sort by atime
-r Reverse sort order
-Z List security context and permission
-w N Format N columns wide
--color[={always,never,auto}] Control coloring
ls: invalid option -- 'q'
BusyBox v1.31.0 (2020-03-27 07:18:49 UTC) multi-call binary.
Usage: ls [-1AaCxdLHRFplinshrSXvctukZ] [-w WIDTH] [FILE]...
List directory contents
-1 One column output
-a Include entries which start with .
-A Like -a, but exclude . and ..
-x List by lines
-d List directory entries instead of contents
-L Follow symlinks
-H Follow symlinks on command line
-R Recurse
-p Append / to dir entries
-F Append indicator (one of */=@|) to entries
-l Long listing format
-i List inode numbers
-n List numeric UIDs and GIDs instead of names
-s List allocated blocks
-lc List ctime
-lu List atime
--full-time List full date and time
-h Human readable sizes (1K 243M 2G)
--group-directories-first
-S Sort by size
-X Sort by extension
-v Sort by version
-t Sort by mtime
-tc Sort by ctime
-tu Sort by atime
-r Reverse sort order
-Z List security context and permission
-w N Format N columns wide
--color[={always,never,auto}] Control coloring
iojs -> N/A (default)
node -> stable (-> N/A) (default)
unstable -> N/A (default)
lts/* -> lts/erbium (-> v12.17.0)
lts/argon -> v4.9.1 (-> N/A)
lts/boron -> v6.17.1 (-> N/A)
lts/carbon -> v8.17.0 (-> N/A)
lts/dubnium -> v10.20.1 (-> N/A)
lts/erbium -> v12.17.0
[nemo@Sailfish Templates]$
```
</details>
#### How did you install `nvm`?
<!-- (e.g. install script in readme, Homebrew) -->
I used the `wget` bash script one liner.
#### What steps did you perform?
Prior to installing nvm using the script, I uninstalled node from my system repo (since I believe nvm is supposed to install it itself in user space) and deleted potential remnants from a prior installation:
```sh
sudo zypper remove node
rm -rf ~/.nvm
rm -rf ~/.npm
rm -rf ~/node_modules
```
I also made sure nothing related to node or nvm was in my `~/.bashrc` and `~/.bash_profile`. Then I installed nvm using the script instructions in the readme ans sourced `~/.bashrc`.
#### What happened?
I then ran `nvm install --lts`, which downloaded it but I get this at the end:
```sh
v12.17.0 is already installed.
env: can't execute 'node': No such file or directory
nvm is not compatible with the npm config "prefix" option: currently set to ""
Run `nvm use --delete-prefix v12.17.0` to unset it.
```
Running `nvm use --delete-prefix v12.17.0` as instructed did not help.
#### What did you expect to happen?
I tried also other node versions, but every time I will get this `env: can't execute node: No such file or directory`
#### Is there anything in any of your profile files that modifies the `PATH`?
<!-- (e.g. `.bashrc`, `.bash_profile`, `.zshrc`, etc) -->
No:
```sh
# .bashrc
# Source global definitions
if [ -f /etc/bashrc ]; then
. /etc/bashrc
fi
# User specific aliases and functions
alias cmus='export TERM=xterm-color && cmus'
alias cava='export TERM=xterm-color && cava'
# Set default terminal editor
export EDITOR='micro'
export VISUAL='micro'
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
```
```sh
# .bash_profile
# Get the aliases and functions
if [ -f ~/.bashrc ]; then
. ~/.bashrc
fi
```
Note that we have an old version of `node` available for installation through our package repository, and I used to use it successfully, however it's very old and therefore incompatible with many things now. Plus, as far as I understand, installing node from repository is discouraged because it will require root privileges to use npm afterwards, right?
|
needs followup
|
low
|
Critical
|
628,409,500 |
PowerToys
|
[Run] Support press the same key twice to open Run
|
Idea-Enhancement,Product-PowerToys Run
|
medium
|
Major
|
|
628,427,274 |
godot
|
Grabbing and moving Godot window is laggy on Linux (NVIDIA with Allow Flipping enabled)
|
___
***Bugsquad note:** This issue has been confirmed several times already. No need to confirm it further.*
___
**Godot version:**
Godot mono 64 bit 3.x and higher.
**OS/device including version:**
Ubuntu Linux 18.04 with XFCE window manager.
GeForce GTX 750 Ti, NVIDIA driver 435.21, GLES3 and GLES2.
**Issue description:**
When you try to move a Godot window it's very hard to move it because it becomes laggy. Some of the CPU cores become maxed out while you're moving it. This also happens to any OS windows that are moving across the Godot window.
**Steps to reproduce:**
1) Install Ubuntu 18.04 with NVDIA drivers.
2) Get Godot stable Linux Mono 64 bit version.
3) Create any "hello world" project.
4) Run project.
5) Try to grab and drag the Godot window around or any other window across it.
**Video showing the problem**
https://www.dropbox.com/s/3n7vwjqp69274sa/2020-06-01%2015-56-06.mov?dl=0
|
bug,platform:linuxbsd,topic:rendering,confirmed,topic:thirdparty,performance
|
medium
|
Critical
|
628,475,261 |
create-react-app
|
Typescript template dependencies
|
I saw that using cra with typescript template adds the dev dependencies to the `dependencies` property in `package.json` instead of `devdepdencies`
I see it here: [packages/cra-template-typescript/template.json](https://github.com/facebook/create-react-app/blob/master/packages/cra-template-typescript/template.json)
Any reason for that?
Is it possible to move it to `devdependencies` or will it break something?
|
issue: proposal,needs triage
|
low
|
Minor
|
628,535,799 |
node
|
stream: end(callback) does not propagate write after end to callback
|
```js
w.end();
w.end('asd', (err) => {
assert.strictEqual(err.code, 'ERR_STREAM_WRITE_AFTER_END'); // Fails
});
```
|
confirmed-bug,http,stream
|
low
|
Major
|
628,545,452 |
terminal
|
Offer information about the key commands for copy/paste the first time you press ctrl c/v
|
# Description of the new feature/enhancement
Have some kind of popup which lets you know that you can use ctrl + shift + c/v the first time a user presses either of those key commands. I'm a fresh convert from macOS and eventually found out you could right click to paste, but it too searching and finding https://github.com/microsoft/terminal/issues/968 till I found out the key commands.
There's a case for when you ctrl + c to close a process, but I think showing it _once_ is an acceptable trade-off (you _might_ also know if there's a running process which it could close) and for ctrl + v that's a NOOP, so doing something there is an easy win
# Proposed technical implementation details (optional)
Probably not relevant
|
Issue-Feature,Help Wanted,Area-UserInterface,Product-Terminal
|
low
|
Minor
|
628,552,165 |
TypeScript
|
Contextual typing fails to throw error with function return type on left side of equals sign
|
<!-- π¨ STOP π¨ π¦π§π’π£ π¨ πΊπ»πΆπ· π¨
Half of all issues filed here are duplicates, answered in the FAQ, or not appropriate for the bug tracker. Even if you think you've found a *bug*, please read the FAQ first, especially the Common "Bugs" That Aren't Bugs section!
Please help us by doing the following steps before logging an issue:
* Search: https://github.com/Microsoft/TypeScript/search?type=Issues
* Read the FAQ: https://github.com/Microsoft/TypeScript/wiki/FAQ
Please fill in the *entire* template below.
-->
<!--
Please try to reproduce the issue with the latest published version. It may have already been fixed.
For npm: `typescript@next`
This is also the 'Nightly' version in the playground: http://www.typescriptlang.org/play/?ts=Nightly
-->
**TypeScript Version:** I got the same behavior on 3.7.4 and 4.0.0-dev.20200601
<!-- Search terms you tried before logging this (so others can find this issue more easily) -->
**Search Terms:**
left side
right side
contextual typing
type inference
**Code**
```ts
interface ReportState {
reportDefinition: {
[key: number]: ReportDetails;
};
}
interface ReportDetails {
requiredNumber: number;
columnWidths?: number[];
}
export const Report: (state: ReportState, action: any) => ReportState = (
state,
action
) => { // See note below about this line
switch (action.type) {
case "TYPE0": {
const { reportDef, sequenceId } = action.payload;
return {
...state,
reportDefinition: {
...state.reportDefinition,
[sequenceId]: {
...reportDef,
selectedElements: ["0x0"],
},
},
};
}
case "TYPE1": {
return {
...state,
reportDefinition: {
...state.reportDefinition,
[action.payload.sequenceId]: {
// requiredNumber is missing here. This should be an error.
columnWidths: action.payload.columnWidths,
},
},
};
}
default:
return state;
}
};
```
**Expected behavior:**
There should be an error, because requiredNumber is missing from the second return statement.
**Actual behavior:**
Compilation succeeds without error.
_Curiosity 1:_
If you add the return type on the right side of the `=`, i.e. by changing
```) => {```
to
```): ReportState => {```
, it correctly finds the error. This should not be necessary, because the return type is already specified on the left side of the `=`.
_Curiosity 2:_
If you remove the first case block, it correctly finds the error (which is in the second case block).
**Playground Link:** <!-- A link to a TypeScript Playground "Share" link which demonstrates this behavior --> [Click here](https://www.typescriptlang.org/play/?ssl=29&ssc=6&pln=17&pc=1#)
**Related Issues:** <!-- Did you find other bugs that looked similar? -->
No
|
Bug
|
low
|
Critical
|
628,581,136 |
electron
|
showItemInFolder doesn't work on Windows if . is in the path
|
<!-- 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 an issue that matches the one I want to file, without success.
### Issue Details
* **Electron Version:**
* 8.3.0
* 7.1.9
* **Operating System:**
* Windows
* **Last Known Working Electron version:**
* 7.1.8
### Expected Behavior
Use `shell.showItemInFolder` to show the contents of a folder:
```js
shell.showItemInFolder("C:\\repos\\.");
```
### Actual Behavior
The contents of the folder is show in file explorer.
### To Reproduce
```js
// show contents of some folder on your machine
shell.showItemInFolder("C:\\repos\\.");
```
### Additional Information
- I have a feeling that [this](https://github.com/electron/electron/pull/21749) is what broke this.
- Using `shell.openItem` still results in the expected behavior, even if there is a . in the path
- That said, this **does not** repro on macOS and Linux. So while I'd be fine with this not getting fixed and just being told "use `openItem`" because of the 2nd bullet, it'd be nice if the APIs were consistent across platforms. π
|
platform/windows,bug :beetle:,8-x-y,7-3-x,11-x-y
|
medium
|
Critical
|
628,582,024 |
flutter
|
Change Highlight Shape of the Tab component
|
Is there some property to modify the shape ?
Or maybe hide the default tab's highlight and add a new one.
I have tried using InkWell but it doesn't work.
|
framework,f: material design,a: quality,c: proposal,P3,team-design,triaged-design
|
low
|
Minor
|
628,605,867 |
TypeScript
|
Add getMappedTypeParameters to TypeChecker
|
## Search Terms
TypeMapper, mapper, mapper1, mapper2
## Suggestion
To help transformer developers: Add a helper on `TypeChecker` to provide (and expose) the inferred (mapped) types of a source function call. i.e.
given the source:
```ts
function foo<C>(ctor: C)...
foo(Bar);
```
provide transformer developers a method (on ```ts.TypeChecker```)
```ts
getMappedTypeParameters(node: ts.CallExpression): ts.Type[]
```
so they can get the inferred TypeParameter nodes for a given source call-expression. By adding a helper method, it does not expose the internal mapping, limits the required maintenance if the internal mapping is ever changed, yet allows developers access to the inference information.
## Use Cases
During AST code transform, the compiler has already inferred the types passed to source functions, without them being explicitly defined as a type parameter in the call. The current `Signature` (```typeChecker.getResolvedSignature(node)```) does not expose these inferred types and the ```signature.typeParameters`` remains `undefined`
For transformers that require this information, the current work around is use the internal "mapper" field on `Signature`. This mapper field changed in ts 3.9 and potentially could change again. Exposing an ability to access this mapping via the TypeChecker should help reduce breakage and limit maintainance required by the TypeScript Team to expose and execute the mapping.
## Examples
In my [DI transformer](https://github.com/pigly-di/pigly), I attempt to infer Type parameters and use them to manipulate the compiled output to configure the DI Container. For example; the source code:
```ts
function main() {
let kernel = new Kernel();
kernel.bind(toSelf(Bar));
kernel.bind(toSelf(Foo));
}```
is transpiled to:
```js
function main() {
let kernel = new pigly_1.Kernel();
kernel.bind(Symbol.for("Bar"), pigly_1.toSelf(Bar));
kernel.bind(Symbol.for("Foo"), pigly_1.toSelf(Foo, ctx => ctx.resolve({ service: Symbol.for("Bar"), name: "bar" }).first()));
}
```
As relevant to this proposal, my transpiler needs to be able to infer the TypeParameter of `.bind<T>` so that it can emit a symbol for it into the runtime code.
My Current work around to discover the inferred mappings is here:
https://github.com/pigly-di/pigly/blob/develop/packages/pigly-transformer/src/index.ts#L284-L331
Admittedly this is a very niche proposal (if allowed I'd happily do a PR to add it and do the tests) - but at least one other developer had trouble with this: https://stackoverflow.com/questions/48886508/typechecker-api-how-do-i-find-inferred-type-arguments-to-a-function
## Checklist
My suggestion meets these guidelines:
* [x] This wouldn't be a breaking change in existing TypeScript/JavaScript code
* [x] This wouldn't change the runtime behavior of existing JavaScript code
* [x] This could be implemented without emitting different JS based on the types of the expressions
* [x] This isn't a runtime feature (e.g. library functionality, non-ECMAScript syntax with JavaScript output, etc.)
* [x] This feature would agree with the rest of [TypeScript's Design Goals](https://github.com/Microsoft/TypeScript/wiki/TypeScript-Design-Goals).
|
Suggestion,In Discussion
|
low
|
Minor
|
628,623,020 |
flutter
|
RenderOpacity sometimes omits its layer during animations which breaks caching performance
|
The RenderOpacity object will not emit a layer when the opacity is 0 or 255. If that is the static opacity then it might make sense, but when it is the value during an animation, this causes the layer tree to drastically change configuration on each frame.
While this could potentially result in a hit if the tree needs to be constantly reconstructed, an even bigger hit happens because the change of identity of its own layer as it moves from a partial value to either 0 or 255 and back again means that the cached information at the engine level needs to be discarded and then re-rendered.
The layer should absolutely be left intact even for opacity values of 0 or 255 during an animation.
|
framework,a: animation,P2,team-framework,triaged-framework
|
low
|
Major
|
628,628,373 |
flutter
|
Implement scene builder ops when an Android view is involved
|
These are the operations that should be applied to the `android.view.View` corresponding to the `fluter::PlatformViewLayer` instance.
- [x] `pushTransform` P0
- [x] `pushClipRect` P0
- [x] `pushOpacity`
- [x] `pushClipPath`
- [x] `pushClipRRect` p2
- [ ] `pushBackdropFilter`
- [ ] `pushColorFilter`
- [ ] `pushPhysicalShape`
- [ ] `pushShaderMask`
To do this, the [mutator stack](https://github.com/flutter/engine/blob/f5a4ea7d8321014269c83eea2f1b33224068f2da/flow/embedded_views.h#L127) must be passed to the Android embedding through the `FlutterViewOnDisplayPlatformView` JNI call.
For more context, see the "Mapping a Flutter Layer Tree to an Android View" section in go/flutter-hybrid-composition.
|
platform-android,engine,a: platform-views,customer: product,customer: money (g3),P2,team-android,triaged-android
|
low
|
Major
|
628,638,798 |
TypeScript
|
JS global variable suggestion
|
*TS Template added by @mjbvz*
**TypeScript Version**: 4.0.0-dev.20200601
**Search Terms**
- global
- javascript
- suggestions
- intellisense
---
Issue Type: <b>Feature Request</b>
In WebStorm, if I write
```js
global.abc = 123;
```
and then type `a` at anywhere I will get suggestion `abc`, but VSCode cannot do that. Screenshots as follow
VS Code version: Code 1.45.1 (5763d909d5f12fe19f215cbfdd29a91c0fa9208a, 2020-05-14T08:27:35.169Z)
OS version: Windows_NT x64 10.0.18363
<!-- generated by issue reporter -->
|
Suggestion,Awaiting More Feedback
|
low
|
Minor
|
628,656,491 |
rust
|
Abort from malloc on arm-unknown-linux-gnueabihf
|
<!--
Thank you for filing a bug report! π Please provide a short summary of the bug,
along with any information you feel relevant to replicating the bug.
-->
I tried this code:
```rust
joshua@arm:~/libc $ rustc build.rs --crate-type bin -C opt-level=1
malloc(): invalid size (unsorted)
Aborted (core dumped)
```
I expected to see this happen: The crate compiles successfully.
Instead, this happened: rustc aborts.
This is build.rs for the libc crate on master: https://github.com/rust-lang/libc/blob/f10ee11f7043cb38fff02d9085b26009f2fc66aa/build.rs. The error goes away if I don't pass `-C opt-level=1`. It happens on a few different crates, maybe rustc is trying to allocate too much memory?
### Meta
<!--
If you're using the stable version of the compiler, you should also check if the
bug also exists in the beta or nightly versions.
-->
`rustc --version --verbose`:
```
rustc 1.43.1 (8d69840ab 2020-05-04)
binary: rustc
commit-hash: 8d69840ab92ea7f4d323420088dd8c9775f180cd
commit-date: 2020-05-04
host: arm-unknown-linux-gnueabihf
release: 1.43.1
LLVM version: 9.0
```
I can reproduce this issue at least as far back as 1.34.0.
|
I-crash,A-LLVM,O-Arm,T-compiler,C-bug
|
low
|
Critical
|
628,691,610 |
tensorflow
|
tf.io.gfile / GCS fails to work on OpenSUSE
|
**System information**
- Have I written custom code (as opposed to using a stock example script provided in TensorFlow): Yes
- OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Linux OpenSUSE Tumbleweed
- TensorFlow installed from (source or binary): Binary (conda)
- TensorFlow version (use command below): unknown 2.1.0
- Python version: 3.7.5
- CUDA/cuDNN version: N/A
- GPU model and memory: N/A
**Describe the current behavior**
```python
import tensorflow as tf
tf.io.gfile.listdir("gs://some-bucket") # replace w/ bucket of your choice
```
This code gives an error:
```
2020-06-01 15:43:56.684531: W tensorflow/core/platform/cloud/google_auth_provider.cc:178] All attempts to get a Google authentication bearer token failed, returning an empty token. Retrieving token from files failed with "Unavailable: Error executing an HTTP request: libcurl code 77 meaning 'Problem with the SSL CA cert (path? access rights?)', error details: error setting certificate verify locations:
CAfile: /etc/ssl/certs/ca-certificates.crt
CApath: none". Retrieving token from GCE failed with "Aborted: All 10 retry attempts failed. The last failure: Unavailable: Error executing an HTTP request: libcurl code 6 meaning 'Couldn't resolve host name', error details: Couldn't resolve host 'metadata'".
```
After that it hangs for a while, and then raises a `NotFoundError`.
I believe this is because the libcurl packaged with tensorflow doesn't know where to find the ca-certificates bundle file on OpenSUSE, which is at `/etc/ssl/ca-bundle.pem` rather than `/etc/ssl/certs/ca-certificates.crt`. Also, I installed through miniconda so there's another equivalent file at `$CONDA_PREFIX/ssl/cacert.pem`. Neither of these seems to be found by tensorflow.
[This code](https://github.com/tensorflow/tensorflow/blob/5597c17b6a677be5264ebda7cc31404f0ae8a434/tensorflow/core/platform/cloud/curl_http_request.cc#L129-L132) suggests that the bundle file's location can be customized with the `CURL_CA_BUNDLE` env variable. However, this doesn't change the behavior as far as i can tell; the error is still raised.
**Describe the expected behavior**
It should list the contents of the bucket.
|
type:bug,comp:core,TF 2.11
|
medium
|
Critical
|
628,699,951 |
rust
|
Investigation into ASCII ctype inherent methods performance with lookup table
|
Prompted by https://github.com/rust-lang/rust/issues/68983#issuecomment-613753557, I started looking into whether the performance of methods like `is_ascii_alphabetic` could be improved by using a lookup table.
The full list of methods:
- `is_ascii_alphabetic`
- `is_ascii_alphanumeric`
- `is_ascii_control`
- `is_ascii_digit`
- `is_ascii_graphic`
- `is_ascii_hexdigit`
- `is_ascii_lowercase`
- `is_ascii_punctuation`
- `is_ascii_uppercase`
- `is_ascii_whitespace`
## Implementations
Currently, all of these methods are implemented by [matching on characters/ranges](https://github.com/rust-lang/rust/blob/5fd2f06e99a985dd896684cb2c9f8c7090eca1ab/src/libcore/num/mod.rs#L4657-L4659):
```rust
#[inline]
pub const fn is_ascii_alphabetic(&self) -> bool {
matches!(*self, b'A'..=b'Z' | b'a'..=b'z')
}
```
I investigated two ways of implementing these functions with a lookup table:
A lookup table for the entire `u8` range:
```rust
#[inline]
pub const fn is_ascii_alphabetic(&self) -> bool {
const LOOKUP_TABLE : [bool; 258] = ...;
LOOKUP_TABLE[*self as usize]
}
```
A hybrid approach by checking if the byte is ascii first, reducing table size by half:
```rust
#[inline]
pub const fn is_ascii_alphabetic(&self) -> bool {
const LOOKUP_TABLE : [bool; 128] = ...;
self.is_ascii() && LOOKUP_TABLE[*self as usize]
}
```
I will be calling these implementations `branch`, `lookup` and `hybrid` respectively throughout the rest of this report.
Using the features `const_if_match` and `const_loop`, the lookup tables for `lookup` and `hybrid` can be easily generated:
<details>
```rust
macro lookup_table($condition:expr, $n:literal) {
{
const LOOKUP_TABLE: [bool; $n] = {
let mut lookup = [false; $n];
let mut i = 0;
while i < $n {
lookup[i] = $condition(&(i as u8));
i += 1;
}
lookup
};
LOOKUP_TABLE
}
}
#[inline]
pub const fn is_ascii_alphabetic_lookup(byte: &u8) -> bool {
lookup_table!(u8::is_ascii_alphabetic, 256)[*self as usize]
}
#[inline]
pub const fn is_ascii_alphabetic_hybrid(byte: &u8) -> bool {
byte.is_ascii() && lookup_table!(u8::is_ascii_alphabetic, 128)[*self as usize]
}
```
</details>
The instructions these different implementations compile down to can be compared on [godbolt](https://rust.godbolt.org/z/aFnVi4).
## Benchmark
*Note: I do not have enough experience with benchmarking to know if this approach is fully representative, let me know if you know any way the benchmark can be improved.*
The task I used for benchmarking is iterating through the bytes of a source text and counting how many of them satisfy a condition, e.g. `is_ascii_alphabetic`:
```rust
const SOURCE_TEXT : &'static str = ...;
#[bench]
fn is_ascii_alphabetic_branch {
$bench.iter(|| {
let mut total = 0;
for byte in SOURCE_TEXT.bytes() {
if byte.is_ascii_alphabetic { total += 1; }
}
assert_eq!(total, ...);
});
}
```
This results in a tight loop, which due to caching might be favorable to the lookup tables. However, some [quick searching](https://github.com/search?l=Rust&q=is_ascii_alphabetic&type=Code) for the use of the ascii methods in open source projects reveals at least a few instances of them being used in a loop or iterator filter, so the benchmark is representative of at least some real-world usage.
As a source text I used "Hamlet, Prince of Denmark", adapted from https://www.gutenberg.org/files/27761/27761-0.txt, a primarily ascii text of 4k+ lines:
<details>
<summary>Benchmark</summary>
```rust
#![feature(test)]
#![feature(decl_macro)]
#![feature(const_if_match)]
#![feature(const_loop)]
#![feature(const_ascii_ctype_on_intrinsics)]
extern crate test;
macro lookup_table($condition:expr, $n:literal) {
{
const LOOKUP_TABLE: [bool; $ n] = {
let mut lookup_table = [false; $ n];
let mut i = 0;
while i < $ n {
lookup_table[i] = $ condition(&(i as u8));
i += 1;
}
lookup_table
};
LOOKUP_TABLE
}
}
macro bench($condition:ident, $branch_ident:ident, $hybrid_ident:ident, $lookup_ident:ident, $expected:literal) {
#[bench]
fn $branch_ident(bench: &mut test::Bencher) {
bench_impl!(bench, u8::$condition, $expected);
}
#[bench]
fn $hybrid_ident(bench: &mut test::Bencher) {
#[inline]
fn condition(byte: &u8) -> bool {
byte.is_ascii() && lookup_table!(u8::$condition, 128)[*byte as usize]
}
bench_impl!(bench, condition, $expected);
}
#[bench]
fn $lookup_ident(bench: &mut test::Bencher) {
#[inline]
fn condition(byte: &u8) -> bool {
lookup_table!(u8::$condition, 256)[*byte as usize]
}
bench_impl!(bench, condition, $expected);
}
}
macro bench_impl($bench:expr, $condition:expr, $expected:literal) {
$bench.iter(|| {
let mut total = 0;
for byte in SOURCE_TEXT.bytes() {
if $condition(&byte) { total += 1; }
}
assert_eq!(total, $expected);
});
}
// "Hamlet, Prince of Denmark", adapted from https://www.gutenberg.org/files/27761/27761-0.txt
const SOURCE_TEXT: &'static str = include_str!("hamlet.txt");
bench!(is_ascii_alphabetic, is_ascii_alphabetic_branch, is_ascii_alphabetic_hybrid, is_ascii_alphabetic_lookup, 83663);
bench!(is_ascii_alphanumeric, is_ascii_alphanumeric_branch, is_ascii_alphanumeric_hybrid, is_ascii_alphanumeric_lookup, 83718);
bench!(is_ascii_control, is_ascii_control_branch, is_ascii_control_hybrid, is_ascii_control_lookup, 8216);
bench!(is_ascii_digit, is_ascii_digit_branch, is_ascii_digit_hybrid, is_ascii_digit_lookup, 55);
bench!(is_ascii_graphic, is_ascii_graphic_branch, is_ascii_graphic_hybrid, is_ascii_graphic_lookup, 94284);
bench!(is_ascii_hexdigit, is_ascii_hexdigit_branch, is_ascii_hexdigit_hybrid, is_ascii_hexdigit_lookup, 23825);
bench!(is_ascii_lowercase, is_ascii_lowercase_branch, is_ascii_lowercase_hybrid, is_ascii_lowercase_lookup, 76787);
bench!(is_ascii_punctuation, is_ascii_punctuation_branch, is_ascii_punctuation_hybrid, is_ascii_punctuation_lookup, 10566);
bench!(is_ascii_uppercase, is_ascii_uppercase_branch, is_ascii_uppercase_hybrid, is_ascii_uppercase_lookup, 6876);
bench!(is_ascii_whitespace, is_ascii_whitespace_branch, is_ascii_whitespace_hybrid, is_ascii_whitespace_lookup, 33893);
```
</details>
Files: [benches.zip](https://github.com/rust-lang/rust/files/4708850/benches.zip)
### Results
| method | `branch` | `hybrid` | `lookup` |
|---|---|---|---|
| `is_ascii_alphabetic` | 364,130 Β±15,302 | 349,235 Β±32,393 | 402,965 Β±22,624 |
| `is_ascii_alphanumeric` | 310,435 Β±12,975 | 346,025 Β±18,873 | 404,880 Β±23,582 |
| `is_ascii_control` | 154,802 Β±17,761 | 105,160 Β±10,883 | 197,493 Β±73,893 |
| `is_ascii_digit` | 96,971 Β± 23,020 | 98,597 Β± 9,754 | 97,051 Β± 4,354 |
| `is_ascii_graphic` | 274,238 Β± 12,480 | 296,010 Β± 24,268 | 340,445 Β± 32,739 |
| `is_ascii_hexdigit` | 358,390 Β± 21,165 | 280,170 Β± 17,197 | 309,484 Β±9,983 |
| `is_ascii_lowercase` | 329,840 Β± 45,032 | 334,270 Β±21,637 | 398,315 Β±18,173 |
| `is_ascii_punctuation` | 372,600 Β± 22,126 | 157,147 Β±7,353 | 173,555 Β±19,918 |
| `is_ascii_uppercase` | 131,011 Β±26,059 | 135,658 Β±19,618 | 142,463 Β±19,191 |
| `is_ascii_whitespace` | 232,675 Β±40,710 | 292,750 Β±61,629 | 350,440 Β±56,609 |
Adapted from the output of `cargo bench`, results are in ns.
### Analysis
It seems the `hybrid` approach with the smaller lookup table was overall as fast or faster than `lookup`, even though it has to do an extra check (smaller table fits better in cache?).
The `hybrid` approach was significantly faster than the current `branch` implementation for:
- `is_ascii_control` (105,160 Β±10,883 vs. 154,802 Β±17,761)
- `is_ascii_hexdigit` (280,170 Β± 17,197 vs. 358,390 Β± 21,165)
- `is_ascii_punctuation` (157,147 Β±7,353 vs. 372,600 Β± 22,126)
Looking at the current implementation for [is_ascii_hexdigit](https://github.com/rust-lang/rust/blob/5fd2f06e99a985dd896684cb2c9f8c7090eca1ab/src/libcore/num/mod.rs#L4828-L4830) and [is_ascii_punctuation](https://github.com/rust-lang/rust/blob/5fd2f06e99a985dd896684cb2c9f8c7090eca1ab/src/libcore/num/mod.rs#L4865-L4867) and [compiler output](https://rust.godbolt.org/z/TKDJws), these two methods have the most complex branches of all the ascii methods, indicating that there is indeed potential for a faster implementation using a lookup table.
Why the `hybrid` version of `is_ascii_control` is faster I can not say, maybe caching or an artifact of the way I benchmarked?
### Conclusion
From this preliminary investigation it seems that at least some of the ascii methods can be implemented faster, prompting further investigation.
As further steps I propose first evaluating the merits of this benchmark, and then conduct more. There are still a number of questions: are the results of this benchmark correct, or is another factor interfering with the results? Is the benchmark representative of real-world usage? Is the source text influencing the benchmark?
Once we are sure our benchmarks are measuring what we want, we can proceed with the investigation: quantify the trade-off of speed vs. memory, and compare alternate implementations. This information will inform the discussion on whether it makes sense to change the current implementations.
|
I-slow,C-enhancement,T-libs
|
low
|
Major
|
628,706,149 |
TypeScript
|
[lib.dom.d.ts] Include well-known KeyboardEvent.key values
|
## Search Terms
- `keyCode`
- `KeyboardEvent`
- `KeyboardEvent.key`
- `KeyboardEvent.keyCode`
## Suggestion
Include `KeyboardEvent.key` values in the `KeyboardEvent` interface.
The [current implementation](https://github.com/microsoft/TypeScript/blob/master/lib/lib.dom.d.ts#L9639) is:
```ts
interface KeyboardEvent extends UIEvent {
...
readonly key: string;
...
```
See [here](https://developer.mozilla.org/en-US/docs/Web/API/KeyboardEvent/key/Key_Values) for a list of some possible values.
Here is my suggestion for how this type can be usefully updated:
```ts
type ModifierKeys = "Alt" | "AltGraph" | "CapsLock" | "Control" | "Fn" | "FnLock" | "Hyper" | "Meta" | NumLock" | "ScrollLock" | "Shift" | "Super" | "Symbol" | "SymbolLock";
type WhitespaceKeys = "Enter" | "Tab" | " ";
type NavigationKeys = "ArrowDown" | "ArrowLeft" | "ArrowRight" | "ArrowUp" | "End" | "Home" | "PageDown" | "PageUp";
type FunctionKeys = "F1" | "F2" | "F3" | "F4" | "F5" | "F6" | "F7" | "F8" | "F9" | "F10" | "F11" | "F12" | "F13" | "F14" | "F15" | "F16" | "F17" | "F18" | "F19" | "F20" | "Soft1" | "Soft2" | "Soft3" | "Soft4";
type NumericKeypadKeys = "Decimal" | "Key11" | "Key12" | "Multiply" | "Add" | "Clear" | "Divide" | "Subtract" | "Separator" | "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" | "8" | "9";
interface KeyboardEvent extends UIEvent {
...
readonly key: string | ModifierKeys | WhitespaceKeys | NavigationKeys | FunctionKeys | NumericKeypadKeys;
...
}
```
In the above I left out `EditingKeys`, `UIKeys`, `DeviceKeys`, `IMEKeys`, `PhoneKeys`, `MultimediaKeys`, `AudioControlKeys`, `TVControlKeys`, `MediaControllerKeys`, `SpeechRecognitionKeys`, `DocumentKeys`, `ApplicationSelectorKeys`, and `BrowserControlKeys`, but we could either include them all at the start or just as necessary.
## Use Cases
When a user is accessing the `KeyboardEvent.key` API, a sane set of the known defaults should be available for type checking.
## Examples
This will help prevent scenarios like:
```ts
if (event.key === "ArowDown") { // <- "ArrowDown" is misspelled here
```
Since users will get completion on common values when typing.
> note: I had [previously suggested](https://github.com/DefinitelyTyped/DefinitelyTyped/issues/45062) implementing this in the React types, but @ferdaber suggested (and I agree) that the dom types themselves are the better place.
## 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,Experience Enhancement
|
medium
|
Critical
|
628,708,295 |
pytorch
|
Clean up RPC request callback implementation
|
`request_callback_impl.cpp` is gradually becoming a lengthy and overly complicated file with a lot of branches and ugly unwraps. #39216 #39267 and #38748 are adding more complexities into this file. If we keep doing this, that file will be very hard to maintain and debug. We can decompose it by letting each message has its own `process` method, but it would require us to solve the tangled unwrapping part first.
cc @ezyang @gchanan @zou3519 @bdhirsh @jbschlosser @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @gqchen @aazzolini @rohan-varma @jjlilley @osalpekar @jiayisuse @agolynski @SciPioneer @H-Huang @mrzzd @cbalioglu @xush6528
|
high priority,triage review,oncall: distributed,triaged,better-engineering,module: rpc
|
low
|
Critical
|
628,708,927 |
terminal
|
Add support for syntax highlighting
|
# Description of the new feature/enhancement
<!--
A clear and concise description of what the problem is that the new feature would solve.
Describe why and how a user would use this new functionality (if applicable).
-->
Add support for syntax highlighting specific words like MobaXterm has:
This makes finding errors/issues much easier and it's something that I immediately miss when using terminal :-)
|
Issue-Feature,Area-TerminalControl,Area-Extensibility,Product-Terminal
|
medium
|
Critical
|
628,743,369 |
pytorch
|
Clean up GIL that used to guard deleted RRef destructions
|
As RRef now holds IValues instead of `py::object` and `python_ivalue.h` has its own GIL to guard the `py::object` destruction, we no long need additional GILs to guard deleted RRefs.
cc @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @gqchen @aazzolini @rohan-varma @xush6528 @jjlilley @osalpekar
|
triaged,better-engineering,module: rpc
|
low
|
Minor
|
628,749,962 |
pytorch
|
Torch script to support dictionary with keys of type tuple
|
## π Feature
I'm seeing the following error in torch script when scripting an nn.Module that contains a dictionary of tuple (or list) data type as keys.
RuntimeError: Cannot create dict for key type '(int)', only int, float, Tensor and string keys are supported.
I'm looking into scripting a model that uses dictionary of tuple keys (Dict[Tuple[int], float])
to store tuples of ngrams with the corresponding ngram probability.
## Pitch
I expect torch script to successfully script the defined module with an attribute of type
(Dict[Tuple[int], float]).
## Additional context
[Update]
There is a specific use-case for dictionaries of tuple keys. An example is a model which keeps track of ngrams and log probabilities corresponding to each ngram in a dictionary. In this case, ngrams should be represented with a tuple/list.
Repro code:
```python
# Scripting the module below throws the above exception:
import torch
from torch.jit.annotations import Tuple, List, Dict
class MyModule(torch.nn.Module):
def __init__(self):
super(MyModule, self).__init__()
self.dict = torch.jit.Attribute({}, Dict[Tuple[int], float])
def forward(self, x):
return self.dict.get(x)
torch.jit.script(MyModule())
```
cc @suo
|
oncall: jit,triaged
|
low
|
Critical
|
628,753,982 |
pytorch
|
Enable scripting modules with recursive functions
|
## π Feature
Scripting modules with recursive function calls. torch script currently does not support this case.
Current behavior:
During scripting the module below, I see an exception:
RuntimeError:
method 'get_value_' is called recursively. Recursive calls are not supported
## Pitch
To enable scripting modules that contain recursive functions.
## Additional context
To Reproduce:
```python
# Scripting the module below throws the above exception:
import torch
from torch import Tensor
from torch.jit.annotations import Dict
class MyModule(torch.nn.Module):
def __init__(self):
super(MyModule, self).__init__()
self.dict = torch.jit.Attribute({}, Dict[Tensor, float])
def get_value_(self, y):
if y in self.dict:
return self.dict.get(y)
else:
return self.get_value_(y[1:])
def forward(self, x):
return self.get_value_(x)
torch.jit.script(MyModule())
```
cc @suo
|
oncall: jit,feature,triaged
|
low
|
Critical
|
628,755,255 |
pytorch
|
SyncBatchNorm doesn't scale well across multiple nodes on large data sizes
|
See internal details in https://fb.workplace.com/groups/1482616798564458/permalink/1520747598084711/?comment_id=1525863900906414&reply_comment_id=1537231383102999
The issue in general is that there was a reported 2x slowdown (compared to ideal scaling) for SyncBatchNorm on large data sizes while scaling across multiple nodes.
cc @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @rohan-varma @gqchen @aazzolini @xush6528 @osalpekar @jiayisuse @agolynski @VitalyFedyunin @ngimel
|
module: performance,oncall: distributed,triaged
|
low
|
Major
|
628,774,457 |
pytorch
|
The conversion of Inplace aten::append to outplace is incomplete, creating incorrect subgraphs
|
# Description
aten::append is inplace. This creates issues for downstream torchscript users that do not have the inplace semantics, such as ONNX.
For example, when used inside a loop, the output of aten::append is not captured in the loop output. Another example is that wherever the appended list is used as input, it is always using the original (empty) list, instead of the output of latest aten::append.
We have tried using the latest `remove_mutation` pass, and the issue remains as the graph is unchanged.
Below is an example with using append in a loop.
Note that output(%10) of aten::append is not captured in the loop output.
And the aten::cat node outside of loop is using the original list value(%res.1) as input.
As a side note, it is not clear what should be captured as loop output as it is not mentioned https://github.com/pytorch/pytorch/blob/master/torch/csrc/jit/OVERVIEW.md#loops.
From looking at the `If` definition above, I'm assuming that loop would have similar definition for its output as `the outputs are the new values of variables in the outer block whose values were altered in an if-statement`. If the assumption is correct, then aten::append output should also be part of loop output.
# Code to repro
```python
import torch
class AppendModel(torch.nn.Module):
def __init__(self):
super().__init__()
self.linear = torch.nn.Linear(2, 2)
def forward(self, x):
res = []
for i in range(x.size(0)):
res.append(self.linear(x[i]))
return torch.cat(res, dim=0)
m = torch.jit.script(AppendModel())
m.eval()
m._c.dump()
# freezing
m_freezed = torch._C._freeze_module(m._c)
m_freezed.dump()
# remove mutation
graph = m_freezed._get_method('forward').graph
torch._C._jit_pass_remove_mutation(graph)
print(graph)
"""
Output graph. Note that output(%10) of aten::append is not captured in the loop output.
And the aten::cat node outside of loop is using the original list value(%res.1) as input.
graph(%self : __torch__.___torch_mangle_0.AppendModel,
%x.1 : Tensor):
%27 : Float(2:1, 2:2) = prim::Constant[value=-0.1994 0.3128 -0.4812 0.3948 [ CPUFloatType{2,2} ]]()
%13 : int = prim::Constant[value=2]() # /home/bowbao/repos/pytorch/torch/nn/functional.py:1672:22
%12 : int = prim::Constant[value=1]()
%3 : int = prim::Constant[value=0]() # repro_append.py:10:30
%2 : bool = prim::Constant[value=1]() # repro_append.py:10:8
%self.linear.bias : Float(2:1) = prim::Constant[value=-0.4483 0.6543 [ CPUFloatType{2} ]]()
%res.1 : Tensor[] = prim::ListConstruct()
%5 : int = aten::size(%x.1, %3) # repro_append.py:10:23
= prim::Loop(%5, %2) # repro_append.py:10:8
block0(%i.1 : int):
%8 : Tensor = aten::select(%x.1, %3, %i.1) # repro_append.py:11:35
%16 : int = aten::dim(%8) # /home/bowbao/repos/pytorch/torch/nn/functional.py:1672:7
%17 : bool = aten::eq(%16, %13) # /home/bowbao/repos/pytorch/torch/nn/functional.py:1672:7
%ret : Tensor = prim::If(%17) # /home/bowbao/repos/pytorch/torch/nn/functional.py:1672:4
block0():
%ret.1 : Tensor = aten::addmm(%self.linear.bias, %8, %27, %12, %12) # /home/bowbao/repos/pytorch/torch/nn/functional.py:1674:14
-> (%ret.1)
block1():
%output.1 : Tensor = aten::matmul(%8, %27) # /home/bowbao/repos/pytorch/torch/nn/functional.py:1676:17
%output.3 : Tensor = aten::add_(%output.1, %self.linear.bias, %12) # /home/bowbao/repos/pytorch/torch/nn/functional.py:1678:12
-> (%output.3)
%10 : Tensor[] = aten::append(%res.1, %ret) # repro_append.py:11:12
-> (%2)
%11 : Tensor = aten::cat(%res.1, %3) # repro_append.py:12:15
return (%11)
"""
```
cc @suo
|
oncall: jit,triaged
|
low
|
Major
|
628,775,458 |
TypeScript
|
Enable node typings if file starts with node shebang
|
<!-- β οΈβ οΈ Do Not Delete This! bug_report_template β οΈβ οΈ -->
<!-- Please read our Rules of Conduct: https://opensource.microsoft.com/codeofconduct/ -->
<!-- Please search existing issues to avoid creating duplicates. -->
<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: https://code.visualstudio.com/insiders/ -->
<!-- Use Help > Report Issue to prefill these. -->
- VSCode Version:
- OS Version:
Steps to Reproduce:
Here is intellisense working correctly for a node script in a project:
Here is another screenshot showing the support not working for a global script that does have a project:
As node.js is a perfectly valid choice for scripting, and has no need for verbose project structures to work, the lack of support is a bug.
## System Info
```
CPUs | Intel(R) Core(TM) i5-6267U CPU @ 2.90GHz (4 x 2900)
-- | --
GPU Status | 2d_canvas: enabledflash_3d: enabledflash_stage3d: enabledflash_stage3d_baseline: enabledgpu_compositing: enabledmetal: disabled_offmultiple_raster_threads: enabled_onoop_rasterization: disabled_offprotected_video_decode: unavailable_offrasterization: enabledskia_renderer: disabled_off_okvideo_decode: enabledviz_display_compositor: enabled_onviz_hit_test_surface_layer: disabled_off_okwebgl: enabledwebgl2: enabled
Load (avg) | 8, 6, 6
Memory (System) | 8.00GB (0.08GB free)
Process Argv | Β
Screen Reader | no
VM | 0%
Extension | Author (truncated) | Version
-- | -- | --
better-comments | aar | 2.0.5
interactive-scala | als | 1.2.1
better-toml | bun | 0.3.2
gitignore | cod | 0.6.0
bracket-pair-colorizer-2 | Coe | 0.1.3
scala | dal | 0.0.5
vsc-material-theme | Equ | 32.6.0
vsc-material-theme-icons | equ | 1.1.4
language-julia | jul | 0.15.32
crystal-ide | kof | 0.0.4
dotty | lam | 0.1.16
Lisp | mat | 0.1.10
vscode-docker | ms- | 1.2.0
python | ms- | 2020.5.78807
remote-containers | ms- | 0.117.1
remote-ssh | ms- | 0.51.0
remote-ssh-edit | ms- | 0.51.0
remote-wsl | ms- | 0.44.2
vscode-remote-extensionpack | ms- | 0.20.0
powershell | ms- | 2020.4.0
vscode-paste-image | mus | 1.0.4
gradle-language | nac | 0.2.3
java | red | 0.61.0
datetime | rid | 2.0.1
rust | rus | 0.7.8
scala | sca | 0.3.9
ayu | tea | 0.18.0
vscode-lldb | vad | 1.5.0
vscodeintellicode | Vis | 1.2.7
vscode-java-debug | vsc | 0.26.0
vscode-java-dependency | vsc | 0.10.1
vscode-java-pack | vsc | 0.9.1
vscode-java-test | vsc | 0.22.4
vscode-maven | vsc | 0.21.4
org-mode | vsc | 1.0.0
vim | vsc | 1.14.4
debug | web | 0.24.0
markdown-all-in-one | yzh | 2.8.0
```
|
Suggestion,Awaiting More Feedback
|
low
|
Critical
|
628,777,953 |
pytorch
|
Attributes not removed by freeze module
|
## Description
There are still `SetAttr` and `GetAttr` remaining in the graph. Why is freeze module pass not removing them?
## To Reproduce
```python
class AttrMod(torch.nn.Module):
def __init__(self, x):
super().__init__()
self.x = x
def forward(self, x, y):
self.x = x
return self.x + y
x = torch.ones(1,2,3)
s_m = torch.jit.script(AttrMod(x))
print(s_m.graph)
s_m.eval()
f_s_m = torch._C._freeze_module(s_m._c)
f_s_m.dump()
"""
Scripted graph:
graph(%self : __torch__.AttrMod,
%x.1 : Tensor,
%y.1 : Tensor):
%6 : int = prim::Constant[value=1]()
= prim::SetAttr[name="x"](%self, %x.1)
%4 : Tensor = prim::GetAttr[name="x"](%self)
%7 : Tensor = aten::add(%4, %y.1, %6) # repro_ssa.py:29:15
return (%7)
Frozen module graph:
module __torch__.___torch_mangle_0.AttrMod {
parameters {
}
attributes {
x = (1,.,.) =
1 1 1
1 1 1
[ CPUFloatType{1,2,3} ]
}
methods {
method forward {
graph(%self : __torch__.___torch_mangle_0.AttrMod,
%x.1 : Tensor,
%y.1 : Tensor):
%3 : int = prim::Constant[value=1]()
= prim::SetAttr[name="x"](%self, %x.1)
%4 : Tensor = prim::GetAttr[name="x"](%self)
%5 : Tensor = aten::add(%4, %y.1, %3) # repro_ssa.py:29:15
return (%5)
}
}
submodules {
}
}
"""
```
-- Update:
We face this issue when scripting Faster RCNN, Mask RCNN, and Keypoint RCNN models. After freeze module, we still see setAttr and getAttr remaining in the graph. One real model use-case for this issue is when attributes of an object or submodule are set in the forward method. Example:
```
class BoxCoder(object):
def __init__(self, bbox_xform_clip):
# type: (float) -> None
self.bbox_xform_clip = bbox_xform_clip
def decode(self, rel_codes, boxes):
# type: (Tensor, List[Tensor]) -> Tensor
boxes_per_image = [b.size(0) for b in boxes]
boxes = torch.cat(boxes, dim=0)
pred_ctr_x = torch.clamp(rel_codes[:, 0::4], max=self.bbox_xform_clip)* boxes[:, 2]
pred_ctr_y = torch.clamp(rel_codes[:, 1::4], max=self.bbox_xform_clip)* boxes[:, 3]
return torch.stack((pred_ctr_x, pred_ctr_y), dim=2)
class MyModule(torch.nn.Module):
def __init__(self):
super(MyModule, self).__init__()
def forward(self, box_regression, proposals):
box_coder = BoxCoder(math.log(1000. / 16))
return box_coder.decode(box_regression, proposals)
model = MyModule()
model.eval()
box_regression = torch.randn([4, 4])
proposal = [torch.randn(2, 4), torch.randn(2, 4)]
output = model(box_regression, proposal)
script_m = torch.jit.script(model)
script_m.eval()
f_s_m = torch._C._freeze_module(script_m._c)
f_s_m.dump()
# module __torch__.___torch_mangle_0.MyModule {
# parameters {
# }
# attributes {
# }
# methods {
# method forward {
# graph(%self : __torch__.___torch_mangle_0.MyModule,
# %box_regression.1 : Tensor,
# %proposals.1 : Tensor):
# %15 : None = prim::Constant()
# %14 : int = prim::Constant[value=9223372036854775807]()
# %13 : int = prim::Constant[value=0]() # pytorch_test.py:13:34
# %12 : int = prim::Constant[value=4]() # pytorch_test.py:16:49
# %11 : int = prim::Constant[value=2]() # pytorch_test.py:16:89
# %10 : int = prim::Constant[value=1]() # pytorch_test.py:17:46
# %9 : int = prim::Constant[value=3]() # pytorch_test.py:17:89
# %3 : float = prim::Constant[value=4.1351665567423561]()
# %box_coder.1 : __torch__.BoxCoder = prim::CreateObject()
# = prim::SetAttr[name="bbox_xform_clip"](%box_coder.1, %3)
# %6 : Tensor[] = prim::ListConstruct(%proposals.1)
# %boxes.4 : Tensor = aten::cat(%6, %13) # pytorch_test.py:14:16
# %17 : Tensor = aten::slice(%box_regression.1, %13, %13, %14, %10) # pytorch_test.py:16:33
# %18 : Tensor = aten::slice(%17, %10, %13, %14, %12) # pytorch_test.py:16:33
# %19 : float = prim::GetAttr[name="bbox_xform_clip"](%box_coder.1)
# %20 : Tensor = aten::clamp(%18, %15, %19) # pytorch_test.py:16:21
# %21 : Tensor = aten::slice(%boxes.4, %13, %13, %14, %10) # pytorch_test.py:16:80
# %22 : Tensor = aten::select(%21, %10, %11) # pytorch_test.py:16:80
# %pred_ctr_x.1 : Tensor = aten::mul(%20, %22) # pytorch_test.py:16:21
# %25 : Tensor = aten::slice(%17, %10, %10, %14, %12) # pytorch_test.py:17:33
# %26 : float = prim::GetAttr[name="bbox_xform_clip"](%box_coder.1)
# %27 : Tensor = aten::clamp(%25, %15, %26) # pytorch_test.py:17:21
# %29 : Tensor = aten::select(%21, %10, %9) # pytorch_test.py:17:80
# %pred_ctr_y.1 : Tensor = aten::mul(%27, %29) # pytorch_test.py:17:21
# %31 : Tensor[] = prim::ListConstruct(%pred_ctr_x.1, %pred_ctr_y.1)
# %32 : Tensor = aten::stack(%31, %11) # pytorch_test.py:19:15
# return (%32)
# }
# }
# submodules {
# }
# }
```
cc @suo
|
oncall: jit,triaged
|
low
|
Minor
|
628,796,181 |
pytorch
|
Draw quantized tensors in tensorboard
|
Hi, I'd like to visualize the quantized model weights with TensorBoard:
```
for k, v in model.state_dict().items():
writer.add_histogram(k, v)
```
and got the error `TypeError: Got unsupported ScalarType QInt8`.
Is there a way to add the tensor to TensorBoard please? Thank you!
|
triaged,enhancement,module: tensorboard
|
low
|
Critical
|
628,817,265 |
kubernetes
|
Cron job staggering / randomization / concurrency limiting
|
<!-- Please only use this template for submitting enhancement requests -->
**What would you like to be added**: CronJobs should have fields allowing for their start time to be randomized (as systemd's Timers have), or even distributed within a range to spread out in equal intervals with other CronJobs of a set. At the very least, it would be useful to specify concurrency policies like "do not run more than two Jobs with the selected annotation", so that scheduling will be postponed until that policy could be met.
**Why is this needed**: I have a cluster with very few cores, and KubeApps adds a CronJob to sync every repo in the cluster, one for each repo, each one scheduled to run every 10 minutes. This would only take a few seconds per repo, but when it tries to run the CronJob at the same time for every repo, it causes a "Three Stooges effect", and most of the synchronization jobs end up timing out for lack of available processing. If I could stagger these synchronizations, they could all succeed, and my cluster wouldn't be momentarily starved of resources while it happens.
|
kind/feature,sig/apps,area/workload-api/cronjob,needs-triage
|
high
|
Critical
|
628,830,640 |
flutter
|
Prompt does not return when exiting desktop app in release mode
|
Thank you for supporting release mode.
## Steps to Reproduce
1. $ git clone https://github.com/google/flutter-desktop-embedding.git
2. $ cd flutter-desktop-embedding/testbed
3. $ flutter run -d linux --release
**Expected results:** <!-- what did you want to see? -->
The prompt should be returned when you exit the app window.
Debug mode is no problems.
**Actual results:** <!-- what did you see? -->
```
$ flutter run -d linux --release
Launching lib/main.dart on Linux in release mode...
Building Linux application...
Flutter run key commands.
h Repeat this help message.
c Clear the screen
q Quit (terminate the application on the device).
<= It will stop here even if you quit the app!!
```
<details>
<summary>flutter doctor log</summary>
```
$ flutter doctor
Doctor summary (to see all details, run flutter doctor -v):
[β] Flutter (Channel master, 1.19.0-3.0.pre.13, on Linux, locale ja_JP.UTF-8)
[!] Android toolchain - develop for Android devices (Android SDK version 29.0.2)
β Android licenses not accepted. To resolve this, run: flutter doctor --android-licenses
[β] Chrome - develop for the web
[β] Linux toolchain - develop for Linux desktop
[!] Android Studio (version 3.5)
β Flutter plugin not installed; this adds Flutter specific functionality.
β Dart plugin not installed; this adds Dart specific functionality.
[β] VS Code (version 1.45.1)
[β] Connected device (3 available)
! Doctor found issues in 2 categories.
```
</details>
|
tool,platform-mac,platform-windows,platform-linux,a: desktop,P3,team-tool,triaged-tool
|
low
|
Critical
|
628,831,561 |
pytorch
|
Pickling a Tensor or a Storage is not deterministic
|
## π Bug
When calling pickle.dump on a Tensor, Storage, or a module, the resulting bytes change between two runs. I was unable to get byte for byte reproducibility.
## To Reproduce
Steps to reproduce the behavior:
1. Run `python -c "import torch, pickle; print(pickle.dumps(torch.IntTensor([1])))"`
1. Re-run it
1. The results are not the same, [cf the diff](https://gist.github.com/MatthieuBizien/cf6fc3265d879c1a25e6bdaa152b9eaf)
Same issue with torch.save:
```
f = io.BytesIO()
torch.save(torch.IntTensor([1]), f)
print(f.getvalue())
```
The bug is still present with `PYTHONHASHSEED` fixed.
## Expected behavior
That snippet should give the same results every time.
Deterministic pickling should be made possible, to allow reproducible pipelines, eg when transforming a model from a format to another format. Converting in Numpy array works but we lose the ability to simply to `torch.load`.
## Environment
```
Collecting environment information...
PyTorch version: 1.5.0
Is debug build: No
CUDA used to build PyTorch: 10.1
OS: Ubuntu 18.04.4 LTS
GCC version: Could not collect
CMake version: Could not collect
Python version: 3.7
Is CUDA available: Yes
CUDA runtime version: Could not collect
GPU models and configuration: GPU 0: Tesla V100-SXM2-16GB
Nvidia driver version: 418.87.00
cuDNN version: Could not collect
Versions of relevant libraries:
[pip] numpy==1.18.1
[pip] torch==1.5.0
[pip] torchvision==0.6.0a0+82fd1c8
[conda] blas 1.0 mkl
[conda] cudatoolkit 10.1.243 h6bb024c_0
[conda] mkl 2020.0 166
[conda] mkl-service 2.3.0 py37he904b0f_0
[conda] mkl_fft 1.0.15 py37ha843d7b_0
[conda] mkl_random 1.1.0 py37hd6b4f25_0
[conda] numpy 1.18.1 py37h4f9e942_0
[conda] numpy-base 1.18.1 py37hde5b4d6_1
[conda] pytorch 1.5.0 py3.7_cuda10.1.243_cudnn7.6.3_0 pytorch
[conda] torchvision 0.6.0 py37_cu101 pytorch
```
|
module: pickle,triaged
|
low
|
Critical
|
628,856,579 |
deno
|
add flag to enable deterministic execution
|
This consists of
1. Setting the random number generator seed. We have this already.
2. Disabling all I/O - so turn off allow-read, allow-write, allow-net, allow-run, allow-plugin - etc.
3. Controlling the Date object
4. need a custom v8::Platform which will allow us to control the date precisely. Then we can add some APIs to allow people to change it. (Probably from the command line would be ideal?)
https://github.com/v8/v8/blob/cc1b741ef34b56d1b29cb24cb391ef38a1172ce8/include/v8-platform.h#L532-L545 So the fix is really in rusty_v8 - we need to expose those primatives
|
feat,cli
|
low
|
Minor
|
628,865,350 |
pytorch
|
RuntimeError: NCCL error in: /pytorch/torch/lib/c10d/ProcessGroupNCCL.cpp:514, unhandled system error, NCCL version 2.4.8
|
I met an error when I use DDP for multi node (two node, two GPUs each) training and 'nccl' backend (It runs perfect when I use 'gloo'). The environment is Ubuntu 16.04+python3.5+pytorch1.5.0+cuda10.1.
My code is based on the demo code on the offical website to test distributed training.
```
def setup(rank, world_size):
os.environ['NCCL_DEBUG'] = 'INFO'
os.environ['NCCL_SOCKET_IFNAME'] = 'eno1'
os.environ['NCCL_IB_DISABLE'] = '1'
dist.init_process_group(
"nccl", rank=rank, init_method='tcp://162.105.146.176:22222', world_size=world_size)
class ToyModel(nn.Module):
def __init__(self):
super(ToyModel, self).__init__()
self.net1 = nn.Linear(10, 10)
self.relu = nn.ReLU()
self.net2 = nn.Linear(10, 5)
def forward(self, x):
return self.net2(self.relu(self.net1(x)))
...
```
The error code when using 'nccl' is as following (the error occurs on the master node),
```
ptwop-176:1755:1755 [0] NCCL INFO Bootstrap : Using [0]eno1:162.105.146.176<0>
ptwop-176:1755:1755 [0] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so).
ptwop-176:1755:1755 [0] misc/ibvwrap.cc:63 NCCL WARN Failed to open libibverbs.so[.1]
ptwop-176:1755:1755 [0] NCCL INFO NET/Socket : Using [0]eno1:162.105.146.176<0>
NCCL version 2.4.8+cuda10.1
ptwop-176:1756:1756 [1] NCCL INFO Bootstrap : Using [0]eno1:162.105.146.176<0>
ptwop-176:1756:1756 [1] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so).
ptwop-176:1756:1756 [1] misc/ibvwrap.cc:63 NCCL WARN Failed to open libibverbs.so[.1]
ptwop-176:1756:1756 [1] NCCL INFO NET/Socket : Using [0]eno1:162.105.146.176<0>
ptwop-176:1755:1870 [0] NCCL INFO Setting affinity for GPU 0 to 5555,55555555
ptwop-176:1756:1871 [1] NCCL INFO Setting affinity for GPU 1 to 5555,55555555
ptwop-176:1756:1871 [1] include/socket.h:390 NCCL WARN Connect to 162.105.146.178<35007> failed : No route to host
ptwop-176:1756:1871 [1] NCCL INFO bootstrap.cc:100 -> 2
ptwop-176:1756:1871 [1] NCCL INFO bootstrap.cc:337 -> 2
ptwop-176:1755:1869 [0] include/socket.h:390 NCCL WARN Connect to 162.105.146.178<54473> failed : No route to host
ptwop-176:1756:1871 [1] NCCL INFO init.cc:695 -> 2
ptwop-176:1755:1869 [0] NCCL INFO bootstrap.cc:100 -> 2
ptwop-176:1756:1871 [1] NCCL INFO init.cc:951 -> 2
ptwop-176:1755:1869 [0] NCCL INFO bootstrap.cc:226 -> 2
ptwop-176:1756:1871 [1] NCCL INFO misc/group.cc:69 -> 2 [Async thread]
Traceback (most recent call last):
File "test.py", line 73, in <module>
run_demo(demo_basic, 2, 3)
File "test.py", line 45, in run_demo
join=True)
File "/home/xukun/setsuna/lib/python3.5/site-packages/torch/multiprocessing/spawn.py", line 200, in spawn
return start_processes(fn, args, nprocs, join, daemon, start_method='spawn')
File "/home/xukun/setsuna/lib/python3.5/site-packages/torch/multiprocessing/spawn.py", line 158, in start_processes
while not context.join():
File "/home/xukun/setsuna/lib/python3.5/site-packages/torch/multiprocessing/spawn.py", line 119, in join
raise Exception(msg)
Exception:
-- Process 1 terminated with the following error:
Traceback (most recent call last):
File "/home/xukun/setsuna/lib/python3.5/site-packages/torch/multiprocessing/spawn.py", line 20, in _wrap
fn(i, *args)
File "/home/xukun/graph/multi_node/test.py", line 57, in demo_basic
ddp_model = DDP(model.to(rank), device_ids=[rank])
File "/home/xukun/setsuna/lib/python3.5/site-packages/torch/nn/parallel/distributed.py", line 285, in __init__
self.broadcast_bucket_size)
File "/home/xukun/setsuna/lib/python3.5/site-packages/torch/nn/parallel/distributed.py", line 483, in _distributed_broadcast_coalesced
dist._broadcast_coalesced(self.process_group, tensors, buffer_size)
RuntimeError: NCCL error in: /pytorch/torch/lib/c10d/ProcessGroupNCCL.cpp:514, unhandled system error, NCCL version 2.4.8
```
How to solve this problem?
|
module: crash,triaged,module: nccl
|
medium
|
Critical
|
628,878,578 |
TypeScript
|
TS: Option to not show imports in references search
|
*TS Template added by @mjbvz*
**TypeScript Version**: 3.9.4
**Search Terms**
- references
- imports
- f12
---
Often I do a find references search and if the item is only used in one place, cmd+click on it will take me straight to it. However if the item is imported and used, I need to navigate the peek widget, expand the tree entry, then click on it. It would be nice if I can just skip over that if the only other reference is the import.
|
Suggestion,Awaiting More Feedback
|
low
|
Minor
|
628,900,265 |
next.js
|
Scrolling happens when user returns to a page with hash using browser back button
|
# Bug report
## Describe the bug
When user returns to a page which have url with hash (e.g. `http://localhost:3000/test#anchor`) from another page, unexpected scrolling to the anchor element happen.
It seems `componentDidUpdate` of Container triggers `scrollToHash` even when one returns by clicking browser back button.
(https://github.com/vercel/next.js/blob/v9.4.4/packages/next/client/index.js#L135-L137)
## To Reproduce
Example repo: https://github.com/monae/nextjs-scrolltohash-bug
1. `git clone [email protected]:monae/nextjs-scrolltohash-bug.git` and `npm run dev`.
2. Click on 'Open the test page without hash'. You are on top of the page.
3. Scroll down and click on 'Open another page'.
4. Click on browser back button, then you return to the preserved scroll position (Chrome, Safari) or top of the page (Firefox).
5. Back to the first page and click on 'Open the test page with hash'. You are on the anchor element.
6. Scroll down and click on 'Open another page'.
7. Click on browser back button, then you return to the anchor element.
## Expected behavior
The results of 4. & 7. should be same (although the behavior of Firefox is not desirable).
## Screenshots
## System information
- OS: macOS
- Browser (if applies) Chrome, Safari, Firefox
- Version of Next.js: 9.4.4
- Version of Node.js: 14.3.0
## Additional context
|
good first issue
|
low
|
Critical
|
628,973,799 |
neovim
|
UI protocol: scrolling up sent as grid_line events instead of grid_scroll
|
<!-- Before reporting: search existing issues and check the FAQ. -->
- `nvim --version`: NVIM v0.5.0-519-g74fed7d50
- UI extensions: ext_linegrid=true, ext_multigrid=false, ext_messages=false, rgb=true
### Steps to reproduce using `nvim -u NORC`
Create a 20 lines x 80 columns UI showing a buffer containing ~1000 lines and `:set nu`. Go to the bottom of the screen and press `k` to scroll up.
### Actual behaviour
Neovim sends the following events:
```
["grid_line",[1,29,71,[["g",1],["k"]]]]
["grid_line",[1,29,71,[[" ",1,2]]]]
["grid_line",[1,1,0,[[" ",67],["9"],["7"],["2"],[" "],[" ",0,77]]],[1,2,0,[[" ",67],["9"],["7"],["3"],[" "],[" ",0,77]]],[1,3,0,[[" ",67],["9"],["7"],["4"],[" "],[" ",0,77]]],[1,4,0,[[" ",67],["9"],["7"],["5"],[" "],[" ",0,77]]],[1,5,0,[[" ",67],["9"],["7"],["6"],[" "],[" ",0,77]]],[1,6,0,[[" ",67],["9"],["7",67,2],[" "],[" ",0,77]]],[1,7,0,[[" ",67],["9"],["7"],["8"],[" "],[" ",0,77]]],[1,8,0,[[" ",67],["9"],["7"],["9"],[" "],[" ",0,77]]],[1,9,0,[[" ",67],["9"],["8"],["0"],[" "],[" ",0,77]]],[1,10,0,[[" ",67],["9"],["8"],["1"],[" "],[" ",0,77]]],[1,11,0,[[" ",67],["9"],["8"],["2"],[" "],[" ",0,77]]],[1,12,0,[[" ",67],["9"],["8"],["3"],[" "],[" ",0,77]]],[1,13,0,[[" ",67],["9"],["8"],["4"],[" "],[" ",0,77]]],[1,14,0,[[" ",67],["9"],["8"],["5"],[" "],[" ",0,77]]],[1,15,0,[[" ",67],["9"],["8"],["6"],[" "],[" ",0,77]]],[1,16,0,[[" ",67],["9"],["8"],["7"],[" "],[" ",0,77]]],[1,17,0,[[" ",67],["9"],["8",67,2],[" "],[" ",0,77]]],[1,18,0,[[" ",67],["9"],["8"],["9"],[" "],[" ",0,77]]],[1,19,0,[[" ",67],["9",67,2],["0"],[" "],[" ",0,77]]],[1,20,0,[[" ",67],["9",67,2],["1"],[" "],[" ",0,77]]],[1,21,0,[[" ",67],["9",67,2],["2"],[" "],[" ",0,77]]],[1,22,0,[[" ",67],["9",67,2],["3"],[" "],[" ",0,77]]],[1,23,0,[[" ",67],["9",67,2],["4"],[" "],[" ",0,77]]],[1,24,0,[[" ",67],["9",67,2],["5"],[" "],[" ",0,77]]],[1,25,0,[[" ",67],["9",67,2],["6"],[" "],[" ",0,77]]],[1,26,0,[[" ",67],["9",67,2],["7"],[" "],[" ",0,77]]],[1,27,0,[[" ",67],["9",67,2],["8"],[" "],[" ",0,77]]],[1,28,0,[[" ",67],["9",67,3],[" "],[" ",0,77]]],[1,0,0,[[" ",67],["9"],["7"],["1"],[" "],[" ",0,77]]]]
```
### Expected behaviour
Neovim sends the following events:
```
["grid_line",[1,29,71,[["g",1],["k"]]]]
["grid_line",[1,29,71,[[" ",1,2]]]]
["grid_scroll",[1,0,29,0,82,-1,0]]
["grid_line",[1,28,0,[["1",67],["0",67,3],[" "],[" ",0,77]]]]
```
Since UI bugs aren't easy to reproduce I'll investigate myself. I'm just opening this here so that I don't forget.
|
bug,ui-extensibility
|
low
|
Critical
|
628,996,598 |
scrcpy
|
Remote connection via VS Code
|
VS Code offers a very handy feature named [Shared servers](https://docs.microsoft.com/en-us/visualstudio/liveshare/use/vscode#share-a-server) essentially allowing anyone participating in my Live Share session to access ports that I choose to expose.
I tried exposing ports 5037 and 27183 and the participant partially managed to connect using scrcpy. The issues are:
- when he taps on the screen scrcpy crashes
- both of us cannot connect to the same device at the same time. If I connect first then he cannot connect at all.
This would be a very handy feature to support to enable remote pair programming sessions.
PS: We made sure that the ports we mapped correctly on the client and that we killed his local ADB before attempting connection.
|
feature request
|
low
|
Critical
|
629,038,867 |
opencv
|
Wrong eroding 3 pixels in seamlessclone
|
##### System information (version)
<!-- Example
- OpenCV => 3.4
- Operating System / Platform => Windows 64 Bit
- Compiler => Visual Studio 2017
-->
- OpenCV => 3.4
- Operating System / Platform => :grey_question:
- Compiler => :grey_question:
##### Detailed description
https://github.com/opencv/opencv/blob/c3e8a82c9c0b18432627fe100db332b42d64b8d3/modules/photo/src/seamless_cloning_impl.cpp#L261
It leads to undesired color propagation into the mask area, for the laplacian on the mask boarder to be used in poisson solver is same as the dstβs. It's not consistent with the algorithm proposed in the [paper](https://www.cs.virginia.edu/~connelly/class/2014/comp_photo/proj2/poisson.pdf), which directly uses the original mask to compute Laplacian.
|
bug,category: photo,Hackathon
|
low
|
Minor
|
629,070,679 |
flutter
|
Creating a paragraph is expensive, which makes animating the color of many individual glyphs prohibitively expensive
|
## Use case
I need to draw some complicated text on a Canvas, in Flutter. I have an external C library that calculates the positioning of all of the characters in the text, which I call with dart:ffi, and which then performs callbacks into Flutter with the data `(character_to_draw, x, y, font/style)`.
Currently I can use either `Canvas.drawParagraph()` or `TextPainter.paint()` to then paint the character at the required position on the canvas. However, I have to call the `layout()` method in either case, which I gather is an expensive operation, and which seems superfluous when I'm just laying out individual characters.
## Proposal
I would like there to be a `Canvas.drawCharacter()` method, that draws just a single character at a given position, in a given style, but without any separate `layout()` considerations.
It isn't clear to me if one can access the underlying drawing library (Skia?) directly, to bypass `layout()`.
If the proper solution to this is to have the C-library directly interface with a Skia canvas, that'd be fine -- I can write a wrapper for that -- but it's not clear to me (a) how to get a reference to the Skia canvas or (b) how I would import Flutter's copy of Skia into my C library. Any pointers in these directions would also be appreciated.
|
c: new feature,engine,c: performance,a: typography,c: proposal,perf: speed,P2,team-engine,triaged-engine
|
low
|
Major
|
629,108,006 |
PowerToys
|
[Preview Pane] HEIC, WEBP File Explorer support
|
# Summary of the new feature/enhancement
Since native SVG support is now part of this (YES!), would love to see a few other common non-supported image types added as well. HEIC and WEBP.
# Proposed technical implementation details (optional)
I believe there are already Explorer apps that can render WEBP. Unsure of HEIC.
|
Idea-Enhancement,Help Wanted,Product-File Explorer
|
medium
|
Major
|
629,116,439 |
pytorch
|
Morphological operations
|
## π Feature
Basic morphological operations (e.g. dilation and erosion)
## Motivation
Many pytorch applications deal with images. It would therefore be great if pytorch would natively support basic morphological operations, so that the user does not have to resort to other libraries (e.g. scipy) to carry them out, which involves moving the data from the GPU to the CPU and can therefore result in a bottleneck.
## Pitch
Implement basic morphological operations in pytorch. This should be very straight-forward, since the code can basically be translated from numpy to torch based on other implementations (e.g. scipy). If this feature request is accepted, then I would be glad to help with the implementation.
|
feature,triaged
|
low
|
Major
|
629,136,039 |
vscode
|
Full view title isn't shown when a contextual title is defined
|
Issue Type: <b>Bug</b>
Found while testing: https://github.com/microsoft/vscode/issues/98996
Create a view with a contextual title
Move the view from activity bar to horizontal panel
Note that the contextual title is shown but full title is not
Click on view
Notice now that full view name is shown
Click on another view in horizontal panel
Notice that full name of newly added view remains shown even though this was not the case when the view was first added to the panel
See animated gif:
VS Code version: Code - Insiders 1.46.0-insider (1bfa086adb9271ff56be11821da686e4cfffb672, 2020-06-02T06:55:33.079Z)
OS version: Windows_NT x64 10.0.18363
<details>
<summary>System Info</summary>
|Item|Value|
|---|---|
|CPUs|Intel(R) Core(TM) i7-8650U CPU @ 1.90GHz (8 x 2112)|
|GPU Status|2d_canvas: enabled<br>flash_3d: enabled<br>flash_stage3d: enabled<br>flash_stage3d_baseline: enabled<br>gpu_compositing: enabled<br>multiple_raster_threads: enabled_on<br>oop_rasterization: disabled_off<br>protected_video_decode: enabled<br>rasterization: enabled<br>skia_renderer: disabled_off_ok<br>video_decode: enabled<br>viz_display_compositor: enabled_on<br>viz_hit_test_surface_layer: disabled_off_ok<br>webgl: enabled<br>webgl2: enabled|
|Load (avg)|undefined|
|Memory (System)|15.93GB (1.56GB free)|
|Process Argv||
|Screen Reader|no|
|VM|0%|
</details><details><summary>Extensions (24)</summary>
Extension|Author (truncated)|Version
---|---|---
vscode-eslint|dba|2.1.5
EditorConfig|Edi|0.15.1
vscode-github-releases-updater|joa|0.18.0
vscode-peacock|joh|3.7.2
chat|kar|0.33.0
vscode-docker|ms-|1.2.1
csharp|ms-|1.22.0
python|ms-|2020.5.80290
remote-containers|ms-|0.119.0
remote-ssh|ms-|0.51.0
remote-ssh-edit|ms-|0.51.0
remote-wsl|ms-|0.44.2
vscode-remote-extensionpack|ms-|0.20.0
azure-account|ms-|0.8.11
js-debug-companion|ms-|1.0.1
vscode-github-issue-notebooks|ms-|0.0.25
cloudenv-extensions|ms-|1.0.1369
vsliveshare|ms-|1.0.2236
vsliveshare-audio|ms-|0.1.85
vsliveshare-pack|ms-|0.3.4
vsonline|ms-|1.0.2268
debugger-for-chrome|msj|4.12.8
test-search-rg|rob|0.0.3
code-spell-checker|str|1.9.0
</details>
<!-- generated by issue reporter -->
|
bug,workbench-views
|
low
|
Critical
|
629,149,584 |
angular
|
ServiceWorker doesn't update `index.html` when headers (e.g. CSP) change
|
<!--π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
Oh hi there! π
To expedite issue processing please search open and closed issues before submitting a new one.
Existing issues often contain information about workarounds, resolution, or progress updates.
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
π
-->
# π bug report
### Affected Package
The issue is caused by package @angular/service-worker:9.1.7
### Is this a regression?
Nope, seems to always been there
### Description
Supposition: It seems that the service worker use body response only to generate its hash, corresponding to versioned resources ; update some headers server side doesn't update these hashes and resources are served with old headers.
## π¬ Minimal Reproduction
When deploying Angular application with service-worker (default configuration):
- Load the main page in a browser
- Service worker do its job and cache resources (including index.html)
- Change global resources header (in my case, nginx configuration, like `add_header X-OneThing "hello there"`
- Let the service worker try to do a cache bust
- **Take a look on the answer : no update detected**
- Refresh the page : **the new header isn't applied** (but the ngsw.json HTTP call got it, of course)
Sorry but cannot do a stackblitz with that :/
## π₯ Exception or Error
None
## π Your Environment
**Angular Version:**
<pre><code>
> ng "version"
_ _ ____ _ ___
/ \ _ __ __ _ _ _| | __ _ _ __ / ___| | |_ _|
/ β³ \ | '_ \ / _` | | | | |/ _` | '__| | | | | | |
/ ___ \| | | | (_| | |_| | | (_| | | | |___| |___ | |
/_/ \_\_| |_|\__, |\__,_|_|\__,_|_| \____|_____|___|
|___/
Angular CLI: 9.1.6
Node: 12.14.1
OS: linux x64
Angular: 9.1.7
... animations, common, compiler, compiler-cli, core, forms
... language-service, platform-browser, platform-browser-dynamic
... router
Ivy Workspace: Yes
Package Version
-----------------------------------------------------------
@angular-devkit/architect 0.901.6
@angular-devkit/build-angular 0.901.6
@angular-devkit/build-optimizer 0.901.6
@angular-devkit/build-webpack 0.901.6
@angular-devkit/core 9.1.6
@angular-devkit/schematics 9.1.6
@angular/cli 9.1.6
@angular/platform-server 9.1.9
@angular/service-worker 9.1.9
@ngtools/webpack 9.1.6
@schematics/angular 9.1.6
@schematics/update 0.901.6
rxjs 6.5.5
typescript 3.8.3
webpack 4.42.0
</code></pre>
|
freq1: low,workaround2: non-obvious,area: service-worker,state: confirmed,type: use-case,P3
|
medium
|
Critical
|
629,175,481 |
vscode
|
Settings sync - Cannot dismiss "Configure..." quick pick dialog
|
Testing #98962
* Click `Configure...` action in the `Preferences sync` toolbar
* Click anywhere else in the workbench
Expected: Quick pick dialog should hide (same as the command palette)
Actual: Only way to dismiss the quick pick dialog is using the `Ok` button
|
quick-pick,under-discussion
|
medium
|
Major
|
629,253,023 |
flutter
|
showDatePicker ignore ThemeData.primaryColor
|
## Steps to Reproduce
1. Update flutter from 1.12 to 1.17 channel stable
2. Our widget with showDatePicker does no more work as before
Code :
```
Future<DateTime> _getDate(BuildContext context, [DateTime firstDate, DateTime lastDate, DateTime initialDate]) async{
DateTime fDate = firstDate != null ? DateTime(firstDate.year, firstDate.month, firstDate.day) : DateTime.now().subtract(Duration(days: 36500));
DateTime lDate = lastDate != null ? DateTime(lastDate.year, lastDate.month, lastDate.day) : DateTime.now().add(Duration(days: 3650));
DateTime iDate = initialDate != null ? DateTime(initialDate.year, initialDate.month, initialDate.day) : null;
return await showDatePicker(
initialDatePickerMode: isDateYear ? DatePickerMode.year : DatePickerMode.day,
context: context,
firstDate: fDate,
lastDate: lDate,
initialDate: (iDate != null && (iDate.isAfter(fDate) || iDate == fDate))
? iDate
: (iDate == null && DateTime.now().isBefore(fDate))
? fDate
: DateTime.now(),
locale: Locale('fr', 'FR'),
**//code added for 1.17**
builder: (BuildContext context, Widget child) {
return Theme(
data: Theme.of(context),
child: child,
isMaterialAppTheme: true,
);
},
);
}
```
There are no errors but
- theme is lost
- "SELECT DATE" displayed in the header, it's in english whereas my locale is french. And we can't set it to null to remove that helpText.
**Expected results:** <!-- what did you want to see? -->
**Actual results:** <!-- what did you see? -->
<details>
<!-- Finally, paste the output of running `flutter doctor -v` here. -->
```
[β] Flutter (Channel stable, v1.17.2, on Mac OS X 10.14.6 18G4032, locale fr-FR)
β’ Flutter version 1.17.2 at /Users/s28091/Tools/flutter
β’ Framework revision 5f21edf8b6 (5 days ago), 2020-05-28 12:44:12 -0700
β’ Engine revision b851c71829
β’ Dart version 2.8.3
[β] Android toolchain - develop for Android devices (Android SDK version 29.0.3)
β’ Android SDK at /Users/s28091/Library/Android/sdk
β’ Platform android-R, build-tools 29.0.3
β’ ANDROID_HOME = /Users/s28091/Library/Android/sdk
β’ Java binary at: /Applications/Android Studio.app/Contents/jre/jdk/Contents/Home/bin/java
β’ Java version OpenJDK Runtime Environment (build 1.8.0_212-release-1586-b4-5784211)
β’ All Android licenses accepted.
[β] Xcode - develop for iOS and macOS (Xcode 11.3.1)
β’ Xcode at /Applications/Xcode.app/Contents/Developer
β’ Xcode 11.3.1, Build version 11C505
β’ CocoaPods version 1.9.2
[β] Android Studio (version 3.6)
β’ Android Studio at /Applications/Android Studio.app/Contents
β’ Flutter plugin version 45.1.1
β’ Dart plugin version 192.8052
β’ Java version OpenJDK Runtime Environment (build 1.8.0_212-release-1586-b4-5784211)
[β] IntelliJ IDEA Ultimate Edition (version 2020.1.1)
β’ IntelliJ at /Applications/IntelliJ IDEA.app
β’ Flutter plugin version 45.1.3
β’ Dart plugin version 201.7223.99
[β] Proxy Configuration
β’ HTTP_PROXY is set
β’ NO_PROXY is localhost,127.0.0.1
β’ NO_PROXY contains 127.0.0.1
β’ NO_PROXY contains localhost
[β] Connected device (2 available)
β’ Android SDK built for x86 β’ emulator-5554 β’ android-x86 β’ Android 10 (API 29)
(emulator)
β’ iPhone 11 Pro Max β’ F27C72A5-9D88-40D0-8F6B-5B624E0239A6 β’ ios β’
com.apple.CoreSimulator.SimRuntime.iOS-13-3 (simulator)
β’ No issues found!
```
</details>
Thanks
|
framework,f: material design,f: date/time picker,a: annoyance,has reproducible steps,P2,found in release: 3.3,found in release: 3.7,team-design,triaged-design
|
low
|
Critical
|
629,256,822 |
flutter
|
Flutter web dropdown not suggesting item when i type keyboard keys
|
On web, it is expected that when you are filling a form and there is a dropdown, that you can type and get the item that matches what you type.
This issue was already created at #55438 but it was wrongly closed as a duplicate of #46824
That issue is talking about providing a text autocomplete widget, which is not the case here. This needs to keep the Dropdown widget, but respond to the keyboard strokes.
Tried to have the original issue reopened but since I had no response I'm opening a new one to get more attention.
|
c: new feature,framework,f: material design,a: fidelity,P3,team-design,triaged-design
|
low
|
Major
|
629,278,958 |
pytorch
|
Update RPC doc to recommend async user functions for non-blocking server execution
|
When the PR that enables async user function for `rpc.remote` is added to the stack started from #39216, update the master `rpc.rst` doc to mention it in `rpc.rpc_sync`, `rpc.rpc_async`, and `rpc.remote` APIs.
cc @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @gqchen @aazzolini @rohan-varma @xush6528 @jjlilley @osalpekar
|
module: docs,triaged,module: rpc
|
low
|
Minor
|
629,305,906 |
vscode
|
Custom menus should absorb keystrokes and dismiss if not mnemonic (letter)
|
Testing #97801
Choose the step into targets command
See a list of targets appear
Ignore list and press F10
List of targets remains in view
Expected: list of targets to close
|
bug,menus
|
low
|
Minor
|
629,317,801 |
flutter
|
[web] Textfield 'autofocus: true' doesn't show up the keyboard
|
It only gets in the focused mode which I know by changing the TextField `focusedBorder` property that it's being listened to but no keyboard there.
<details>
<summary>doctor</summary>
```
[β] Flutter (Channel dev, 1.19.0-2.0.pre, on Mac OS X 10.15.5 19F101, locale es-ES)
β’ Flutter version 1.19.0-2.0.pre at /Users/tomas/Development/flutter
β’ Framework revision 1d395c5e18 (2 days ago), 2020-05-31 07:41:50 -0700
β’ Engine revision 923687f0e7
β’ Dart version 2.9.0 (build 2.9.0-11.0.dev 6489a0c68d)
[β] Android toolchain - develop for Android devices (Android SDK version 29.0.3)
β’ Android SDK at /Users/tomas/Library/Android/sdk
β’ Platform android-29, build-tools 29.0.3
β’ Java binary at: /Applications/Android Studio.app/Contents/jre/jdk/Contents/Home/bin/java
β’ Java version OpenJDK Runtime Environment (build 1.8.0_212-release-1586-b4-5784211)
β’ All Android licenses accepted.
[β] Xcode - develop for iOS and macOS (Xcode 11.5)
β’ Xcode at /Applications/Xcode.app/Contents/Developer
β’ Xcode 11.5, Build version 11E608c
β’ CocoaPods version 1.9.1
[β] Chrome - develop for the web
β’ Chrome at /Applications/Google Chrome.app/Contents/MacOS/Google Chrome
[β] Android Studio (version 3.6)
β’ Android Studio at /Applications/Android Studio.app/Contents
β’ Flutter plugin version 44.0.2
β’ Dart plugin version 192.7761
β’ Java version OpenJDK Runtime Environment (build 1.8.0_212-release-1586-b4-5784211)
[β] VS Code (version 1.45.1)
β’ VS Code at /Applications/Visual Studio Code.app/Contents
β’ Flutter extension version 3.11.0
[β] Connected device (3 available)
β’ iPhone 11 β’ 04298E38-2633-4586-B868-1FBD5C708AF7 β’ ios β’ com.apple.CoreSimulator.SimRuntime.iOS-13-5 (simulator)
β’ Web Server β’ web-server β’ web-javascript β’ Flutter Tools
β’ Chrome β’ chrome β’ web-javascript β’ Google Chrome 83.0.4103.61
β’ No issues found!
```
</details>
|
a: text input,framework,customer: crowd,platform-web,f: focus,has reproducible steps,P3,browser: safari-ios,browser: chrome-android,workaround available,team-web,triaged-web,found in release: 3.19,found in release: 3.20
|
high
|
Critical
|
629,362,037 |
vscode
|
Find all references is not highlighting the match
|
Testing #98821
Sorry that I did not look for dups, if it is please close it.
<img width="1049" alt="image" src="https://user-images.githubusercontent.com/10746682/83547731-f0ff4400-a502-11ea-8295-4a70faa07508.png">
|
feature-request,under-discussion,references-viewlet,notebook-workbench-integration
|
low
|
Minor
|
629,367,397 |
flutter
|
Fuchsia debug .build-id archive expects to only have the unstripped binaries
|
Currently we have `1e/1317a6715a2639ed43e257e258c754cf514a08` (stripped) and `1e/1317a6715a2639ed43e257e258c754cf514a08.debug` (unstripped). Fuchsia expects
to only have `1e/1317a6715a2639ed43e257e258c754cf514a08.debug`.
cc: @joshuaseaton
|
customer: fuchsia,engine,P3,team-engine,triaged-engine
|
low
|
Critical
|
629,370,907 |
PowerToys
|
[FancyZones] window must resist dragging
|
# Summary of the new feature/enhancement
The feature should work like (or replace) Windows Snap in every single way. So, when a window is opened inside a FancyZone it remains static unless considerable dragging happens (just like Windows Snap)
# Proposed technical implementation details (optional)
Just make it like Windows Snap! xD
|
Idea-Enhancement,FancyZones-Dragging&UI,Product-FancyZones
|
low
|
Minor
|
629,373,173 |
vscode
|
Problems view do not show markers when Show Active File filter is on
|
Testing #98821
- Enable the filter `Show Active File Only` in Problems view using the filter dropdown in filter input box
- Open any notebook with problems
π Problems are not shown
<img width="1346" alt="image" src="https://user-images.githubusercontent.com/10746682/83549389-8e5b7780-a505-11ea-9682-894c69534932.png">
|
feature-request,under-discussion,languages-diagnostics,notebook-workbench-integration
|
low
|
Minor
|
629,379,401 |
vscode
|
Partial cross-file undo if a file is changed in the meantime
|
Testing #98987
Repro:
1. Rename symbol to update 3+ files
2. Edit one file on disk
3. Undo on another file
:bug: => cross-file undo does not work at all. Could we fix a majority of the files that don't have this issue?
|
feature-request,undo-redo
|
low
|
Critical
|
629,394,618 |
TypeScript
|
Export Typescript libraries as individual interfaces as well
|
## Search Terms
export library interface
## Suggestion
Right now Typescript declares it's libraries using `d.ts` (e.g. `lib.dom.d.ts`) so if you include this in your project you get the entire DOM library and it's interfaces. This is great however it makes development of isomorphic code harder.
Your only option is to add the `dom` library to your NodeJS project. This means that consumers could accidentally write unsafe code.
My proposal is really simple. Along with existing d.ts imports, we also export (in a separate file) the interfaces for those libraries
## Use Cases
Isomorphic code could be be implemented in an easier way. Right now we need to either include the whole library or use dependency injection. Which is not always trivial.
Here is an example of an issue encountered from the puppeteer repo - https://github.com/DefinitelyTyped/DefinitelyTyped/issues/24419#issuecomment-637700421
## Examples
```ts
import { Window } from 'typescript/interface/lib.dom'; // proposed path
window.alert('something'); // this will throw TS errors
page.evaluate(() => {
// this will be fine and type checked according to window
(window as Window).alert('something');
});
```
## 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
|
629,398,464 |
PowerToys
|
Tool to redo file organization
|
# Invert the order of a directory tree
Windows Explorer is frequently used as an archive system for files. After years of using a specific system sometimes the method becomes cumbersome or the access patterns change.
It would be nice if PowerRename could change the organizational method of a directory tree and rename the files with respect to their parent folder.
## For example:
Files organized by: Year > Subject > Class. Maybe user would want to invert to: Subject > Year > Class.
Directory
ββββ1
β ββββA
β ββββB
β ββββC
β ββββD
ββββ2
β ββββA
β ββββB
β ββββC
ββββ3
β ββββB
β ββββC
β ββββD
β ββββE
β ββββF
ββββ4
β ββββB
β ββββC
β ββββD
β ββββE
β ββββF
ββββ5
ββββD
ββββE
ββββF
**Could become:**
Inverted
ββββA
β ββββ1
β ββββ2
ββββB
β ββββ1
β ββββ2
β ββββ3
β ββββ4
ββββC
β ββββ1
β ββββ2
β ββββ3
β ββββ4
ββββD
β ββββ1
β ββββ3
β ββββ4
β ββββ5
ββββE
β ββββ3
β ββββ4
β ββββ5
ββββF
ββββ3
ββββ4
ββββ5
|
Idea-New PowerToy,Product-File Explorer,Product-File Actions Menu
|
low
|
Minor
|
629,423,342 |
godot
|
Massive visual bug (pink screen) with ssao and blur on Android
|
**Godot version:**
3.2 from Steam
**OS/device including version:**
iPhone 11
iPhone 7
Samsung Galaxy a51
**Issue description:**
On Android I get a pink screen whenever I turn ssao or blur on in my Environment. This does not happen on iOS. I have no idea what causes this since I only have experience with iOS.
The performance on Anrdoid is also terrible compared to iOS. On an iPhone 11 I get 50+ fps while the a51 is struggling to reach 10 fps. This makes no sense since the a51 is 2x faster on paper.
I'm using GLES3 wich is normally not advised and I've seen comments on every other Anrdoid performance issue to use GLES2. I find this feedback useless since it doesn't solve the issue, it's only avoiding it. Why can't it run on Android when iOS is at 60 fps?
So the two issues are the pink screen and the bad performance.
**Steps to reproduce:**
Go to your project settings >Rendering > Quality > Itended Usage and set it to 3D. Then you just need to add a WorldEvironment and turn on ssao or blur. You always get a pink screen.
|
bug,topic:rendering
|
low
|
Critical
|
629,436,233 |
godot
|
macOS crash when press play button to build the game [Mono]
|
<!-- Please search existing issues for potential duplicates before filing yours:
https://github.com/godotengine/godot/issues?q=is%3Aissue
-->
**Godot version:**
<!-- Specify commit hash if using non-official build. -->
Godot_mono_3.2.2 Beta 3, using C#
**OS/device including version:**
<!-- Specify GPU model, drivers, and the backend (GLES2, GLES3, Vulkan) if graphics-related. -->
GLES2 , Macbook pro, Catalina (check the [crash.txt](https://github.com/godotengine/godot/files/4719073/crash.txt)
for more details
**Issue description:**
<!-- What happened, and what was expected. -->
Sometimes (seems to happen randomly), when I press the "play" button to build and run the game, an emergent dialog asking for admin user/password shows. If you press cancel button, godot editor crash (check the crash report), if you enter user/password, a process in the background uses 100% of CPU, and you can't kill, even with kill -9 pid. The only way to kill is by rebooting the macbook pro.
**Steps to reproduce:**
It's randomly
**Crash Report**
[crash.txt](https://github.com/godotengine/godot/files/4719073/crash.txt)
|
bug,platform:macos,topic:dotnet,crash
|
low
|
Critical
|
629,494,349 |
rust
|
Add an attribute for asserting variance of generic parameters
|
Currently, we don't have any kind of programmatic way of ensuring that we don't accidentally change the variances of any `libcore`/`libstd` types. While some types do have explicit tests for variance (e.g. [here](https://github.com/rust-lang/rust/commit/bf592cefde313aed68bccc91c053f5912350249b#diff-f9e71018920a56a1532da2e8fb8d02cdR1005-R1007) and in PR https://github.com/rust-lang/rust/pull/72902), this is not done in any kind of systematic way.
I think it would be useful to have an internal `#[variance(Co/Contra/In)]` (names bikesheddable) which causes an error if the actual and expected variances do not match.
We could require that every generic parameter on every public type have a `#[variance]` attribute, similar to how we require `#[stable]`/`#[unstable]`. However, this could end up being unnecessarily verbose, especially if most types are covariant. Instead, we might want to treat the absence of a `#[variance]` annotation as an implicit `#[variance(Covariant)]` (or `#[variance(Invariant)]` for const generics). If covariance is normally what we want, then this would only require us to annotate cases where something 'unusual' is happening. Of course, we would run the risk of cases like https://github.com/rust-lang/rust/pull/71814 'slipping under the radar', where the generic parameter really should have been invariant to begin with.
|
C-enhancement
|
low
|
Critical
|
629,495,755 |
pytorch
|
Allow to_here to interrupt blocking wait in the case of rpc.remote timeout
|
## π Feature
https://github.com/pytorch/pytorch/pull/38590/files is adding support for RRef timeouts, although, this can be improved in the case of `to_here`. Right now, we only check if the rpc.remote call for `to_here` timed out before making a blocking call on waiting on the corresponding future for `to_here`. However, we instead have to_here wait on 2 conditions:
1) [Happy path] the `to_here` future successfully finishes
2) [Timeout]: `isTimedOut()` (added in the mentioned PR) is set to true. We can probably handle these conditions by signalling on a condvar.
This will make timeouts in the case of `to_here` more robust, as we will be able to interrupt the blocking call when we get notified that the original rpc.remote() creation has failed.
cc @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @gqchen @aazzolini @rohan-varma @jjlilley @osalpekar @jiayisuse @xush6528
|
feature,triaged,module: rpc,pt_distributed_rampup
|
low
|
Critical
|
629,523,448 |
rust
|
Recursive diagnostics when using consts in arrays
|
<!--
Thank you for filing a bug report! π Please provide a short summary of the bug,
along with any information you feel relevant to replicating the bug.
-->
I tried this code:
```rust
trait A {
const N: usize;
const ARRAY: [u8; Self::N];
}
trait B<T: A> {
fn f() -> [u8; <T as A>::N];
}
```
[Playground](https://play.rust-lang.org/?version=nightly&mode=debug&edition=2018&gist=967077056c818dba1e90e0dfc42ebc96)
I expected to see this happen: the compiler gives useful diagnostics
Instead, this happened: the compiler gives diagnostics which don't work
```
Compiling playground v0.0.1 (/playground)
error[E0599]: no associated item named `N` found for type parameter `Self` in the current scope
--> src/lib.rs:3:29
|
3 | const ARRAY: [u8; Self::N];
| ^ associated item not found in `Self`
|
= help: items from traits can only be used if the type parameter is bounded by the trait
help: the following trait defines an item `N`, perhaps you need to add a supertrait for it:
|
1 | trait A: A {
| ^^^
error[E0277]: the trait bound `T: A` is not satisfied
--> src/lib.rs:7:20
|
2 | const N: usize;
| --------------- required by `A::N`
...
7 | fn f() -> [u8; <T as A>::N];
| ^^^^^^^^^^^ the trait `A` is not implemented for `T`
|
help: consider further restricting this bound
|
6 | trait B<T: A + A> {
| ^^^
error: aborting due to 2 previous errors
Some errors have detailed explanations: E0277, E0599.
For more information about an error, try `rustc --explain E0277`.
error: could not compile `playground`.
To learn more, run the command again with --verbose.
```
### Meta
<!--
If you're using the stable version of the compiler, you should also check if the
bug also exists in the beta or nightly versions.
-->
`rustc --version --verbose`:
```
rustc 1.45.0-nightly (74e804683 2020-05-30)
binary: rustc
commit-hash: 74e80468347471779be6060d8d7d6d04e98e467f
commit-date: 2020-05-30
host: x86_64-pc-windows-msvc
release: 1.45.0-nightly
LLVM version: 10.0
```
|
A-diagnostics,A-trait-system,A-associated-items,T-compiler,D-papercut
|
low
|
Critical
|
629,528,880 |
rust
|
Refactor `non_ssa_locals` to remove `LocalAnalyzer::process_place`
|
While doing an initial implementation of rust-lang/compiler-team#300, I began replacing all uses of `{Non,}MutatingUseContext::Projection`. One user was [`rustc_codegen_ssa::mir::analyze::LocalAnalyzer`](https://github.com/rust-lang/rust/blob/fe10f1a49f5ca46e57261b95f46f519523f418fe/src/librustc_codegen_ssa/mir/analyze.rs#L49), which is doing a custom recursive traversal in [`process_place`](https://github.com/rust-lang/rust/blob/master/src/librustc_codegen_ssa/mir/analyze.rs#L99). There's a [`HACK` comment](https://github.com/rust-lang/rust/blob/fe10f1a49f5ca46e57261b95f46f519523f418fe/src/librustc_codegen_ssa/mir/analyze.rs#L184-L186) from @eddyb suggesting that this should be rewritten.
@eddyb, did you have something specific in mind? I spent some time trying to decipher the logic here, but there's parts I don't understand. For example, there's [a long comment](https://github.com/rust-lang/rust/blob/fe10f1a49f5ca46e57261b95f46f519523f418fe/src/librustc_codegen_ssa/mir/analyze.rs#L150-L170) about why we need to call `visit_local` for `NonUseContext::VarDebugInfo`, but `visit_local` [does nothing for `NonUse`s](https://github.com/rust-lang/rust/blob/fe10f1a49f5ca46e57261b95f46f519523f418fe/src/librustc_codegen_ssa/mir/analyze.rs#L277). What's the indended behavior here?
Additionally, `LocalAnalyzer` visits basic blocks in numerical order, but [this code](https://github.com/rust-lang/rust/blob/fe10f1a49f5ca46e57261b95f46f519523f418fe/src/librustc_codegen_ssa/mir/analyze.rs#L286-L291) assumes we will visit the assignment to a local before any uses of it. This isn't a soundness issue; The worst thing that can happen is that locals are placed onto the stack unnecessarily. However, I think we should be visiting basic blocks in RPO, no?
|
C-cleanup,A-codegen,T-compiler
|
low
|
Critical
|
629,559,409 |
electron
|
Crash on start leaves orphaned processes
|
### 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:** Electron 9, 8
* **Operating System:** Windows 10 (1909)
### Expected Behavior
I expect that Electron cleans up processes if the app crashes on start
### Actual Behavior
It leaks the main, GPU, and network service processes
### To Reproduce
main.js:
```
throw new Error("test");
```
Run that program. In [Process Explorer](https://docs.microsoft.com/en-us/sysinternals/downloads/process-explorer), search for "electron.exe" (doesn't seem to show up in Task Manager for me) and observe the 3 leaked processes.
It seems that I'm forced to add an `uncaughtException` handler at the very top to exit.
```
process.on("uncaughtException", (err) => {
throw err;
});
```
Which I'd prefer not to do. I was seeing a crash in my app at the very top of main.js because of one of the requires, so I'd have to add this handler above the requires even which is strange to me.
Can Electron handle this gracefully?
|
platform/windows,bug :beetle:,8-x-y,9-x-y,10-x-y
|
medium
|
Critical
|
629,567,196 |
electron
|
Implement BrowserWindow.setOpacity on Linux
|
<!-- 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. -->
BrowserWindow.setOpacity is not implemented on Linux, only on Mac and Windows.
### Proposed Solution
I have a local patch and I will submit a PR.
### Alternatives Considered
<!-- A clear and concise description of any alternative solutions or features you've considered. -->
None
### Additional Information
<!-- Add any other context about the problem here. -->
None
|
enhancement :sparkles:,platform/linux
|
low
|
Minor
|
629,574,330 |
electron
|
Implement electron.DesktopCapturer.SetSkipCursor(DesktopCapturerSource.id)
|
<!-- 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. -->
Until https://bugs.chromium.org/p/chromium/issues/detail?id=1007177 is implemented we need a way to hide the cursor when capturing a Window. And we need to hide and show dynamically as the source continued to be captured (i.e. without restarting the capture)
### Proposed Solution
<!-- Describe the solution you'd like in a clear and concise manner -->
I have a local patch and I will submit a PR, the api looks like:
electron.DesktopCapturer.SetSkipCursor(DesktopCapturerSource.id)
### Alternatives Considered
<!-- A clear and concise description of any alternative solutions or features you've considered. -->
None
### Additional Information
<!-- Add any other context about the problem here. -->
Older bug https://github.com/electron/electron/issues/7584
|
enhancement :sparkles:
|
low
|
Critical
|
629,577,888 |
bitcoin
|
Implement PayJoin / Pay-to-EndPoint
|
PayJoin (also called pay-to-end-point or P2EP) is a special type of CoinJoin between two parties where one party pays the other. This coinjoin type has different (probably better) privacy properties. The transaction then doesn't have the distinctive multiple outputs with the same value, and so is not obviously visible as an equal-output CoinJoin.
An overview of PayJoin with additional information and links can be found here https://en.bitcoin.it/wiki/PayJoin
Implementing PayJoin would potentially need OpenSSL to communicate with the PayJoin server (unless I am misunderstanding how PayJoin works) as such, implementing this would have similar objections to BIP 70.
Looking forward to hearing thoughts.
|
Feature
|
low
|
Major
|
629,584,426 |
pytorch
|
torch::jit::script::Module::to(torch::kDouble) also casts buffers
|
## π Documentation
The C++ API documentation for ``torch::jit::script::Module::to(at::Device, at::ScalarType, bool)``
indicates that the function casts all Parameters to ``at::ScalarType``, however, the function also casts the Module's buffers.
The documentation in question is in:
https://pytorch.org/cppdocs/api/structtorch_1_1jit_1_1_module.html#_CPPv4N5torch3jit6Module2toEN2at6DeviceEN2at10ScalarTypeEb
And you can see the behaviour of the function in:
https://github.com/pytorch/pytorch/blob/master/torch/csrc/jit/api/module.cpp
relevant code:
```
void Module::to_impl(
const c10::optional<at::Device>& device,
const c10::optional<at::ScalarType>& dtype,
bool non_blocking) {
for (at::Tensor e : parameters()) {
module_state_to(e, device, dtype, non_blocking);
}
for (at::Tensor e : buffers()) {
module_state_to(e, device, dtype, non_blocking);
}
}
```
I'm not sure if this is a bug, a documentation issue, or I'm being a bit pedantic, so I'm raising it as a documentation issue, but I believe the behaviour of the python API is different. IIRC model.double() leaves buffers of type ``torch.long`` unchanged, but buffers of type ``torch::kLong`` are cast to ``torch::kDouble`` by to() using the C++ API.
cc @suo
|
oncall: jit,triaged
|
low
|
Critical
|
629,584,468 |
flutter
|
When front end server unexpectedly exits "the Dart compiler exited unexpectedly", no additional crash info
|
Copied from https://github.com/flutter/flutter/issues/58468
```
[ +19 ms] "flutter run" took 9,975,116ms.
the Dart compiler exited unexpectedly.
[ +49 ms] Running shutdown hooks
```
No info about what the exit was or post-exit stdout/err draining before the `throwToolExit`.
|
team,tool,P2,team-tool,triaged-tool
|
low
|
Critical
|
629,598,308 |
neovim
|
Keep command-line pop-up menu open while typing
|
- `nvim --version`:
<details>
<summary>
output
</summary>
<p>
NVIM v0.4.3
Build type: Release
LuaJIT 2.0.5
Compilation: /usr/local/Homebrew/Library/Homebrew/shims/mac/super/clang -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=1 -DNDEBUG -DMIN_LOG_LEVEL=3 -Wall -Wextra -pedantic -Wno-unused-parameter -Wstrict-prototypes -std=gnu99 -Wshadow -Wconversion -Wmissing-prototypes -Wimplicit-fallthrough -Wvla -fstack-protector-strong -fdiagnostics-color=auto -DINCLUDE_GENERATED_DECLARATIONS -D_GNU_SOURCE -DNVIM_MSGPACK_HAS_FLOAT32 -DNVIM_UNIBI_HAS_VAR_FROM -I/tmp/neovim-20191107-13403-1or2rj3/neovim-0.4.3/build/config -I/tmp/neovim-20191107-13403-1or2rj3/neovim-0.4.3/src -I/usr/local/include -I/tmp/neovim-20191107-13403-1or2rj3/neovim-0.4.3/deps-build/include -I/usr/local/opt/gettext/include -I/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.15.sdk/usr/include -I/tmp/neovim-20191107-13403-1or2rj3/neovim-0.4.3/build/src/nvim/auto -I/tmp/neovim-20191107-13403-1or2rj3/neovim-0.4.3/build/include
Compiled by [email protected]
Features: +acl +iconv +tui
See ":help feature-compile"
system vimrc file: "$VIM/sysinit.vim"
fall-back for $VIM: "/usr/local/Cellar/neovim/0.4.3/share/nvim"
Run :checkhealth for more info
</p>
</details>
- `vim -u DEFAULTS` (version: ) behaves differently? Irrelevant since vim does
not have `wildoptions=pum`
- Operating system/version: Mac OS X 10.14.6
- Terminal name/version: kitty 0.17.4
- `$TERM`: xterm-kitty
### Steps to reproduce using `nvim -u NORC`
```
nvim -u NORC
:e<Tab><C-p>c
```
Or
```
nvim -u NORC
:set wildmode=longest:full
:e<Tab>c
```
### Actual behaviour
When you type the `c`, the completion popup menu closes.
### Expected behaviour
When you type the `c`, the completion popup menu should stay open and just be
filtered.
Like in insert mode in the command-line window (BTW this is with `set completeopt+=noselect`)
|
enhancement,completion,cmdline-mode
|
low
|
Critical
|
629,617,927 |
flutter
|
[web]: filter/blur incorrectly rendered by the HTML backend
|
To produce the effect on the red waves (that actually slowly continuously animate like waves when run) a MaskFilter.blur(BlurStyle.outer, 10) is used, I suspect it is the culprit that looks bad on DomCanvas, but until I isolate it and demonstrate it in a stand-alone example Iβm not 100% sure.
Expected (Canvaskit rendering):
Actual (Dom rendering):
|
engine,platform-web,c: rendering,e: web_html,has reproducible steps,customer: web10,P2,customer: house,found in release: 3.3,found in release: 3.6,team-web,triaged-web
|
low
|
Major
|
629,632,826 |
pytorch
|
Format issue in `torch.quantization.add_quant_dequant` documentation parameter section
|
## π Documentation
#### Link
https://pytorch.org/docs/stable/quantization.html#torch.quantization.add_quant_dequant
#### API
`torch.quantization.add_quant_dequant`
#### Issue
In the "Parameters" section, there is an input :
**we want to quantize** (*that*) β
which should be part of the description for input argument `module`
|
module: docs,triaged
|
low
|
Minor
|
629,643,303 |
godot
|
ViewportTexture flashes when mapped into emission
|
<!-- Please search existing issues for potential duplicates before filing yours:
https://github.com/godotengine/godot/issues?q=is%3Aissue
-->
**Godot version:**
<!-- Specify commit hash if using non-official build. -->
v3.2.1.stable.official
**OS/device including version:**
<!-- Specify GPU model, drivers, and the backend (GLES2, GLES3, Vulkan) if graphics-related. -->
macOS, Radeon Pro 560 4gb Graphics, GLES3.
**Issue description:**
<!-- What happened, and what was expected. -->
I'm mapping display a control node on a mesh in 3d by using ViewportTexture (as both albedo
and emission, with transparency on). The mesh in question is dynamically added to the scene.
When the mesh becomes visible - it seems that the texture is not immediately computed and it
flashes rather brightly once loaded.
**Steps to reproduce:**
1. Download [MeshFlash.zip](https://github.com/godotengine/godot/files/4720690/MeshFlash.zip)
2. Unzip
3. Open it in Godot
4. Run several times until it flashes (sometimes it doesn't flash on the first run).
**Minimal reproduction project:**
See above.
|
bug,topic:rendering,confirmed
|
low
|
Minor
|
629,672,570 |
TypeScript
|
import type from from "./specifier" fails to parse correctly
|
**TypeScript Version:** 4.0.0-dev20200616
**Search Terms:** import type declaration from name parsing
**Code**
```ts
import type from from "./a"
```
**Expected behavior:** Should parse as `import from from "./a"` parses fine.
**Actual behavior:** Parses as an import declaration and expression statement
This is such an edge case. Not sure if any user will actually run into this (doesn't affect me, but thought I should open it anyway... perhaps this is not even worth fixing).
|
Bug
|
low
|
Minor
|
629,691,698 |
godot
|
the game suddently crash without error
|
Godot version:
Godot v3.2.1.stable.official download from weisbite
Mono binding including version:
Running platform: pc windows 10
Language: C#
Issue description:
I followed a tutorial on youtube to make a terraria like tile light system, and found out when I tried to use SetShaderParam to enable the shader, the app will crash without error.
Steps to reproduce:
1. Add a Sprite with a texture and set up material and shader's file with below code:
shader_type canvas_item;
uniform sampler2D light_values;
void fragment() {
float alpha = 1.0 - texture(light_values, UV).g;
COLOR = vec4(vec3(0.0), alpha);
}
in the _Process method, call SetShaderParam to update its texture.
2.Added another sprite and a camera as its child. put this sprite at same level as the step 1 sprite.
3.For step 2 sprite, reset the sprite's scale or texture in its _Process method.
4.when the app is launched, you can see it will stopped for no reason. and can not find any error in the editor.
PS: This will only happened in the editor running mod, once we pack the app, and generate exe with release mod, it wont happen.
|
bug,topic:dotnet,crash
|
low
|
Critical
|
629,711,183 |
go
|
x/build/maintner: GitHubIssue.ID field is never populated
|
Similarly to #28745, the `ID` field of a `GitHubIssue` is always zero:
```Go
corpus, err := godata.Get(context.Background())
if err != nil {
log.Fatal(err)
}
num := 0
corpus.GitHub().ForeachRepo(func(gr *maintner.GitHubRepo) error {
return gr.ForeachIssue(func(gi *maintner.GitHubIssue) error {
if gi.ID != 0 {
num++
}
return nil
})
})
fmt.Printf("%d GitHub issues with non-zero IDs.\n", num)
// Output:
// 0 GitHub issues with non-zero IDs.
```
The field was added in [CL 37888](https://golang.org/cl/37888), but `GithubIssueMutation.ID` is not set anywhere. It should probably be set in the `githubIssueDiffer.Diff` method, with care taken that it's populated for both new and existing issues (it would be unhelpful to populate it only for new issues, as no one could rely on the field then).
An alternative fix is to document that it's always zero and/or to remove it.
This isn't high priority because the `Number` field is usable as a workaround; filing so I can look it up. (Fixing this is made harder by #37603.)
|
Builders,NeedsInvestigation
|
low
|
Critical
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.