I have successfully built a native image with Quarkus/Graal and I can run it in the terminal (no container yet). However, it crashes with "Segmentation fault (core dumped)" message. How can I get more information about the error? Are there flags to pass to the native executable or while building the image?
I am using GraalVM Version 20.2.0 (Java Version 11.0.8)
Any help is much appreciated.
Thanks.
This could be a bug in the native image implementation. Please report it to the GraalVM team: github.com/oracle/graal.
You can also try debugging the executable: https://github.com/oracle/graal/blob/master/substratevm/DebugInfo.md
You can also build native image with a different level of optimizations passing the flag -H:Optimize=N where N is 0, 1, 2. Maybe there will be a difference in behavior -- suggesting a bug in the compiler.
When building the native image are you using something like --report-unsupported-elements-at-runtime? If yes, try removing it, maybe it tells you what can be wrong. The same with the --allow-incomplete-classpath?
In general, the best is to report it to the GraalVM team.
It was a condition in my code that caused this behavior. After I fixed the code, native image works like a charm now. Thanks for your help.
Related
I'm using intel realsense SDK2.10.0 with D410 depth module to build a project. But faced with an immediate problem that I can't run even the hello world example they give us. With no problem is building the project, there is always internal error as:
Unhandled exception at 0x00007FFB60434008 in hello-world.exe: Microsoft C++ exception: rs2::error at memory location 0x000000B0AD1EF670.
This is my first-time use visual studio so that could be a naive problem but I can't identify it. Any help and idea?
I solved the problem. The point is that those samples didn't expect infrared image as input and the render code part is for RGB frames. I only have the depth module with the infrared camera with no RGB sensor. so the sample program can't handle it.
You have to heavily modify the example.hpp file to make it functional. So if you are not a professional user, don't try individual modules on your own, or at least for now.
Suggest to use Windows CMAKE tool to configure the build x64 environment for VS2015.
The CMAKE tool will help create the required librealsense VS projects for you to build those RealSense sample apps.
BR,
Freeman
I'm using node-webkit to build a self-contained web app demo, and for some reason, the -webkit-backface-visibility CSS3 property doesn't work – but only when I run the app on Windows, not on Mac.
When I say it doesn't work, I mean it's exactly as if the property isn't supported: when I flip an element over on its back, I see it reversed, rather than hidden.
The issue isn't only with my particular code, either, as online demos of -webkit-backface-visibility (like this one) don't work when I run them in the app on Windows:
Is this a bug in node-webkit? Is there any way I can get this to work?
Well, it sounds like a bug. It might be the same issue that causes WebGL not to be available with specific hardware/driver combinations on Windows:
https://github.com/rogerwang/node-webkit/wiki/Webgl-support-on-windows
On Windows with some hardware and driver, WebGL won't work until you copy D3DCompiler_43.dll and d3dx9_43.dll to node-webkit's directory, or install DirectX 9 redistributable.
For license reasons we cannot ship those DLLs.
This seems to have fixed a similar problem with backface-visibility reported here:
CSS: backface-visibility and perspective doesn't work
It WORKS! Thanks a lot! I've added two dll files two node-webkit package and it works!
There used to be a gui toolkit for experimenting with OpenCV image processing pipelines including code generation. I think it started with an "h". What is it called, and what happened to it?
I think I saw in the past a nice tool to test opencv functions, not so sure about the code generation.
Now i found this one, it looks decent:
http://code.google.com/p/piaf/
This one looks nice but experimental:
http://siggiorn.com/videopipe/
And these, but look poor and ancient:
http://code.google.com/p/cvpreprocessor/
http://code.google.com/p/open-rtvision/
EDIT:
This one is dedicated to feature detecting module:
http://introlab.github.io/find-object/
I think you refer to Harpia. It's available in Ubuntu applications if you enter OpenCV in SW repository (Ubuntu Software Center) search box
This is called HighGUI. You can find many docs in the web. Here's one of them: Load and Display an Image.
As we know that there are two modes to execute javascript in firefox:interpreter and JIT.
My question is that whether I can disable JIT while building firefox in linux,and if so,how?
Any useful suggestion will be appreciated , thanks!
Using the --disable-methodjit configure option might work: SpiderMonkey supports it, but it's not obvious to me whether it'll propagate from the main build to the SpiderMonkey configure step.
Building Firefox without JIT is broken. Sources do not track ENABLE_JIT & friends carefully.
I am trying to get an OpenGL ES demo app to build on Windows with Mesa. I'm confused about the EGL side of things.
When my project links, it can't find eglTerminate() (and other EGL functions).
I've checked for a libEGL.lib file but I don't have one. Do I have to build the src\egl code into a library as well?
Unless something changed recently, I'm not even sure how you got an ES DLL in the first place.