I am using VS2017 to develop a cross platform (ARM linux) project with OpenCc and I would like to use "Image Watch" extension to debug my images.
I've installed the extension, but when I try to open the Image Watch window (while debugging) Visual Studio freezes, with no going back.
Is there any incompatibility between VS/Image Watch/Linux development that I am not aware of?
PS: I have also tried with VS2015 with the same result.
Related
I am trying to run C++ code for Azure Kinect Body Tracking samples on Windows available from Microsoft on Git. I have latest NVIDIA drivers installed on my system. While trying to build the project in Visual Studio, NuGet restore for Microsoft.Azure.Kinect.BodyTracking.Dependencies.cuDNN fails for me with a message "could not find a part of path build/native/Microsoft.Azure.Kinect.BodyTracking.Dependencies.cuDNN.targets".
On the other hand, the body tracking viewer app works fine for me.
I was able to solve this by reinstalling the Visual Studio. Not sure what was failing but it got resolved after fresh install.
I am developing an app using Flutter, and I want to be able to test some functionality that behaves slightly different depending on the device directly on both an iPhone and an Android device.
I am using a Linux machine with Visual Studio Code as my main development machine (and do not want to change that), and on that machine Visual Studio Code is deploying to an Android emulator (or a real device).
Then I have a Mac, which has also got Visual Studio Code installed. I mounted the "lib" directory from the Linux machine via sshfs to the Mac (the rest of the Flutter project is just copied to there, since for example I do not want to share the build-folder). From there Visual Studio Code deploys the app to either the simulator or a real iPhone.
With this setup, whenever I change a file, Visual Studio Code on the Linux maching deploys directly to Android. However, since file system events do not get picked up on the Mac due to sshfs, Visual Studio Code on the Mac does not directly deploy. I only have to trigger a hot deployment though (which I mapped to F5), so even now it is only one keypress I have to do on the Mac.
My question is, can I somehow automate this last step? I was thinking about using ssh to connect from the Mac to Linux, and then pick up the filesystem events via inotify. But how could I then trigger a hot deployment on the Mac?
So, I have a project being built and test on Windows platform, since Visual Studio 2017 works and run program most comfortable for me in Windows. Now, that I have completed debug most part of the code, I'll have to change my platform built target to Linux.
I used an external project (lib) to help with the internet control, which it's built target is also Windows.
Both the project are written with under C++17 standards so cross-platform should work fine.
Following the instruction on Microsoft website, I had successfully built a test project on my remote Linux server.
However, all the tutorials seem to ask people to start up a new "Linux Project", in which I did for the test project. The question is, since I already have a project/solution created and the build target is Windows platform, is there a way to switch the build to Linux Remote Build?
I've tried opening project settings, but the "target build" option is always Windows 10 and is not changeable.
I followed the instruction here https://askubuntu.com/questions/616075/how-to-install-visual-studio-code-on-ubuntu to run Visual Studio Code https://code.visualstudio.com/
but nothing happens (no error message) when I double-click on application (named Code).
In Ubuntu 16.04 and later Visual Studio Code should show up in Ubuntu Software application if you search for "vscode" or it can be installed from the terminal by running the command sudo snap install code --classic. A snap in classic confinement behaves as a traditionally packaged application with full access to the system, and Visual Studio Code extensions are installed into the user's home directory.
Visual Studio Code is not an IDE, so it is not possible to create a project in Visual Studio Code, like it is in Visual Studio. On the other hand it is possible to run Python, C, C++, JavaScript, PHP, Java, R and some other programming language code blocks directly in Visual Studio Code using the Code Runner extension. Other extensions are required to get the debugging features of each programming language working.
Where you extracted code, you cannot double click on it. Easiest way is to right-click and select run. You can make the "code" icon executable by using the terminal.
I have a VC++ (unmanaged) project to which I added performance counters using manifest method. Everything is working great and I even have build steps that do manifest compilation and automatic registration using lodctr.exe tool with /m:manifest switch.
The problem that I'm running into is that I now need to make installer (VS2008 Installer project) do registration on install (unregistration on uninstall).
I tried to set it up by having a custom action that referenced lodctr.exe tool with CustomActionData set to /m:[TARGETDIR]PerfMon.manifest and similar setup using unlodctr.exe for custom uninstall action.
When installer runs, I do see a brief flash of cmd window (which is ok), but counters are not getting registered. I ran SysInternals' ProcMon tool, and I don't see lodctr.exe process even created.
I tested that running lodctr.exe manually on target machine registers the counters.
I googled for an example or explanation how to do it with Visual Studio installer project, but came up empty. It has some .NET oriented examples, but nothing for unmanaged, hard and cold VC++...
How can I register performance counters manifest using installer project?
Visual Studio installer project has very limited functionality. Starting from Visual Studio 2012 Visual Studio installer project is not supported, you should consider to use Windows Installer XML (WIX) instead.
Include your manifest into WIX project and use WIX's PerfCounterManifest with it.