{"id":35268,"date":"2025-03-25T17:19:55","date_gmt":"2025-03-25T17:19:55","guid":{"rendered":"https:\/\/devblogs.microsoft.com\/cppblog\/?p=35268"},"modified":"2025-03-25T17:21:33","modified_gmt":"2025-03-25T17:21:33","slug":"boost-your-cmake-development-with-copilot-custom-instructions","status":"publish","type":"post","link":"https:\/\/devblogs.microsoft.com\/cppblog\/boost-your-cmake-development-with-copilot-custom-instructions\/","title":{"rendered":"Boost Your CMake Development with Copilot Custom Instructions"},"content":{"rendered":"

Introduction<\/h3>\n

Creating a new CMake project that uses unfamiliar libraries can be daunting and time-consuming. This blog post takes you along on my journey using Copilot to make this easier, and leveraging Custom Instructions<\/a> to tailor Copilot responses. For a sneak peek, see the outcome of this journey in this repository<\/a>.<\/p>\n

Background<\/h3>\n

As C++ developers who use Copilot every day, we\u2019ve experienced how it can improve our day-to-day productivity. Copilot really shines in scenarios like new codebases or unfamiliar topics, significantly reducing ramp-up. There are multiple ways to engage with it in Visual Studio, including Completions<\/a> and Copilot Chat<\/a>. This blog is focused on leveraging Copilot Chat to be more productive.<\/p>\n

What project should I create?<\/h3>\n

Being a maintainer of the CMake experience, I have always been aware of a popular computer vision library, OpenCV<\/a>, that is built on CMake to power 2500+ algorithms. However, I\u2019ve never actually played with it and used it in my own code. The first idea that came to mind, being a dog-lover myself, was to create my own real-time dog detector using OpenCV. I set out to create a Dog Detector using CMake and Copilot in Visual Studio.<\/p>\n

Import OpenCV into CMake Project<\/h3>\n

Starting with a new CMake Project in Visual Studio, the first problem to solve was: \u201cHow do I import OpenCV into my CMake project?\u201d. Using CMake can sometimes be complex, and I wanted to make sure I used OpenCV in a cross-platform way, so to figure out how to do this, I asked Copilot.<\/p>\n

\"Asking<\/a><\/p>\n

Copilot was able to give me CMakeLists.txt modifications that I can make to my project. However, you\u2019ll notice that it mentions that OpenCV must be installed and that I can use vcpkg. Vcpkg is a cross-platform C\/C++ package manager<\/a>.<\/p>\n

Prior to doing this, while the suggested CMakeLists.txt changes are correct, CMake fails the configure step with an error indicating that OpenCV cannot be found. This is because I haven\u2019t yet installed OpenCV.<\/p>\n

Use vcpkg to install OpenCV<\/h3>\n

Now, how do I take Copilot\u2019s suggestion to use vcpkg to install OpenCV? Let\u2019s ask Copilot.<\/p>\n

Copilot gave the following response:<\/p>\n

\"Image<\/a><\/p>\n

You\u2019ll notice that it gave only command-line suggestions for how to install OpenCV with vcpkg. Our team typically uses vcpkg in manifest mode<\/a> and I\u2019d prefer to use that. Also, you\u2019ll notice that Copilot suggested a CMake command line. We are in Visual Studio, so the command line is driven by our CMake Presets file, and I\u2019d rather make modifications there.<\/p>\n

Let\u2019s ask Copilot again with those requests in mind.<\/p>\n

\u00a0<\/em>\ud83d\udca1Hm, wouldn\u2019t it be nice if I didn\u2019t have to specify manifest mode?<\/em><\/p>\n

\ud83d\udca1It would also be nice if I didn\u2019t have to specify that I prefer using CMakePresets.json when I ask Copilot questions? <\/em><\/p>\n

\"Asking<\/a><\/p>\n

 <\/p>\n

These suggestions are much more aligned with what I\u2019d like to do in my project. Following a combination of the Copilot suggestions and this documentation<\/a>, I did the following:<\/p>\n

    \n
  1. Installed vcpkg and ran the bootstrap script.<\/li>\n
  2. Created the necessary vcpkg.json and vcpkg-configuration.json files defining opencv as a dependency.<\/li>\n
  3. Modified the CMAKE_TOOLCHAIN_FILE<\/code> using the CMakePresets file.<\/li>\n<\/ol>\n

    At this point, we were closer, but we still failed to configure. After asking Copilot for help, I realized that it wasn\u2019t finding the toolchain file, and that I should specify the VCPKG_ROOT<\/code> in an environment variable in the preset and use that variable in the CMAKE_TOOLCHAIN_FILE<\/code> setting to successfully find it.<\/p>\n

    At this point, CMake configured successfully and OpenCV had been included in my CMake project.<\/p>\n

    Use OpenCV to detect dogs<\/h3>\n

    Now that OpenCV was properly included in my CMake project, the next step was to figure out how to use the library and how to get it to detect dogs. From my little background knowledge of OpenCV, I knew that it would need a model trained to detect certain objects. To start, I asked Copilot if OpenCV can detect dogs already, and it told me that there were already publicly available models and, upon further questioning, pointed me to some links where I could download configuration files, weights files, and class names from this repository<\/a>.<\/p>\n

    Now that I had downloaded models I could use, I asked Copilot to give me a code snippet that would allow me to detect dogs.<\/p>\n

    One thing that I hit errors on and had to modify was that Copilot suggested relative paths, and this didn\u2019t work, since it was running the executable from the build location. I found it easier to always use absolute paths to make this simpler.<\/p>\n

    \"Asking<\/a><\/p>\n

    \ud83d\udca1It would be great if I could tell Copilot to always use absolute paths.<\/em><\/p>\n

    After modifying the paths, I was able to detect dogs in images!<\/p>\n

    \"Using<\/a><\/p>\n

    Switch to livestream<\/h3>\n

    I didn\u2019t want to only be able to detect dogs in images, I also wanted to be able to detect dogs using a livestream. I went back to Copilot and asked it to help me make these changes: \u201cHow can I modify this code to livestream from a camera rather than reading a single image?\u201d.<\/p>\n

    Copilot was able to suggest slight modifications to the code, and I was able to integrate a while loop and slightly modify the camera capture mechanism to accomplish this. See the original code versus the suggested livestream code below.<\/p>\n

    \"Copilot<\/a><\/p>\n

    To my surprise, the model could detect other objects as well.<\/p>\n

    \"Image<\/a><\/p>\n

    Add UI<\/h3>\n

    To make a more well-rounded application, I wanted to add a user interface to make it easier to control what models are used and to make it easier to operate overall. Through some conversations with Copilot, I was able to add qt5 as a vcpkg.json dependency and add components that allow you to select the .cfg, the .weights, the class names file, and to start and stop the livestream.<\/p>\n

    The below partial screenshot shows the initial suggestions that copilot gave:<\/p>\n

    \"Image<\/a><\/p>\n

    After refining the UI with more conversation with UI and code improvements, the pictures below show The Dog Detector when the camera isn\u2019t on versus when it is.<\/p>\n

    \"Image<\/a><\/p>\n

    Also, the ultimate test was to test out the \u201cDog Detector\u201d with my dog.<\/p>\n

    \"GIF<\/a><\/p>\n

    Build the project statically<\/h3>\n

    Finally, I wanted to be able to share the executable of this project easily with others so that they could use this project without having to worry about the runtime, the build process, etc. To do this, I asked Copilot: \u201cHow to statically link a library with CMake and vcpkg?\u201d<\/p>\n

    \"Asking<\/a><\/p>\n

    Copilot gave a suggestion to set the vcpkg triplet, but it gave suggestions to modify it in CMakeLists.txt or in the command line. To get better suggestions, I asked again, adding that we want to modify the CMakePresets.json.<\/p>\n

    \ud83d\udca1This is another instance where it\u2019d be nice to not have to specify this every time I ask.<\/em><\/p>\n

    I took these suggestions and modified them to use CMakePresets.json, so I set the VCPKG_TRIPLET_TARGET<\/code> in CMakePresets.json.<\/p>\n

    With these changes, configuration succeeded, but I still got the following build errors:<\/p>\n

    mismatch detected for \u2018RuntimeLibrary\u2019: value \u2018MTd_StaticDebug\u2019 doesn\u2019t match value \u2018MDd_DynamicDebug\u2019 in DogDetector.cpp.obj.<\/pre>\n
    To help me fix this problem, I asked Copilot to help. Copilot told me to ensure I set the variable OpenCV_STATIC<\/code> to ON<\/code>. Since this variable is specific to only the target that includes OpenCV, I set this in the CMakeLists.txt. Additionally, CMake told me that I should set the CMAKE_MSVC_RUNTIME_LIBRARY<\/code> to MultiThreaded$<$<CONFIG:Debug>:Debug><\/code>. However, after making these changes, I was still getting the error.<\/p>\n
    After much conversation with Copilot, I realized it may have something to do with CMake Policies associated with the CMAKE_MSVC_RUNTIME_LIBRARY<\/code>. Looking at the documentation<\/a>, the variable only takes effect when the CMP0091 policy is set to \u201cnew\u201d, which is only \u201cnew\u201d by default in version 3.15 and beyond.<\/p>\n
    \ud83d\udca1It would have been great to be informed about CMake Policies that affect CMake variables that Copilot suggested I use.<\/em><\/p>\n
    After updating the CMake project to have a minimum required version of 3.15, the project built statically with no errors.<\/p>\n
    Custom Instructions<\/h3>\n
    After completing the project, looking back at all the questions and interactions I had with Copilot through Copilot Chat, there were many times where I thought, \u201cHm, it would be great if I didn\u2019t have to continually specify this information in each question.\u201d. You may have noticed this internal dialog above.<\/p>\n
    To solve this, I used the Copilot Custom Instructions feature.<\/a> This feature allows you to create a .github\/copilot-instructions.md file that contains additional context you want to be added to every question that you ask Copilot Chat.<\/p>\n
    Based on the various observations I made, I created a copilot-instructions.md file and added the following custom instructions:<\/p>\n