-
Notifications
You must be signed in to change notification settings - Fork 1
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add robotology software ports to this repository #15
Comments
I think that make totally sense for all the software that depends on other ports that are only stored here, so for example libraries that depend on IPOPT. For libraries that instead can be directly submitted to vcpkg main repo, I would not rule out the possibility of submitting them directly to the vcpkg main repo, as with that we automatically get by free CI and maintenance for every possible change of vcpkg itself. For example, whenever a new version of the MSVC compiler is released, it is tested on all ports of the vcpkg main repo (see microsoft/vcpkg#13748), so having our libraries directly in vcpkg would help catching any kind of early compiler-related regression. If instead for any reason the vcpkg ports get stale, we can always add in this repo our version of those ports, so that we can overlay the vcpkg main repo version. In the end, I think that any library mantainer that went to submit the vcpkg port for its library in this repo is welcome, but on the specific case it could also make sense to submit it directly to vcpkg main repo. |
For reference, there is a working version for YCM here: https://mp-force.ziti.uni-heidelberg.de/asc/infrastructure/vcpkg-ports/-/tree/master/vcpkg/ports/ycm |
For YARP that needs to run its own code generators as part of the build process, in microsoft/vcpkg#14464 I added the possibility for vcpkg ports to set |
As vcpkg installs the ports in a directory that is different from the final one, I think solving this two issues is necessary to easily support writing a vcpkg YARP port: |
For the sake of easily target a arm64-uwp target (HoloLens2) with a YARP using application, I implemented two simple ports for Using the version from microsoft/vcpkg@386aaab, ycm works fine, while yarp fails for the following reason:
The reasons are the following:
|
Unfortunately, the carriers are needed for sure, and that can't be disabled. So we need to find a better solution. |
Whatever vcpkg is doing seems very wrong... Perhaps have a look at how the components are handled in Qt5
What is Qt5 doing with the plugins? I think that the issue is very similar, and I don't think they are just installed in lib...
This looks like a bug in yarp to me... (fixed in robotology/yarp#2598) |
That ship has sailed a long time ago (related discussion: microsoft/vcpkg#77 ).
Qt5 is handled different as any component can be built on its own, and this is what is happening (see https://github.com/microsoft/vcpkg/blob/1bb0fe61dfecd1450a1ff44c7fc8d424365366aa/ports/qt5-base/cmake/qt_build_submodule.cmake#L21). However, this is not a big problem as I remember I had already implemented a piece of code to manage this case while packaging ignition robotics libraries: https://github.com/microsoft/vcpkg/blob/e81a56708035a7612127d6152d16bf863b29df78/ports/ignition-modularscripts/ignition_modular_library.cmake#L14 so I just adapted that.
Thanks!
At least in the vcpkg version used in https://github.com/robotology/robotology-superbuild-dependencies-vcpkg/releases/tag/v0.6.5, the plugins that for example in debian are installed in |
I am trying again to compile yarp for
This is not bad, it means we need to pass those values from the configuration. As I have no idea about them, I inspected the
|
I fixed the configuration issue, and now there are two more issues:
|
I added yarp and icub-main in this repo (#24), while idyntree in vcpkg itself (microsoft/vcpkg#30810). More ports can be added as necessary, so I think we can close the issue. |
I was wondering if it would make sense to add robotology software (YCM, RobotTestingFramework, YARP, etc) ports to this repository. What do you think?
The text was updated successfully, but these errors were encountered: