-
Notifications
You must be signed in to change notification settings - Fork 48
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
Remove pinning of swig to 4.1.0 to workaround swig regressions and add "libopencv>=4.10.0" constraint #1607
Conversation
Cool, removing the pinning still installs swig 4.2.0 instead of swig 4.2.1, I guess there is a pcre2 migration going on:
|
Exactly, the problem is conda-forge/swig-feedstock#49 that was done on swig side and conda-forge/glib-feedstock#159 that still neesd to be done on the glib side. |
This was eventually done in conda-forge/glib-feedstock#164 . |
Ok, now YARP compiles fine, the failure is caused by casadi 3.6.3 instead:
|
Cool, that is fixed by casadi 3.6.5, see https://github.com/casadi/casadi/blob/3.6.5/swig/casadi.i#L84 . So this PR is blocked by #1614 . |
f00ae59
to
5bbbcf4
Compare
#1614 was merged, so this is ready for new CI. |
CI still installs swig 4.2.0 as:
fails with:
|
This turned out to be a bit complicated, the core issue is:
In a nutshell, a recent libgdal build (required by a recent gazebo) requires expat 2.6, but this is not compatible with vtk (see conda-forge/vtk-feedstock#321, conda-forge/conda-forge-repodata-patches-feedstock#682 and conda-forge/conda-forge-pinning-feedstock#5640). I guess the best solution now is to wait until vtk is again compatible with expat 2.6 . |
Something changed with conda-forge/vtk-feedstock#328, once pcl migrates to vtk 9.3.0 we should be ready to resume this PR. |
We have the new pcl since conda-forge/pcl-feedstock#65, let's see if we can finally close this. |
I also added a constraint on |
There is now a strange installation conflict that is due to conda-forge/zziplib-feedstock#11, hopefully once that PR is merged it will be solved. |
Cool, now |
On Windows there is a casadi bindings failure, that could be related to a swig update:
|
Perhaps it is related to casadi/casadi#3723 and casadi/casadi#3712 . I wonder why this does not happen with older swig. I guess it should be solved by casadi/casadi#3724 . |
Cool, new zlib excessive constraint problem:
|
This was reported in conda-forge/libtheora-feedstock#21 and should be solved by conda-forge/libtheora-feedstock#22 . |
Windows is failing with:
not sure what changed recently. macOS osx-arm is still failing with the |
Ah, this was discussed in #1607 (comment) . |
We need to unblock this to fix #1664 . |
I guess we can backport casadi/casadi#3724 in our version of casadi and just set QT_HOST_PATH to workaround conda-forge/qt-main-feedstock#273 . |
Related issues: