Mike Sarahan
Mike Sarahan
Happy to reopen this, but we have no intention of doing development on this any time soon. The shell scripts are devilishly complicated, and we don't have time to get...
We have discussed this at length internally. @mcg1969 is very hesitant to change anything about how features work. There's a lot of pain there, and his opinion is that it's...
I don't think #2181 is related. I think your issue is a regression in some new-ish behavior, which requires that conda-build make two passes at parsing setup.py data. The first...
PS: unless you are using numpy's distutils stuff, I strongly advise you to prevent numpy from being imported at build time. At best, it will take much longer for your...
Thanks for re-raising this issue. We'll try to fix it soon.
Please consider a PR. We don't use this functionality and are overly booked with internal work right now. This is not feature that we use, and we can't justify spending...
Almost not at all. That's a develop-time feature, and we are very much focused on building only tagged, released software, where the model is that you provide the information in...
I'm not familiar with nb_conda_kernels, but I'm happy to help implement any necessary changes on the conda side.
after conda installs something, it reactivates your environment. One of the things that Keras brings in is theano. Theano needs a compiler to operate. Unfortunately, we don't currently have a...
As a hack, you can edit the .bat file in `/etc/conda/activate.d` to change the reference to VS2015 to point instead to your VS2017 install.