Sebastian Pietras
Sebastian Pietras
> Would you wanna take a stab at this and submit a PR? Sure, I'll give it a try soon. > Looks like it's actually possible to use the downloaded...
Update: It would be so easy to just use ```JAVABASE```... if only we had access to it. But there is a problem. ```JAVABASE``` is a [make variable from Java toolchain](https://docs.bazel.build/versions/master/be/make-variables.html#custom_variables)....
> If you remove the float specifier and have only TensorType['h', 'w'] does that still produce an error? I checked it and yes, there is still an error. Doing `torch.randn(5,...
See also: https://github.com/metallb/metallb/issues/1681
Hey, thanks for contributing. Can you provide some simple example of when this would be needed? Just so that I can understand the context better. I guess in your case...
Thanks, I will get back to it in a couple of days.
Not really, but I guess disabling ```conda clean``` (with ```clean = False``` in ```conda_create```) is a nice workaround. That way: - conda caches all packages (the downloaded tarballs) - if...
@sfc-gh-zpeng Can you split it into two pull requests? One with "-n base" fix, and another with version changes.
Hi, thanks for the feedback. I'm afraid the ```--prune``` option is of no use for us. Bazel is very strict on correctness and reproducibility. Rule outputs can only depend on...