Error cloning dataset with Datalad on Windows10

  • Please describe the problem.

While trying to clone a dataset hosted on GIN to a Windows laptop, my console shows:

λ datalad clone git@gin.g-node.org:/jadecci/CBPP.git cbpp_project
Clone attempt:   0%|          | 0.00/1.00 [00:00<?, ? Candidate locations/s]getsockname failed: Not a socket

The same error shows up if I clone the repository with git clone and then try to download the content with datalad:

λ datalad get -n -r fc-behaviour-prediction
Clone attempt:   0%|          | 0.00/2.00 [00:00<?, ? Candidate locations/s]getsockname failed: Not a socket
  • What steps will reproduce the problem?

Any clone or get command gives me the same warning. For example:

datalad clone git@github.com:jadecci/cbpp.git temp
  • What version of DataLad are you using (run datalad --version)? On what operating system (consider running datalad plugin wtf)?

I’m using datalad 0.13.3. datalad wtf shows that the operating system is mingw32 x86_64

  • Please provide any additional information below.

I get no error if I used the HTTP link instead of the SSH link so it would seems that it’s an ssh error. However, I would like the remote link to be an SSH link, because otherwise git annex would not be able to sync the content properly.

  • Have you had any luck using DataLad before? (Sometimes we get tired of reading bug reports all day and a lil’ positive end note does wonders)

I have been syncing my work data, scripts and results with DataLad datasets across my Macbook and Linux environment on HPC for quite some time. Everything has been working real good (until I suddenly came out with the idea of running stuff on Windows)!

Hey @jadecci, thanks for the report! Please feel free to post bug reports like this as a GitHub issue to github.com/datalad/datalad/issues/new - we’ll see them much faster like this. The error is known and there has been some work towards a fix. Could you retry with the most recent DataLad 0.13.5 release?
If you happen to be working frequently on Windows with DataLad, you would do us a huge favor if you would paste all errors and bugs into issues. Currently, there is much work being done to establish and improve functionality on Windows, and having real users report what fails in their usecases would be invaluable! Thanks!