Closed corneliusroemer closed 4 months ago
Hi @corneliusroemer,
Thanks for opening this issue. It looks like we introduced this bug today and we will start working on a fix soon.
You have two options to get this working:
--force
with dataformat to ignore the errorBest, Eric
Eric Cox, PhD [Contractor] (he/him/his) NCBI Datasets NIH/NLM/NCBI eric.cox@nih.gov
Thanks @ericcox1 for your quick reply! I'll try upgrading to 16.13.0 and see if that fixes it.
Are you sure --force
fixes it? The output says "Use --force to remove this warning." but then below there's an "Error" which is usually considered worse than a warning: "Error: proto: (line 10:7): unknown field "uncompressedMd5Hex""
I'm wondering whether you're running tests in CI to catch such errors before releasing? If you have a staging environment, you could ensure old versions (of the same major version) still function against the new server code (assuming this is what happens: you changed something on the server and it breaks old clients).
This issue appeared across multiple independent pipelines (Nextstrain and non-Nextstrain) that I'm monitoring, so it's likely a high-profile bug that's breaking things for a lot of users.
I can confirm that upgrading to 16.13.0 works. I haven't tested adding --force
.
Hi @corneliusroemer,
We rolled back the change that caused this bug, so previous versions of the client should work normally again.
Thanks again for your report.
Best, Eric
Before opening an issue, please:
datasets --version
: not using the latest version, but I'm on a 16. something version so this should work nonetheless. One can't always update to the latest version immediately.Describe the bug
Dataformat suddenly throws a cryptic error:
To Reproduce
Expected behavior No error. Things worked fine just a few hours ago.