Michael Terry

This was closed - but isn't solved yet, correct? Like #281 hasn't landed, for example. Can this be re-opened, re-examined?

Good idea - plus that way they can be integrated as required tests for PRs. Got a PR to do this above.

I'm going to optimistically close this old ticket, on the theory that this was a python2 issue. But please re-open if this still happens.

That would likely solve my current use cases yes, as long as that sort of prefix-stripping option could be provided multiple times for different prefixes. But: - for the case...

Do you know if Athena complains if you also send it basic auth? Curious if we could just send both unconditionally - would need to test on other servers too

OK thank you - so we need to either somehow detect / handle this case or allow more customization of the auth flow.

I'm going to close this in favor of #151 which is a more targeted "just fix python2-isms like importing" - I cherry-picked your original commit here for proper credit. But...

OK using an rclone backend, this error doesn't happen. So it's only on local FUSE filesystems, I guess. Restic is thinking it should be a fully capable filesystem, but I'm...

I started looking at where we spend time in the code, but realized that even if I optimize it, it's pointless if we're still just using one core. So that's...

While investigating, I discovered that ever since we started using the MS de-id tool, we don't really need the internal validation provided by `fhirclient`. By skipping that de-serialization and re-serialization,...