Closed jwodder closed 1 year ago
Patch and project coverage have no change.
Comparison is base (
b310e3e
) 97.71% compared to head (1542d74
) 97.71%.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.
Sad part of the story is that then pip
would likely just to install some prior release of dandi-schema
if some newer "bad" pydantic would already be installed, so overall this specification is good to have but is not really even a workaround. Looking forward, I guess we better indeed strive to keep such specification in the future to prevent "known to break something next major release" from being installed. Meanwhile will also add release label and merge.
@yarikoptic
Sad part of the story is that then pip would likely just to install some prior release of dandi-schema if some newer "bad" pydantic would already be installed
A (somewhat drastic) way to prevent this from happening would be to yank (not delete!) all older dandischema releases from PyPI, which would prevent pip 22.0+ from installing those versions unless the user specifically requested the older version with ==
. If someone then tried installing this latest dandi-schema in an environment in which pydantic 2.0 had already been installed, I believe pip would fail without installing anything.
We can update the code to use pydantic 2.0 after it comes out.
See #176.