You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
kbcstorage has a hard dependency constraint urllib3<2.0.0 in its metadata, which prevents downstream projects from upgrading to urllib3 2.x.
The constraint was added as a workaround for boto/botocore#2926, which has since been resolved and closed.
Evidence
The comment in pyproject.toml reads:
"urllib3<2.0.0", # Frozen until fixed: https://github.com/boto/botocore/issues/2926
The referenced botocore issue is closed.
Current botocore (1.42.x) handles urllib3 version constraints itself via markers:
urllib3<1.27 for Python <3.10
urllib3>=1.25.4,<3 for Python >=3.10
kbcstorage does not import or use urllib3 anywhere in its source code — it's a phantom dependency.
Suggested fix
Remove the direct urllib3<2.0.0 pin from pyproject.toml and let the transitive dependencies (boto3/botocore) manage their own urllib3 version constraints. This is the correct approach since botocore already specifies the appropriate urllib3 markers for each Python version.
Impact
This constraint blocks all downstream projects from resolving to modern versions of urllib3, cryptography, requests, and other packages that have moved to urllib3 2.x, causing numerous unresolvable Dependabot security alerts.
I’d like to voice my support for merging this change.
Removing the phantom dependency version constraints would allow users to work with current versions of many widely used Python libraries without having to pin older releases or rely on workarounds. This would make the client much easier to integrate into existing projects, reduce dependency conflicts, and improve long-term maintainability.
This change would provide a significant benefit with zero downside.
Activity
Remove outdated urllib3<2.0.0 pin
stanekjiri commentedon Aug 27, 2026
I’d like to voice my support for merging this change.
Removing the phantom dependency version constraints would allow users to work with current versions of many widely used Python libraries without having to pin older releases or rely on workarounds. This would make the client much easier to integrate into existing projects, reduce dependency conflicts, and improve long-term maintainability.
This change would provide a significant benefit with zero downside.
Thanks,
Jiri Stanek, Seznam.cz