poetry stores repo credentials by host, it should include path

if you have multiple pypi repositories off the same domain but with different paths, poetry will fail to authenticate properly because it will store the credentials for the repository based on the domain, but not inclusive of path, meaning it will use the wrong credentials for a given domain.

tldr poetry credential cache keys are not correct per python spec on pypi, as poetry doesn't address repositories by suffix.
https://www.python.org/dev/peps/pep-0503/

separately there's an issue is that the url doesn't get normalized early, so subsequent lookups in different parts of the codebase but same cli exec will do lookups with and without + 'simple' on the url.

• [ x ] I am on the latest Poetry version.
• [ x ] I have searched the issues of this repo and believe that this is not a duplicate.
• [ x ] If an exception occurs when executing a command, I executed it again in debug mode (-vvv option).

$ poetry config --list | grep repo

repositories.assetdb.url = "https://corp-1122334455.d.codeartifact.us-east-1.amazonaws.com/pypi/assetdb/"
repositories.c7n-next.url = "https://corp-1122334455.d.codeartifact.us-east-1.amazonaws.com/pypi/c7n/"

traceback

  • Installing c7n (0.1.2): Pending...
  • Installing c7n (0.1.2): Failed

  RepositoryError                                                                                         
                                                     
  401 Client Error: Unauthorized for url: https://corp-1122334455.d.codeartifact.us-east-1.amazonaws.com/pypi/c7n/simple/c7n/
                                                     
  at ~/.local/share/pypoetry/venv/lib/python3.9/site-packages/poetry/repositories/legacy_repository.py:393 in _get
      389│             if response.status_code == 404:
      390│                 return                                                                         
      391│             response.raise_for_status()                                                                                                                                                                  
      392│         except requests.HTTPError as e:                                                        
    → 393│             raise RepositoryError(e)                                                           
      394│                                           
      395│         if response.status_code in (401, 403):
      396│             self._log(
      397│                 "Authorization error accessing {url}".format(url=response.url),

Versions

• Linux/Ubuntu 20.04
• Poetry 1.12

The underlying issue appears to be how credentials are retrieved, it simply grabs the credentials for the first repository that has a domain match
https://github.com/python-poetry/poetry/blob/master/src/poetry/utils/authenticator.py#L151

Start in src/poetry/utils/authenticator.py around the credential lookup at line 151, then inspect the request flow in legacy_repository.py around _get. Compare repository URL handling with PEP 503, including the differing /simple/ forms. Done means repositories sharing a host use credentials keyed by their paths and lookups behave consistently within one CLI run.