Skip to content

E1 cleanup: delete the Python HuggingFace commit path #75

Description

@tamnd

Spec: notes/Spec/2115/07-enhancements.md, item E1. Follow up to #40.

Problem

v0.6.0 made the Go commit path the default and left the Python one reachable through CCRAWL_HF_COMMIT=python, on purpose, so anyone who hit a problem in the first release had a way back. That was the deal in #40: keep it for one release, then delete it. The escape hatch is dead weight now. It keeps ccrawl/embed/hf_commit.py in the binary, it keeps hfResolveUV looking for an interpreter that nothing needs, and it keeps a second code path that no test run in CI exercises end to end.

What to do

  • Delete ccrawl/embed/hf_commit.py and the embed directive that pulls it in.
  • Delete hfScriptPath, hfResolveUV and createCommitPython from ccrawl/hf.go.
  • Drop the CCRAWL_HF_COMMIT branch in CreateCommit and the row for it in docs/content/reference/configuration.md.
  • Take out the script extraction into ~/.cache/ccrawl/, which is the last thing writing executable content there.

Done when

  • grep -ri python ccrawl/ cli/ finds nothing that runs an interpreter
  • CCRAWL_HF_COMMIT is gone from the code and the docs
  • the release notes for v0.7.0 say the escape hatch is gone

Metadata

Metadata

Assignees

No one assigned

    Labels

    area: hfHuggingFace client and uploadpriority: P0Integrity, the binary must keep the claims it makestype: featureNew capability

    Projects

    No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions