Failures during add_hookspecs / historic replay leave partial state
Maintainers usually reply within 1 day
Assessment
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Newbie friendliness
- 50/100
Research direction
Start with add_hookspecs() and the _verify_hook call that can raise for an already-registered impl, then see where set_specification() installs the spec before that check runs. Next, look at register() and the comment near _maybe_apply_history, which admits a failed replay leaves the plugin half-registered. Done means a failed add_hookspecs leaves no spec behind so a retry succeeds, and a failed replay in register() rolls the registry back. The issue does not say how rollback should be structured, so confirm the approach with maintainers before writing it.
Written by the indexing model from the issue text.
Description
🤖 Written by Claude Opus 5.5 via Claude Code for the pluggy maintainers; I prompted it, it did the work, I read it.
add_hookspecs(): if_verify_hookraises for an already-registered impl, the spec stays installed viaset_specification(). A retry then raisesValueError: Hook 'h' is already registered within namespace ..., even after unregistering the bad plugin.register(): if a historic replay (_maybe_apply_history) raises, the plugin stays in the registry with only the hookimpls processed before the failure. The code comment there acknowledges this.
Proposal: validate first and install afterwards in add_hookspecs. For the replay case, roll back on exception.
Found while working on #759.
- Dominant language
- Python
- Stars
- 1.7k
- Forks
- 172
- Avg merge
- 2d 18h
- Merged PRs (30d)
- 21
Getting set up
This project ships no dev container, Dockerfile or contributing guide, so setting up is up to you: start from its README, and see our first-contribution guide for the general steps.
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from pytest-dev/pluggy
-
Plugin registry compares plugins by equality, not identityPossibly taken @00200200 claimed this 1 day ago. Open
Difficulty 2/5 1-3 hours Newbie friendliness 62/100
pytest-dev/pluggy#760 ·
Maintainers usually reply within 1 day
-
Defaulted positional-only hook parameters are forwarded as keywordsPossibly taken @adenzhou1350 claimed this 1 day ago. Open
Difficulty 4/5 1-2 days Newbie friendliness 45/100
pytest-dev/pluggy#774 · 1 comment ·
Maintainers usually reply within 1 day
-
list_plugin_distinfo() keeps unregistered pluginsPossibly taken @RonnyPfannschmidt claimed this 3 days ago. Open
Difficulty 2/5 1-3 hours Newbie friendliness 35/100
pytest-dev/pluggy#763 ·
Maintainers usually reply within 1 day
-
unregister() argument validation relies on assertPossibly taken @ekanshul claimed this 3 days ago. Open
Difficulty 2/5 1-3 hours Newbie friendliness 42/100
pytest-dev/pluggy#762 ·
Maintainers usually reply within 1 day
-
[Downstream] devpi broken - trying to register `None`Possibly taken @bluetech claimed this 12 days ago. Open
Difficulty 3/5 1-2 days Newbie friendliness 58/100
pytest-dev/pluggy#749 · 4 comments ·
Maintainers usually reply within 1 day
All issues in pytest-dev/pluggy
Similar issues
-
Difficulty 1/5 1-3 hours Newbie friendliness 87/100
adri0/scadustats#128 ·
Maintainers usually reply within 1 day
-
`TIMESTAMP`, `TIMESTAMP_NTZ`, `TINYINT` and the other `TypeDecorator`s don't declare `python_type`Open
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
-
RPBCOpenawaiting policy acceptance precheck-passed
Difficulty 1/5 Under an hour Newbie friendliness 78/100
Bioconductor/BiocContributions#240 · 1 comment ·
-
bug
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
Maintainers usually reply within 1 day
-
documentation good first issue launch
Difficulty 2/5 1-3 hours Newbie friendliness 83/100
FlologixAI/etchwork#4 ·