Fix dep_updates for v1 recipes using python_min #5306

janjagusch · 2025-12-29T12:48:20Z

Summary

Fix the dep_updates migrator failing for v1 recipes that use python_min (CFEP-25).

Root cause: Grayskull generates skip: match(python, "<3.10") for noarch recipes, but v1 recipe variant configs only define python_min, not python. When rattler-build tries to render the recipe, all variants are skipped because python is undefined.

Fix: Replace match(python, with match(python_min, in grayskull-generated v1 recipes before rendering.

Changes

Add workaround in _make_grayskull_recipe_v1 to replace match(python, with match(python_min,
Add integration test case for dominodatalab feedstock
Add unit test test_get_grayskull_comparison_v1_python_min_mismatch
Add assert_pr_body_not_contains helper method for integration tests

Test Evidence

Successful version update PR created by the bot during integration testing:
janjagusch-conda-forge-bot-staging/dominodatalab-feedstock#28

This test case reproduces the issue where the dep_updates migrator fails for v1 recipes when grayskull generates a recipe with `skip: match(python, ...)` but the feedstock's variant config only has `python_min` (not `python`). The test adds: - dominodatalab feedstock as a git submodule (pinned to v1.4.7) - Test case that validates version update from 1.4.7 to 2.0.0 - New `assert_pr_body_not_contains` helper method to check PR body - Assertion that the PR body doesn't contain the dep_updates error message The test currently fails, demonstrating the bug where rattler-build skips all variants when the `python` variable is not in the variant config. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

This test reproduces the issue where get_grayskull_comparison fails for v1 recipes when grayskull generates `skip: match(python, "<3.10")` but the variant config only has `python_min` (not `python`). The test is marked as xfail since it currently fails - rattler-build skips all variants when the `python` variable is not set in the variant config. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

The xfail marker was causing pytest to consider the failing test as "passing" (expected failure). Removing it ensures the test actually fails in CI until the bug is fixed. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Replace `match(python, ...)` with `match(python_min, ...)` in grayskull-generated v1 recipes. This fixes the dep_updates migrator for feedstocks using CFEP-25's python_min variable. Grayskull generates skip conditions like `skip: match(python, "<3.10")` but v1 recipe variant configs only define `python_min`, not `python`. This caused rattler-build to skip all variants with: "Failed to render recipe YAML! No output recipes found!" See: conda/grayskull#574 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

- Add mock endpoint for version-specific PyPI API (/pypi/dominodatalab/2.0.0/json) - Include 'urls' field in main PyPI endpoint for grayskull sdist discovery - Add spdx.org/* to transparent URLs (grayskull license discovery) - Store full PyPI metadata response for reproducible testing 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Add documentation to help Claude Code understand the codebase structure, common commands, and integration test setup. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

beckermr · 2025-12-29T16:30:15Z

@wolfv rattler build should define python, right? Is this an issue somewhere in rattler-build or smithy maybe?

…egration-test # Conflicts: # .gitmodules

CLAUDE.md

beckermr · 2026-01-19T17:55:40Z

conda_forge_tick/update_deps.py

+    # Grayskull generates `match(python, ...)` for noarch recipes, but v1 recipes
+    # use `python_min` in their variant configs (CFEP-25). Replace to make the
+    # recipe renderable. See: https://github.com/conda/grayskull/issues/574
+    recipe_str = recipe_str.replace("match(python,", "match(python_min,")


This filter is rather broad and I am concerned it might sweep up other parts of the recipe with this syntax. I suggested we write a more targeted edit that ensures the change is in a noarch: python recipe.

Guard the match(python, ...) to match(python_min, ...) replacement with a check for package_is_noarch to avoid unintended changes to arch-specific recipes. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

- Rename CLAUDE.md to AGENTS.md and create symlink for backwards compat - Make content agent-agnostic (remove Claude Code-specific references) - Remove "(recommended for development)" from environment.yml option - Use local conda-lock.yml instead of wget from GitHub - Remove unmaintained test feedstocks list Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

janjagusch · 2026-01-31T12:16:56Z

@beckermr, @ytausch, took me a bit to get back to it, but it should be ready for another round of reviews now. :)

codecov · 2026-01-31T12:34:53Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 80.60%. Comparing base (d529852) to head (17fe1d8).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5306      +/-   ##
==========================================
- Coverage   80.62%   80.60%   -0.02%     
==========================================
  Files         144      144              
  Lines       17005    17013       +8     
==========================================
+ Hits        13710    13714       +4     
- Misses       3295     3299       +4

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

beckermr · 2026-01-31T12:44:44Z

conda_forge_tick/update_deps.py

+
+    # Grayskull generates `match(python, ...)` for noarch recipes, but v1 recipes
+    # use `python_min` in their variant configs (CFEP-25). Replace to make the
+    # recipe renderable. See: https://github.com/conda/grayskull/issues/574


This issue is now marked as resolved. Let's test this PR without the change below to see if it works now.

Wow, that's surprising! It had been open for more than a year.

@beckermr, we have to bump grayskull to v3.1.0:

https://github.com/conda/grayskull/releases/tag/v3.1.0

grayskull is at 3.1.0 in the lockfile, so something is still broken or something else broke.

conda_forge_tick/update_deps.py

beckermr · 2026-01-31T13:00:41Z

This error looks odd?

tests/test_update_deps.py::test_get_grayskull_comparison_v1_python_min_mismatch - RuntimeError: Failed to render recipe YAML! No output recipes found!
[]

beckermr · 2026-01-31T13:01:29Z

/relock-conda

conda_forge_tick/update_deps.py

Co-authored-by: Matthew R. Becker <beckermr@users.noreply.github.com>

beckermr · 2026-01-31T13:52:03Z

@janjagusch We need to diagnose the error here, update the comment in the code, and report it upstream to grayskull before we can merge this.

Do you know what is still broken in the generated recipe even in 3.1.0?

janjagusch · 2026-01-31T15:01:49Z

@janjagusch We need to diagnose the error here, update the comment in the code, and report it upstream to grayskull before we can merge this.

Do you know what is still broken in the generated recipe even in 3.1.0?

Sure, I can report that the Grayskull patch seemingly doesn't work for us. It might take me a few days before I can get back to it, though. :)

janjagusch and others added 4 commits December 29, 2025 13:37

janjagusch changed the title ~~Add failing tests for dominodatalab dep_updates issue~~ Fix dep_updates for v1 recipes using python_min Dec 29, 2025

janjagusch marked this pull request as draft December 29, 2025 14:55

janjagusch force-pushed the add-dominodatalab-integration-test branch from 12a2004 to 09ca7cc Compare December 29, 2025 16:21

docs: add CLAUDE.md for Claude Code guidance

5d9b386

Add documentation to help Claude Code understand the codebase structure, common commands, and integration test setup. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

janjagusch marked this pull request as ready for review December 29, 2025 16:23

ytausch self-requested a review January 19, 2026 17:19

Merge remote-tracking branch 'origin/main' into add-dominodatalab-int…

f672537

…egration-test # Conflicts: # .gitmodules

ytausch approved these changes Jan 19, 2026

View reviewed changes

CLAUDE.md Show resolved Hide resolved

CLAUDE.md Outdated Show resolved Hide resolved

CLAUDE.md Outdated Show resolved Hide resolved

CLAUDE.md Outdated Show resolved Hide resolved

beckermr requested changes Jan 19, 2026

View reviewed changes

janjagusch and others added 4 commits January 31, 2026 13:10

Only apply python_min replacement for noarch recipes

2883823

Guard the match(python, ...) to match(python_min, ...) replacement with a check for package_is_noarch to avoid unintended changes to arch-specific recipes. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Always use lockfile

c746b18

Merge branch 'main' into add-dominodatalab-integration-test

5a19be5

janjagusch requested review from beckermr and ytausch January 31, 2026 12:16

beckermr reviewed Jan 31, 2026

View reviewed changes

conda_forge_tick/update_deps.py Show resolved Hide resolved

Apply suggestion from @beckermr

74c24a4

Update grayskull version to 3.1.0

40dbc97

beckermr reviewed Jan 31, 2026

View reviewed changes

conda_forge_tick/update_deps.py Outdated Show resolved Hide resolved

Update conda_forge_tick/update_deps.py

17fe1d8

Co-authored-by: Matthew R. Becker <beckermr@users.noreply.github.com>

Fix dep_updates for v1 recipes using python_min #5306

Are you sure you want to change the base?

Fix dep_updates for v1 recipes using python_min #5306

Uh oh!

Conversation

janjagusch commented Dec 29, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Changes

Test Evidence

Related

Uh oh!

beckermr commented Dec 29, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

beckermr Jan 19, 2026

Choose a reason for hiding this comment

Uh oh!

janjagusch commented Jan 31, 2026

Uh oh!

codecov bot commented Jan 31, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

beckermr Jan 31, 2026

Choose a reason for hiding this comment

Uh oh!

janjagusch Jan 31, 2026

Choose a reason for hiding this comment

Uh oh!

janjagusch Jan 31, 2026

Choose a reason for hiding this comment

Uh oh!

beckermr Jan 31, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

beckermr commented Jan 31, 2026

Uh oh!

beckermr commented Jan 31, 2026

Uh oh!

Uh oh!

beckermr commented Jan 31, 2026

Uh oh!

janjagusch commented Jan 31, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

janjagusch commented Dec 29, 2025 •

edited

Loading

codecov bot commented Jan 31, 2026 •

edited

Loading