[JarryShaw]

Please follow the guide below

• You will be asked some questions, please read them carefully and answer honestly

• Put an x into all the boxes [ ] relevant to your pull request (like that [x])

• Use Preview tab to see how your pull request will actually look like

• Searched for similar pull requests

• Followed the coding style (make pylint, make mypy, make isort)

• make test passes, and a test case covers the change. Narrow runs only, listed below.

• Added a changelog entry under docs/source/changelog/ and regenerated CHANGELOG.md, if the change is user-visible: N/A — centralised in docs(changelog): shared 1.5.0 changelog — long-lived, merges last (#610, #616, #617, #618, #620) #657

What is the purpose of your pull request?

Tick the commit type your subject line carries.

• fix — corrects a defect
• feat — adds a feature
• perf — changes performance, not behaviour
• refactor — changes neither behaviour nor performance
• test — tests only
• docs — documentation only
• ci — workflows or build tooling
• release — bumps the version or rolls up a distribution
• chore — anything else

---

• util/conda_vendor.py: parse() kept only the last line per package and
  marker, so a conflicting duplicate pin passed --check while pip download
  -r rejected the file. It now raises Conflicting naming each such pair,
  and compare() reports each as a conflict: problem, so --check exits 1.
• An identical duplicate line is accepted, since pip resolves it.
• tests/project/test_conda_vendor_1814.py: a conflicting aenum pin appended
  to the committed file is reported through parse() and compare(), the
  committed file and an identical duplicate are not, markers are compared
  once normalised, and --check exits 1 on a conflict in its own output.

Closes #1852

[JarryShaw]

Verdict: GOOD TO GO at 72a388a38. A Sonnet cross-review (author: Opus) confirmed:

• --check exits 0 on the committed file and 1 on a conflicting duplicate, with a conflict: line. main exits 0 on the same input, which is the bug.
• An identical duplicate is accepted, and markers are compared after normalisation.
• The writer path (cron-conda.yml:177) never calls parse, so it cannot raise. Running it on a scratch copy gives output that --check passes.
• --check runs with sockets patched to fail, so it stays offline.
• The new tests fail on main. Mutating the fix (no raise, no normalisation, no marker in the key, compare dropping conflicts) is caught.

Non-blocking nits:

• No test pins that the first line's pin is kept.
• The parse docstring calls conflicts pins that "cannot both hold", then says it flags them whether or not they could.
• Specifiers are compared as strings, so ==1.0 against ==1.0.0 is flagged.