fix(remote): deduplicate artifact names in remote-build success message

[kirtisingh05]

• Have you followed the guidelines for contributing?
• Have you signed the CLA?
• Have you successfully run make lint && make test?
• Have you added an entry to the changelog (docs/reference/changelog.rst)?

---

Summary

When remote-build is run multiple times, Launchpad returns historical
builds with different URLs but the same output filenames. Since
artifact_downloads keys on URL, duplicates slip through into .values(),
causing the success message to list the same snap multiple times.

Fix

Deduplicate artifact names using a set comprehension before sorting:

artifact_names = sorted({path.name for path in artifacts})

Test

Added a regression test that mocks fetch_artifacts to return duplicate
filenames and asserts each artifact appears only once in the success message.

Fixes #1093
Related: canonical/snapcraft#6230

[steinbro]

This change would fix the visual quirk of duplicate filenames, but I'm concerned that it's masking a root cause. Why would historical builds be included in the data that's being fetched? Is the client downloading a bunch of paths to the same target unnecessarily? If so, are we sure that the resulting file left on disk is the most recent build?

Maybe @mr-cal knows if we need to adjust our Launchpad API calls to only fetch the artifact that resulted from this specific remote-build invocation..

[kirtisingh05]

Thanks for the review. you're right, this is masking the root cause rather than fixing it. The real issue is that get_builds() returns all historical builds for the recipe, not just the current invocation. I'm working on filtering by build request ID for fresh runs, with a fallback to newest-per-arch for resumed builds. Will update the PR once I've traced the right fix with @mr-cal's input on the Launchpad API.

[mr-cal]

Hi @kirtisingh05,

Thanks for the PR. I agree with @steinbro that filtering at the output is indeed masking a deeper problem and not the direction we should go. I've put some of my thoughts into the original issue into how this could be solved.