[shbernal]

Closes #1520

Summary

The comment on RESULTS_PER_LINKEDIN_PAGE described only the classic search page, which rendered 10 or 11 cards per navigation. It now also says that the redesigned /jobs/search-results/ page renders 25, the same as the stride, and that the match is a coincidence: the offset still follows what each page rendered. Comment only, no behaviour change.

Verification

• ruff, ty, tests/linkedin/test_job_policy.py, the policy-trace and architecture-doc tests pass.
• Full suite in fork CI on this head (d5a81d71): every job passed, the test job included (run).

Synthetic prompt

Update the comment above RESULTS_PER_LINKEDIN_PAGE in job_policy.py: the classic /jobs/search/ page rendered 10 or 11 cards per navigation, the redesigned /jobs/search-results/ page renders 25, so the match with the stride is a coincidence and only the past-the-last-page check may use the constant.

Implemented with Claude Opus 5.5 (high) in Claude Code via tmux.

Note

Docs: clarify card counts per page in RESULTS_PER_LINKEDIN_PAGE comment

Rewrites the comment in job_policy.py to distinguish the two LinkedIn job search pages. The classic /jobs/search/ page shows 10 or 11 jobs despite advertising 25 per page, while the redesigned /jobs/search-results/ page shows 25. The note that the stride only checks for past-last-page search is kept. Comment-only change.

^{Macroscope summarized d5a81d7.}

[greptile-apps]

Confidence Score: 5/5

[Low impact] This comment-only change appears safe to merge.

Summary

Updates the comment above RESULTS_PER_LINKEDIN_PAGE.

• The job-search note explains why 25 cards can match the offset by chance.

_{Reviews (1) · Last reviewed commit: "docs(jobs): Give cards per page on each ..." · Reviewed by Greptile}