Mining: a PR that names its blocker in Unblocks: or the text, without a closing keyword, isn't counted as a fix #52

Input text

- [x] Unblocks: PPSA99011 #4808; PPSA99995 (same stop; titles not rerun)

(boykopovar/AnyPS5#5163, open)

Propose a host-provided certificate bundle for `/<sandbox word>/common/cert/CA_LIST.cer`, addressing #5001.

(boykopovar/AnyPS5#5091, open)

- [x] Unblocks: PPSA97510, PPSA98510, PPSA99003 #1578 (title evidence in #4478 and the linked retest above)

(boykopovar/AnyPS5#4525, merged 2026-10-13)

Expected

In the first two, the PR names an open blocker issue in its Unblocks: line or with "addressing", so the blocker should count as Fix in review (#4808 → #5163, #5001 → #5091). The third is a merged PR whose evidence issue it fixes (#4478 → #4525). A link at least as strong as mentions there would show the blocker as fixed, or flag it as "fixed by a merged PR but still open" for review.

Actual

Only PRs with a closing keyword get link: "fixes", so all three blockers show under No open fix (#/blockers?status=nofix, data of 2026-10-13T12:33Z). Of the 88 blockers in that view, 46 have open PRs that reference them without a closing keyword. Others where that reference is the fix (open PR in brackets):

• #3036 (#4820)
• #4814 (#4822)
• #3553 (#4417)
• #3410 (#3619)
• #2746 (#3972)
• #4205 (#4827)
• #4285 (#4451)
• #4013 (#4593)
• #3880 (#4702)
• #4463 (#4467, #4469, #4514)

Some of the 46 are plain mentions, for example #1 macOS support cited by unrelated PRs. So a heuristic would need to tell the two apart: the blocker number in a PR's Unblocks: line, or the PR title naming the same function as the blocker.

Separately, a few blockers in the view are already fixed on main and only stay open because no closing keyword was used: #4478 (#4525), #3624 (#3640 and #3651), #3625 (#4226) and #3249 (localeconv is exported on main). I'm leaving notes on those issues so they can be closed upstream.

Where the text comes from

boykopovar/AnyPS5#5163