StrictMark list entries soft-wrap: a following line indented to the SAME depth continues the entry's text (StrictMark; the wiki line budget — bullet 2 lines — makes such continuations normative across /wiki and /todo). mark instead flushes the continuation as a nested <p> inside the still-open <li>, so one logical entry renders as bare inline text PLUS a paragraph block (wrong structure, wrong spacing). Reported by gritzko 2026-07-07; reproduced with ~/bin/mark same day. Scope: beagle/mark (C + ragel). Method Issues.
- first entry line + next line continuation indented to the same depth renders <li>first entry line<p>\ncontinuation…\n</p>\n</li> — entry text bare, continuation wrapped in <p>.MARK.c:38 buffers soft-wrapped PARAGRAPH text (joined into one line at flush), but a list entry's own text is emitted bare while its <li> is held open for nesting (MARK.c:396-405 frame model, mark_para_flush MARK.c:412); the continuation line then enters the paragraph buffer and flushes as <p> INSIDE the li.<p> (the deferred-</li> design) — only the SAME-depth continuation is misclassified.<li>first entry line continuation…</li>.beagle/mark (MARK.c / MARKE.c.rl) — line classification is ragel's job, no manual byte scanning (CLAUDE no manual parsing).mark/test (strict-dir harness); wiki/blog pages render via make, so re-render must stay byte-stable except the fixed lists.para_li mode (para_del for wontfix <del>): the para continuation path does the same-depth join, mark_para_flush emits li-text bare (no <p>), budget-checks the JOINED entry as "bullet"; sub-4-space leftover indent trimmed at the join (spec ruling StrictMark.mkd:65 "a block continues on later lines via a 4-space indent").--strict (none in the wiki today).mark/test/MARK.c: 7 new (ul/ol/todo/del join, sibling, deeper-indent still nests, same-depth AFTER a nested block stays nested <p>) + 1 AMENDED existing case nest-item-para (it pinned the buggy nested-<p> render; unsatisfiable with a green suite — veto reverts it). RED evidenced via the pre-fix deployed binary.MARK.c (para_li/para_del + flush branch + leftover-indent trim); verified by orchestrator 2026-07-07: MARKtest+MARKstrictdir 3/3, probe render joins the li and the deeper-indent case is byte-identical old-vs-new; whole-wiki re-render diff = only POST.html + Submodules.html, both this fix; --strict breach lists byte-identical.be log has no MARK-001 yet).make) + redeploy ~/bin/mark (still pre-fix, 06:14 build) — user's call.mark/MARK.c, mark/test/MARK.c), by explicit no-worktree order.