Craig pasted the actual lost Day 43 conversation, confirming the reconstructed diary's account: a breadcrumb
color bug fixed on Learn_by-catagory.html, a Home-prefix breadcrumb added to
sequence-index.html, the three-file cross-link between sequence-index.html,
Learn_by-catagory.html, and EXPANDED-index.html, and the design of a 4-button banner
split (Commands / Processes & Architecture / All Sections / Search Site) that Craig then built solo —
catching and fixing a stray-paste bug himself before saving.
The one real gap left over from Day 43: EXPANDED-index.html linked out to
sequence-index.html via "All Sections," but nothing pointed back. Added a sentence to the existing
.intro-note box — "New here? Try the guided Sequence for a structured path through the basics
instead of browsing" — confirmed rendering correctly via screenshot. Craig also added the page's dark
slate/orange header-card treatment as a small mod of his own; confirmed intentional as a one-off style rather than
a shared class.
sequence-index.html00-sequence-index.html, 01-sequence-index.html,
3-buttons_sequence-index.html, and the sequence-index.html~ backupNote: reported by Craig, not yet independently re-verified against a fresh tar.
check_tree -a on EXPANDED found a real bug: SED-AWK/index.html didn't exist at all
(the directory only held sed-awk-primer-claude.html). Craig fixed it — deliberately as a plain
file rather than a symlink, since with only one file in the directory, a separate descriptively-named file plus a
symlinked index.html was agreed to be redundant. Treated as an intentional exception to the site's usual
index.html-as-symlink convention.
Found reversed relative to every other EXPANDED symlink (Learn_by-catagory.html -> index.html
instead of the other way around). Craig decided to leave it as-is — a known, deliberate anomaly, not a bug
to chase.
Craig checked AI/learning/how-to-use-this-site.html directly and confirmed it's AI Education site
content (the four AI-learning paths), unrelated to LearnLinux's sequence-index.html. Decided it does
not need the sequence link — different section, different topic, no reason to force a connection.
Ran Find-file-NO-breadcrumb.sh across CMDS/, cmds/commands/, the
LearnLinux/ root, and the /AI root. Every flagged item across all four runs turned out
to be either a non-HTML file (PDFs, CSS, shell scripts, no-extension notes) or a file already sitting inside an
established exempt directory (SAVED/, biomes/, remember/,
glossary/NOTES/). No genuine gaps found anywhere. This finally and fully closes the breadcrumb
rollout item that's been carried since Day 30.
Follow-up identified but not yet done: Find-file-NO-breadcrumb.sh should skip non-HTML file types
outright instead of flagging them on every run.
The persistent /cgi-bin/CGI-LinuxSearch.cgi "MISSING" flag turned out to be a real bug in the
script, not a false positive to just live with. .cgi being in SKIP_EXTS had no effect,
because that list is only consulted by the orphan-detection pass (STEP 3) — the broken-link checker (STEP 2)
has its own separate, hardcoded skip logic that never saw it. Craig added
[[ "$href" == /cgi-bin/* ]] && continue to STEP 2. Confirmed via a full clean run afterward:
0 orphans, 0 broken links, 23 symlinks.
check_tree -a on EXPANDED: Clean! No orphans, no broken links. (119 linked, 23
symlinks.) Breadcrumb coverage confirmed complete across the entire site.