The top-level toggle is a tool choice, not the two classification axes (those
are the By-tracking / By-transmittal tabs inside Classify & Copy). Default to
the Classify & copy workflow and relabel the toggle 'Classify & copy' /
'Rename in place' so its purpose is clear; the in-place spreadsheet stays one
click away. 'Use Local Directory' now opens in Classify mode too.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A folder with files but no subfolders got no expand toggle, so in Classify &
Copy mode its files (the drag source) could never be revealed — and leaf
folders full of files are exactly where the work is. Make a folder expandable
when it has files in classify mode; expanding lists the draggable file rows.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The classifier re-scanned the source on every session; on cloud-backed mounts
(OneDrive/Samba) that's minutes of per-op latency. Workspaces fix it: scan a
folder ONCE, snapshot the completed tree, and resume instantly — all
classification runs on the data model; the filesystem is only touched at copy.
- persist.js v2: multi-workspace IndexedDB (tiny 'index' store for the welcome
list + 'data' store holding the source handle, tree snapshot, and map). DB v2.
- scanner.js: snapshotTree()/loadSnapshot() (compact, handle-less, marked done,
totals recomputed) + lazy resolveFileHandle/resolveDirHandle from the root.
- workspace.js: welcome manager (new/open/rename/delete), debounced autosave of
the active workspace, 'Refresh from disk' (re-scan → re-snapshot, path-keyed
map carries over). New workspace = the one slow full scan; reopen = instant.
- copy.js: resolves snapshot files' handles from the workspace root with a
one-click read permission re-grant; missing-on-disk files surface as errors.
- app.js: enterAppShell() shared by rename/workspace flows; exposes setMode;
classify.js decoupled from persistence.
- template/css: welcome workspace list + header 'Workspaces' button.
- tests: snapshot round-trip, persist CRUD + classify-only-preserves-tree,
copy-from-snapshot via mock root handle (28 classify/classifier tests green).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The scan is I/O-bound on cloud-sync / network mounts (OneDrive, Samba) where
each directory read is a high-latency round-trip. More in-flight reads hide
that latency on the many-folders case. (A single large folder is still
enumerated one entry at a time by the File System Access API and can't be
parallelized.)
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The Copy button (enabled once >=1 file is fully classified) copies the mapped
files into a user-chosen output directory under their canonical names/layout
<party>/{received,issued}/<transmittal>/<filename> — reading the source, never
writing it.
- copy.js: plan() (complete, non-excluded files) → conflict scan (two sources
→ same output path are reported + skipped) → copyTo() engine on the generic
FS-Access shape (ensureDir + getFileHandle + createWritable). Per-file dedup:
identical target (sha256) is skipped; existing-but-different is left
untouched and reported; live footer progress; completion toast.
- app.js: restores the saved map on launch (keyed by source-relative path, so
it re-attaches when the same directory is re-opened) and persists the source
handle on open; Copy button wired.
- target-tree.js: enables/labels the Copy button from the done count.
- 2 copy-engine tests with mock FS handles (copy/skip/differ + conflict);
24 classify+classifier tests green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Each source file row shows a classification state dot (unassigned →
has-tracking/transmittal → done), and each folder shows an aggregate dot
over its subtree.
- Right-click a file or folder to Exclude/Include from the copy (folder applies
to its whole subtree) or clear an axis; excluded files are struck through and
never copied.
- Cross-tree find is bidirectional: click a placed file in the target pane to
reveal+flash it in the source tree (expanding its folders); click a source
file to switch the target pane to its placed axis and flash the node.
- Target pane now reverse-looks-up over ALL scanned files (the left tree), not
the selection-scoped grid, with placements grouped in one pass per render.
- classify.getAssignment() read-only accessor; 5 new tests (18 total green).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
In Classify & Copy mode the left tree now lists each folder's files as
draggable rows (with a classification state dot), and folder rows are
draggable for a group-drag of the whole subtree. Target-tree nodes are drop
zones: a tracking folder (any node) or a transmittal bin; dropping assigns the
dragged source key(s) along that axis via classify.place().
- dnd.js: drag-payload bus (keys held in a module var since dataTransfer can't
be read during dragover; carries a marker for the copy cursor).
- tree.js: createFileElement + group-drag dragstart; classify-mode file rows.
- target-tree.js: setupDropZone with dragover highlight + drop assignment
(tracking = any node, transmittal = bins only).
- app.js: source tree re-renders on classify state change.
- 2 DnD drop-handler tests (14 total green).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Header gets a Rename / Classify & Copy switch. In Classify & Copy mode the
spreadsheet pane is replaced by a tabbed target pane (By tracking number /
By transmittal), while the source tree stays on the left.
- target-tree.js: renders both trees from classify state; tracking-folder
create/rename/delete (leaf folders styled as the revision); party CRUD +
per-slot inline transmittal-bin form (date + TRN/SUB + seq + optional
status/title); shows the derived filename + a validation badge for each
placed file; live header stats (done / in progress / unassigned / excluded).
- app.js setMode(): swaps panes, toggles classify mode, re-renders both trees.
- 3 UI smoke tests added to classify.spec.js (12 total green).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Foundation for the non-destructive map+copy workflow: source stays read-only,
files are mapped onto two orthogonal target trees, a later step copies renamed
copies to a separate output dir.
- classify.js: the single source of truth. assignments map keyed by
source-relative path (survives re-pick); tracking tree (positional: ancestors
joined '-' = tracking number, immediate parent 'REV (STATUS)' leaf = rev+status,
title from original name) and transmittal tree (<party>/{received,issued}/<bin>).
deriveTarget() computes filename + output path + validation purely; pub/sub +
debounced autosave; node CRUD with dangling-placement cleanup.
- persist.js: IndexedDB store of the serialized map + the source
FileSystemDirectoryHandle, with queryPermission/requestPermission re-grant on
reload and a re-pick fallback.
- tests/classify.spec.js: 9 in-page unit tests for the derive/assignment logic
(no FS Access needed) — tracking join, leaf REV (STATUS) parse incl. invalid
status, title derivation/override, transmittal path composition, exclude,
cascade delete.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The black-completed vs grey-flashing distinction was too subtle. Completed
numbers (the direct count, always; the +total once final) now render in
var(--primary) — theme-aware blue in both light and dark. While a subtree
is still scanning its +total stays muted grey + pulses, so blue = done,
grey = in progress. Once both numbers are blue the row's folders/files
labels turn blue too (.folder-count.done .ct-label).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Scan concurrency: the scan is I/O-bound — each directory read is a network
round-trip to the share, so the lever is parallel in-flight reads, not CPU
threads. Replace the per-level BFS barrier (which idled workers waiting on
the slowest dir in each level) with a continuous shared-queue pool that
keeps up to SCAN_CONCURRENCY (16, up from 6) reads in flight at once,
pulling newly discovered child dirs as they land. Still roughly
breadth-first (FIFO), so top levels surface first. ensureScanned reuses it.
Error messages: translate File System Access DOMExceptions into accurate,
actionable text keyed on err.name (not the cryptic raw message, which reads
like a permission problem when it isn't). e.g. InvalidStateError now reads
'the folder changed on disk since it was first read … rescan' instead of
'an operation that depends on state cached in an interface object …'. The
raw name+message is appended in parens for copy-paste troubleshooting.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The scan was slow because it OPENED every file (getFile() for size/lastModified
— which the grid doesn't even display) and read every ZIP inline. On a network
share that's a round-trip per file. Now:
- createFileObject builds rows from the directory entry name alone, no
getFile(); size/lastModified load on demand (preview/SHA/rename already call
getFile() themselves). The scan is now a pure directory listing.
- ZIPs are lazy: a .zip is an expandable node read only when opened
(scanZipNode), not during the walk.
- Footer shows live elapsed time (ticks every second), and a success toast
fires at completion with totals: "Scan complete — N folders, M files in Ts."
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The live "Scanning… N folders · M files — <path>" status now lives in a
persistent page footer bar instead of under the folder-tree header.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Counts now read "direct+total" — e.g. "(2+10 folders, 15+300 files)". The
direct number (immediate children) shows as soon as a folder's own directory
is read; the total (whole-subtree) is accumulated progressively and flashes
grey until the subtree is fully scanned, then goes solid. The "+total" is
omitted once done and there's nothing deeper.
- Scan errors (permission denied, network hiccups on a share) now surface as a
toast (de-duped per path) instead of only console noise; a failed folder/zip
is marked done-empty so it doesn't wedge the walk.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Replaces the full depth-first "scan everything, then render once + expandAll +
selectAll" walk (which looked stalled and was a render bomb on a large network
drive) with a progressive, breadth-first scan:
- Walks level-by-level behind a bounded worker pool (6), rendering as it goes —
the top folder levels appear immediately, deeper levels fill in the
background. Workers await between directories so the UI stays responsive.
- Live status line under the tree header: "Scanning… N folders · M files —
<current path>", ending "Scanned … in Ts."
- Per-folder state machine (pending → scanning → children → done) with
immediate subfolder/file counts; the row is greyed (with a faint pulse) until
its whole subtree is scanned, then turns solid — the at-a-glance signal.
- Opening a folder jumps its subtree to the front of the scan (ensureScanned),
so an opened folder always shows complete contents; idempotent vs the
background walk.
- No more auto-expand/auto-select-all (that loaded the entire drive up front);
the root is selected so the grid shows its files immediately.
- ZIPs stay expandable, scanned inline into virtual nodes (already in memory
once read); whole zip subtree marked done at once.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Drop the bottom-right floating "Admin mode" switch in favour of a proper
account menu in the header's upper-right (every tool's .header-right).
New shared/profile-menu.{js,css}: a circular avatar button (email initial)
opening a dropdown with the signed-in email, an "Admin mode" item (only for
can_elevate principals — drives elevation.setOn/setOff, drops on leave),
Profile (/.profile), and Access tokens (/.tokens). The panel is portaled to
<body> + position:fixed so it overlays content reliably regardless of the
app's stacking contexts; the button shows a red ring while elevated.
No logout: authentication is the upstream proxy's concern (oauth2-proxy /
Authelia) — ZDDC owns no session, so the menu doesn't render sign-out.
elevation.js keeps the state machine (cookie, armed banner/frame, ephemeral
pagehide-clear, zddc:elevationchange, ?admin= URL) but no longer renders any
control — the profile menu is the UI. elevation.css drops the floating-
toggle styles (keeps banner + frame). All 7 templates drop the dead
elevation-toggle placeholder; all 7 build.sh bundle profile-menu.{js,css}.
Validated in a containerized browser: menu items, links, elevation arming +
armed ring, dropdown overlays content, no floating toggle.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
saveAllFiles() carried a 200ms inter-operation sleep "to prevent race
conditions" plus a 300ms post-error "settle" sleep. But saveFile() is fully
awaited and each iteration renames a distinct file, so the saves are already
serialized — the sleeps were the band-aid for an earlier missing-await bug
(the "ensure properly awaited" comment marks where that got fixed). Remove
both: correctness comes from the awaits, the operations are independent, and
Save All no longer pays 200ms per file (≈10s on a 50-file batch).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Three small helpers under window.zddc.cap, wired into every tool's
build:
cap.at(path) — Promise<AccessView|null>. Fetches
/.profile/access?path=<urlpath> and
memoises per-path for the session.
Used by tools to gate top-of-page
affordances on path_verbs / path_is_admin
/ path_can_elevate_grant.
cap.has(node, verb) — boolean. Reads the listing entry's
verbs string for the named verb.
Falls back to node.writable for 'w'
when verbs is absent (offline FS-API
listings or pre-promotion clients).
cap.handleForbidden(resp, — parses a 403 response's JSON body for
opts) missing_verb and renders an error
toast. When opts.path is supplied AND
the path-scoped access view reports
path_can_elevate_grant covering the
missing verb, the toast appends an
"Elevate" button that flips the
elevation cookie and reloads.
Browse loader.js + tree.js carry the new verbs field through to the
node objects so context-menu gating can call cap.has(node, 'w'|'d')
without changing the legacy node.writable contract. New CSS rule
.zddc-toast__action styles the inline Elevate button.
Concatenation order: cap.js comes after toast.js + elevation.js so
the dependencies (window.zddc.toast, window.zddc.elevation) are
present at module-load time.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Three UI cleanups against the admin/browse chrome.
Red admin-mode frame (shared/elevation.css)
Was: body { outline: 3px ... ; outline-offset: -3px } — an outline
doesn't reflow content, so in tools that butt their content to the
viewport edge (browse split-pane, archive grid) the frame painted
on top of the first 3px of content.
Now: body.is-elevated::after { position:fixed; inset:0; border:3px;
pointer-events:none; z-index:9200 }. The frame lives in its own
fixed layer above all content, so it never overlaps or steals
clicks; content layout is unchanged.
Project-stage strip (Archive · Working · Staging · Reviewing)
Low-value chrome. Removed entirely:
- delete shared/nav.js + shared/nav.css
- drop the include from every tool's build.sh
(browse, transmittal, form, archive, landing, tables, classifier)
- delete tests/nav.spec.js
- rebuild tables.html (the //go:embed'd baked-in copy)
Project navigation already happens through the directory tree in
browse and the URL bar; the strip duplicated breadcrumb information
without adding capability.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Polish pass after the big refactor in 2d114fc.
== Header elevation slot propagated ==
shared/elevation.{js,css} surface a header checkbox for admins.
30-minute sudo-style cookie window (Max-Age=1800, SameSite=Lax).
Only renders when /.profile/access reports can_elevate=true; quiet
for non-admins. Slot added to all 7 tool templates and concat'd
into all 7 build.sh files; admin in any tool now sees the toggle.
Three text-rename ride-alongs in archive/classifier/transmittal
templates: "Add Local Directory" → "Use Local Directory" (the same
rename that landed in browse earlier in this branch).
== Docs ==
- CLAUDE.md gets an "Admin elevation is sudo-style" paragraph in
the "Things that bite if you forget" section.
- AGENTS.md gets a dedicated "Admin elevation (sudo-style)" section
alongside "Bearer tokens" — same depth as the existing auth docs.
== Helper file splits ==
The retired form editor's shared helpers got bundled into a single
zddc_admin.go in the cleanup; that name is now misleading. Split by
concern:
- admin_helpers.go: hasAnyAdminScope (the only admin-specific helper)
- paths.go: resolvePath, urlPathOf, chainDirs (URL ↔ filesystem path
math — used by several profile / zddc-file handlers)
- profile_assets.go (renamed from zddc_admin_assets.go): custom CSS
pipeline. URL renamed from /.profile/zddc/assets/ → /.profile/assets/
since /.profile/zddc/ no longer hosts an editor.
- treeEntry moves to profilehandler.go (alongside AccessView, its
only consumer).
- writeError moves to profileprojects.go (its only consumer).
== Smell cleanup ==
- zddc.HasAnyAdminGrant(fsRoot, email) — new elevation-independent
primitive that walks the cascade and reports whether email is named
in any admin: list anywhere. Replaces the synthetic-elevated probe
hack in enumerateAccess (`Principal{Email, Elevated: true}` was
"lying" to the elevation gate to ask what it would say). The handler's
hasAnyAdminScope collapses to a 4-line wrapper that gates on
p.Elevated and delegates.
- Access-log middleware records `elevated` per request, so forensics
can distinguish "admin acting as user" from "admin exercising power."
- browse/js/app.js's ?file= deep link walks multi-segment paths. Each
intermediate segment is matched + expanded; the leaf gets
selected/previewed. Auto-shows hidden when any segment starts with
. or _. Silently no-ops on unresolved segments.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
System-default font stack ('-apple-system, BlinkMacSystemFont, Segoe UI,
…') is the textbook generic admin-tool look. The tools have a real point
of view (engineering documents, traceability, immutability); the
typography should reflect that.
Picks:
--font → IBM Plex Sans (400 + 600). UI body text. Distinctive
engineering sans with tabular nums and proper figures.
--font-display → Source Serif 4 (600). Headings, page titles,
.app-header__title. Reads as "document" not "UI label."
--font-mono → unchanged. Platform mono fonts are already excellent
and engineering tools rarely benefit from a custom mono.
Wiring:
- Raw .woff2 files live in shared/fonts/ (~60 KB total, latin subset,
SIL OFL 1.1 — both families)
- shared/fonts.css is base64-inlined data URIs for those three fonts
(~80 KB after b64 overhead). Generated once from the snippet in
shared/fonts/README.md.
- Every tool's build.sh prepends shared/fonts.css before shared/base.css
so @font-face is parsed before any rule references the family names.
- Headings (h1-h6) and .app-header__title now use var(--font-display);
.app-header__title bumped 17→18px and letter-spacing reset since the
serif doesn't need the original sans-text tightening.
- table/code/.tabular-nums get font-variant-numeric: tabular-nums so
tracking-number columns align vertically.
"Ship the record player with the record": zero CDN dependency at render
time. Tools render identically offline and online. Per-tool dist sizes
grew by ~80 KB.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Reshape browse from "tree-as-table with popup preview" into a unified
file-experience tool with three layered behaviors:
Phase A — Two-pane shell
Phase B — Markdown plugin (Toast UI inline)
Phase C — Grid mode (classifier workflow)
Phase D — Deprecation banners on standalone classifier + mdedit
= Phase A: two-pane shell + lightweight preview plugins =
Browse's table view becomes a tree-pane on the left + preview-pane on
the right with a draggable resizer. Click a folder → expand inline.
Click a file → render in the right pane. The previous popup window
becomes an explicit "⤴ Pop out" button in the right-pane header for
users with a second monitor.
Preview rendering reuses shared/preview-lib.js (PDF iframe, image
<img>, TIFF, ZIP listing, text <pre>). Unknown types show a download
link. browse/js/preview.js refactored into renderInline (default) +
renderInPopup (Pop out button); both share the same plugin
dispatch logic.
Filter rows were already removed earlier this session. Sort columns
likewise — the tree is alphabetical by default; the underlying
setSort API still exists for future re-introduction.
= Phase B: markdown plugin =
New browse/js/preview-markdown.js: when a .md or .markdown file is
clicked, the right pane mounts a Toast UI editor (initial-value =
file contents) with a small toolbar containing Save + dirty indicator
+ status text. Save sends PUT through the file API for server-mode
files; non-server sources are read-only for now (deferred to a
follow-up that wires zddc-source.js writes too). Ctrl+S / Cmd+S
inside the editor saves.
Toast UI Editor (~700 KB JS + ~160 KB CSS) was previously bundled
only in mdedit/vendor/. Moved to shared/vendor/ so browse and mdedit
both pull from one location.
= Phase C: grid mode =
View-mode toggle [Browse | Grid] in the toolbar. Grid mode loads the
classifier tool as an iframe scoped to the current directory (server
mode at working/staging/incoming locations) — classifier's full
bulk-rename workflow without leaving browse. v1 implementation; a
future iteration could bundle classifier's modules directly into
browse for tighter integration. Hostile cases (file:// origin, paths
outside working/staging/incoming) show a friendly explanation
instead of a blank iframe.
new browse/js/grid.js handles the activation logic.
= Phase D: deprecation banners =
mdedit and classifier standalones gain a "this tool is being absorbed
into Browse" advisory banner. Both standalones remain fully
functional and continue to ship — they're useful for offline single-
file editing and air-gapped environments. The banner just points
users toward the unified browse experience.
= Files =
+ browse/js/preview-markdown.js (markdown plugin)
+ browse/js/grid.js (grid-mode plugin)
M browse/template.html (two-pane layout, view toggle, banners)
M browse/css/tree.css (two-pane CSS, replaces table styles)
M browse/js/init.js (state additions: selectedId, viewMode)
M browse/js/tree.js (rowHtml: <tr>+<td> → <div>)
M browse/js/preview.js (renderInline / renderInPopup split)
M browse/js/events.js (toggle wiring, resizer, click handlers
adapted from <table> to <div>)
M browse/build.sh (Toast UI vendor + new modules)
R mdedit/vendor/toastui-* → shared/vendor/ (one bundle, two tools)
M mdedit/build.sh (paths)
M mdedit/template.html (deprecation banner)
M classifier/template.html (deprecation banner)
M tests/browse.spec.js (selectors updated for new layout +
new "click file → preview" test)
Bundle sizes after this commit:
browse: ~1020 KB (was ~290 KB; added Toast UI ~700 KB)
classifier: ~1470 KB (unchanged from prior baseline)
mdedit: ~2140 KB (unchanged; vendor location moved but not added)
What's deferred:
- TOC + front-matter pane in browse's markdown plugin (mdedit has
these; browse v1 uses just the editor).
- FS-API writes from browse's markdown plugin (server PUT works).
- Classifier modules bundled directly into browse (v1 uses iframe).
- Sort UI in the new tree (model still supports it; no widget yet).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Removes every runtime CDN load. The "ship the record player with the
record" philosophy: a downloaded .html file works offline against any
file the user can open, with no network dependency at runtime.
Newly vendored under shared/vendor/:
- xlsx.full.min.js (SheetJS, 928 KB) — XLSX/XLS preview
- utif.min.js (UTIF, 57 KB) — TIFF preview
Already there but now used by mdedit too:
- jszip.min.js, docx-preview.min.js
Call sites updated to drop the `await loadLibrary(URL)` pattern —
since the vendor JS is concatenated into the inline <script> at build
time, window.XLSX / window.JSZip / window.UTIF / window.docx are
available synchronously from page load.
Per-tool changes:
- archive/build.sh: +xlsx, +utif
- classifier/build.sh: +xlsx, +utif
- transmittal/build.sh: +xlsx, +utif
- mdedit/build.sh: +jszip, +docx-preview, +xlsx, +utif
(mdedit was the only tool not yet
bundling any of the preview deps)
- browse/build.sh: +utif
- archive/js/table.js, classifier/js/preview.js,
transmittal/js/files-preview.js, mdedit/js/file-tree.js (×2):
drop the `await loadLibrary('…cdn…')` lines.
- shared/preview-lib.js:
drop the loadLibrary(UTIF) / loadLibrary(JSZip) wrappers; assume
window.UTIF and window.JSZip are present.
Net bundle-size delta after baking:
archive: +990 KB → ~1.47 MB
browse: +57 KB → ~292 KB
classifier: +990 KB → ~1.43 MB
mdedit: +1100 KB → ~2.09 MB
transmittal: +990 KB → ~1.63 MB
Docs (AGENTS.md, ARCHITECTURE.md) updated: removed the "runtime CDN
loading exception" paragraph and the table row that flagged xlsx as
CDN-loaded.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
classifier/js/tree.js was inserting a <div class="empty-state">No
folders found</div> inside the folder-tree pane when the tree was
empty. That conflicted with the shared .empty-state rule promoted
in the previous commit — which expects an outer flex container with
a child .empty-state__inner card, used for the top-level welcome
overlay.
The two usages aren't the same thing semantically (one is the
welcome screen; one is a tiny inline "list is empty" placeholder
inside the folder tree). Rename the inline one to .tree-empty to
remove the collision. The spreadsheet.css rule that targeted the
old class is renamed to match; same padding/text-align/color.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Three tools (archive, browse, classifier) independently implemented
an empty-state pattern with three different CSS class naming
conventions and slightly different rules:
archive: .empty-state + .empty-state-content (BEM-less)
browse: .empty-state + .empty-state__inner (BEM)
classifier: .empty-state + .empty-state-content (BEM-less)
Same visual intent ("nothing's loaded yet — here's a welcome card
with instructions"), implemented three times with subtly different
spacing, no shared body styling for h2/p/ul/li, and incompatible
class names that prevented a future tool from copy-pasting the
pattern.
Promote a single consolidated rule set to shared/base.css using
BEM naming throughout:
.empty-state — base (flex centered, padding)
.empty-state--overlay — modifier: position absolute,
top 50px to clear app-header,
z-index 10. Used by archive
and classifier (their empty
states sit OVER the main
layout).
.empty-state__inner — content card (left-aligned,
text-muted, max-width 640)
.empty-state__inner--centered — modifier: tighter max-width
500, centered text, 2rem
padding. Used by tools whose
welcome screen reads as a
centered card.
.empty-state__inner h2/p/ul/ol/li — typography defaults
.empty-state__inner .note — italic small-print
.welcome-list — bullet list with left-aligned
text + auto margins; safe to
nest inside a centered card.
Per-tool changes:
- archive/template.html, archive/js/app.js: rename
.empty-state-content → .empty-state__inner empty-state__inner--centered;
add .empty-state--overlay to the outer .empty-state container.
Also the runtime-injected unsupported-browser markup in
showUnsupportedBrowserMessage() and the showHttpErrorState
selector.
- classifier/template.html: same renames.
- archive/css/layout.css + components.css: delete .empty-state*
and .welcome-list rules (now in shared).
- classifier/css/layout.css: same. Keep .empty-state.drag-over
locally — classifier is the only tool whose empty state acts
as a drop target.
- browse/css/base.css: delete .empty-state* (shared covers it).
browse's template was already using .empty-state__inner so no
template change needed.
LOC: shared/base.css gains ~70 lines; per-tool overrides lose ~85
combined. Net -15. More importantly, future tools can reuse the
pattern by adding two divs and (optionally) the --centered or
--overlay modifiers; no copy-paste required.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Five tools (browse, transmittal, landing, form, tables) were rendering
their headers as vertically-stacked blocks — the .app-header flex
container correctly laid out its left/right groups, but those groups
themselves had no display:flex rule, so their children (logo, title,
build label, action button) defaulted to block-level stacking.
Three tools (archive, mdedit, classifier) hid the bug because they
each carried their own copy of the .header-left/.header-right flex
rule in tool-local CSS. Same intent, slightly different gap values:
archive: left gap 0.75rem, right gap 0.5rem
mdedit: both 0.75rem
classifier: both 0.5rem
Promote the rule to shared/base.css alongside .app-header (where the
class is used in every template anyway): left 0.75rem, right 0.5rem
(matching archive — the layout the user pointed at as the reference).
Delete the three local duplicates.
Now all eight tools use the same header chrome contract: logo + title
group + primary action laid out horizontally with consistent gaps,
icon buttons grouped tighter on the right.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The .app-header__logo SVG was decorative on every tool. Web's
strongest convention is "click logo → go home" — so users tapping
it expecting that fallback got nothing. Now the logo is wrapped in
an anchor whose href reflects the URL the page was loaded from:
file:// → no wrap (no server home to point at)
/ → wrap, href=/ (deployment root)
/index.html / /<tool>.html → wrap, href=/ (root, no project)
/<project>/... → wrap, href=/<project> (project landing)
The wrap happens client-side at DOMContentLoaded via shared/logo.js,
loaded by every tool's build.sh after toast/nav. Idempotent — a
template-supplied anchor or a second mount call is a no-op.
The companion shared/logo.css adds a subtle hover/focus affordance
(opacity 0.82, focus ring) so the logo reads as clickable without
otherwise altering its visual weight. Tools opt out by setting
window.zddc.logo.disabled = true before DOMContentLoaded (e.g. for
deployments that pin the logo to an external destination).
Five Playwright tests (tests/logo.spec.js) lock the contract:
no-wrap on file://, href=/ at root, href=/<project> in project
subtree, aria-label matches target, idempotent re-mount.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a thin nav strip directly under the app-header showing the four
canonical lifecycle stages from the transmittal-workflow spec:
archive · working · staging · reviewing. Each is a link to that
stage's directory under the current project. Current stage is
highlighted (bold + primary color, aria-current="page"). Strip
mounts as a sibling of .app-header on DOMContentLoaded — no
template changes needed in any tool.
Render rules (shared/nav.js shouldRender):
- location.protocol must be http: or https: (file:// has no project
structure to navigate within)
- a project segment must be detectable as the first path segment
(when it isn't a tool HTML file like /index.html or
/archive.html?projects=A,B). Multi-project view at the deployment
root therefore shows no strip.
Stage URL targets:
- Archive → <project>/archive.html (project-root archive view)
- Working → <project>/working/ (directory listing — mdedit auto-served)
- Staging → <project>/staging/ (directory listing — transmittal auto-served)
- Reviewing → <project>/reviewing/ (directory listing)
Convention-driven, not probed: if a deployment doesn't have one of
these folders the link returns 404. Operators on non-standard layouts
can opt out by setting window.zddc.nav.disabled = true before
DOMContentLoaded.
This pairs with the previous landing-tool change (single-project
click → <project>/archive.html). Together they give the user
both URL-bar manipulation AND visible navigation across the four
canonical project stages.
Five Playwright tests in tests/nav.spec.js exercise:
- non-render at deployment root
- render + active stage on <project>/archive.html
- render + active stage deep inside <project>/working/foo/mdedit.html
- canonical link targets
- mount position is sibling of .app-header
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Promote classifier's local toast (classifier/css/base.css + showToast
in classifier/js/excel.js) into shared/toast.{js,css}. Every tool's
build.sh now concatenates them, so window.zddc.toast(msg, level, opts)
is callable from any tool.
API:
window.zddc.toast('Saved.', 'success');
window.zddc.toast('Could not load: ' + err.message, 'error');
window.zddc.toast('Note', 'info', { durationMs: 3000 });
Levels: info (default) | success | warning | error. Single-toast
policy — a second call replaces the first. Click anywhere on the
toast to dismiss. ARIA: error → role=alert/aria-live=assertive,
others → role=status/aria-live=polite.
Class prefix is .zddc-toast (BEM-ish) to avoid colliding with any
tool-local .toast rules. Classifier's existing showToast now
delegates to window.zddc.toast — call sites in excel.js +
selection.js are unchanged. Classifier's local .toast CSS block
deleted in favor of the shared one.
This commit only EXPOSES the API. Replacing the ~25 alert() call
sites scattered across archive/transmittal/mdedit/classifier with
toast calls is left as follow-up — each alert needs per-call review
to decide if it's truly non-blocking.
Five Playwright tests in tests/toast.spec.js lock the contract:
API exposure, level mapping, ARIA roles, single-toast replace,
click-to-dismiss.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces the binary acl.allow/deny model with five permission verbs
(r/w/c/d/a) and first-class roles, and adds an authenticated file API
(PUT/DELETE/POST move/mkdir) so the HTML tools can edit-in-place over
HTTP. Closes the AC-3(7) and AC-6 federal-readiness gaps.
File API (zddc/internal/handler/fileapi.go)
- PUT <new> → action c
- PUT <existing> → action w
- PUT <.zddc> → action a (CanEditZddc strict-ancestor rule)
- DELETE → action d
- POST mkdir → action c (auto-writes creator-owned .zddc when the
parent is Incoming/Working/Staging)
- POST move → action w on src + c on dst, atomic via os.Rename
- Optional If-Match for optimistic concurrency, --max-write-bytes cap,
audit log emits a structured file_write event per operation.
Permission model (zddc/internal/zddc/{acl,file,roles,cascade_mode}.go)
- acl.permissions: { principal → verb-set } map; principals are email
patterns or role names. Empty verb set is an explicit deny.
- roles: { name → members } definitions, available at the level they
declare and all descendants. Closer-to-leaf shadows ancestor.
- Legacy acl.allow/deny still work; they fold into permissions at
parse time (allow → "rwcd", deny → "").
- Cascade walks leaf→root; first level with any matching entry wins;
the union of matching verb sets at that level decides.
- --cascade-mode=strict adds a root→leaf ancestor-deny pre-pass so an
ancestor explicit-deny is absolute (NIST AC-6). Default delegated
preserves the existing commercial behavior.
Special folders (zddc/internal/zddc/special.go)
- Incoming / Working / Staging: mkdir auto-writes a .zddc into the new
subdir granting created_by + that email rwcda directly. Same form
operators write by hand; creator can edit it later to add others.
- Issued / Received: server-enforced WORM split. Cascade grants
inherited from above the WORM folder are masked to r only; grants
placed at-or-below the WORM folder retain r,c. Operators grant
write-once (cr) to the doc controller via an explicit .zddc at the
Issued/Received folder. Admins exempt — only escape hatch.
Browser polyfill (shared/zddc-source.js)
- HttpDirectoryHandle + HttpFileHandle implement the FS Access API
surface (values, getFileHandle, createWritable, removeEntry,
queryPermission/requestPermission) over zddc-server's listing JSON
and file API. Existing tools written against showDirectoryPicker
work unchanged.
- detectServerRoot() returns { handle, status }: tools auto-load on
HTTP, surface a clear "no permission to list" message on 403, and
fall back to the welcome screen on 0.
- classifier renames take the atomic POST move path on HTTP-backed
handles; mdedit and transmittal route reads/writes through the
polyfill so prior FS-API code paths cover both modes.
Tests
- zddc/internal/zddc/{cascade_mode,roles,special,acl}_test.go cover
delegated vs strict, role membership / shadowing / legacy fallback,
WORM split semantics, verb-set parser round-trip.
- zddc/internal/handler/fileapi_test.go now also covers role-based
vendor scenarios, WORM blocking vendor & doc controller writes,
explicit Issued .zddc unlocking the cr drop-box, admin bypass,
auto-ownership on mkdir, and strict-mode lockouts.
Docs
- ARCHITECTURE.md + zddc/README.md document the verb model, role
syntax, special-folder behaviors, cascade-mode flag, and full file
API surface. Federal-readiness gap analysis strikes AC-3(7) and
AC-6.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Same pattern as the browse fix. archive, transmittal, classifier
previously CDN-loaded jszip + docx-preview on first preview of a
.zip / .docx file via shared/preview-lib.js's loadLibrary helper.
That meant each first-preview blocked on a CDN round-trip + parse,
and broke entirely under restrictive networks or CSPs.
Vendor both libs under shared/vendor/ and concat them at the top of
each tool's build, ahead of init.js. window.JSZip + window.docx are
now defined immediately on page load. Drop the redundant loadLibrary
calls (and classifier's stray <script src="cdn..."> tag in the
template, plus archive's bespoke loadJSZip helper in export.js).
xlsx (SheetJS) intentionally stays CDN-loaded — at ~900 KB it's too
large to inline, and only fires on .xlsx preview which is a rarer
path.
Bundle size impact (uncompressed):
archive: 304 KB → 476 KB (+172 KB)
transmittal: 449 KB → 621 KB (+172 KB)
classifier: 252 KB → 424 KB (+172 KB)
With the gzip middleware (~75% reduction on HTML) and ETag-cached
revalidation now in place, the wire-size delta is ~40 KB per tool
on the first load and 0 on every subsequent load until redeploy.
Bring every tool's header in line with archive's pattern:
[logo] [title] [version] [Add Local Directory] [⟳] ............... [◐] [?]
------------- header-left --------------- ----- header-right -
Changes per tool:
* browse: rename "Select Directory" → "Add Local Directory"; add the
red-non-stable wrap to the build label (was missing); add a help
panel + bundle shared/help.js.
* classifier: rename selectDirectoryBtn → addDirectoryBtn,
refreshBtn → refreshHeaderBtn for consistency. Update all JS
callers and welcome-screen copy to the new label.
* mdedit: same id rename. Move the previously-in-pane refresh
button into the header. Stop renaming the dir button to
"Directory: <name>" once a folder is loaded — instead use the
shared btn--subtle variant to de-emphasize while keeping the
standard label.
* transmittal: convert non-standard <div class="app-header"> with
spacer/icons containers to <header class="app-header"> with the
canonical header-left/header-right pair. Move the publish split-
button into header-left (Transmittal-specific primary action).
Remove dead .app-header__spacer/__icons/header-icon-btn CSS now
that nothing references those classes.
* landing, form: add help-btn + help-panel + bundle shared/help.js.
Each panel is tool-specific (project picker docs for landing,
schema-driven form docs for form).
Cross-cutting:
* shared/base.css: promote .btn--subtle from browse/css/tree.css
so any tool with an online mode can de-emphasize Add Local
Directory consistently.
Verified all 7 tools in headless Chromium: header structure correct,
build label red on non-stable cuts, help panel opens + closes via
button + Esc.
Cross-tool header inconsistencies cleaned up after the audit
prompted by the browse Phase 2 work:
- landing/template.html: title was 'ZDDC Archive' (a holdover from
when landing WAS the archive). The page is now the project
picker — title shortened to plain 'ZDDC'. Browser tab title
follows: 'ZDDC Archive — Projects' → 'ZDDC — Projects'. Title +
build label wrapped in title-group div for layout consistency
with archive/classifier/mdedit/browse.
- form/template.html: title was bare; same title-group wrapping.
The id='form-title' stays — its content is overwritten at
runtime by form.js based on the form schema's name.
- classifier/template.html: refresh button text 'Refresh' →
'⟳' icon to match archive + browse. Same title attribute, just
smaller visual weight.
Untouched (intentionally):
- archive's button stays 'Add Local Directory' + addDirectoryBtn
id — semantically different from the others (archive
accumulates multiple directories; everyone else operates on
one). The naming reflects that.
- transmittal — different layout entirely (page-header with
sender/receiver logo cells); not a candidate for app-header
standardization.
User report: opening an .html file with a '../.archive/' hyperlink in
a new tab works (zddc-server intercepts and serves the right file),
but clicking the same link inside the file previewer does nothing.
Two combined causes:
1. The previewer's iframe was loaded from a blob: URL (built from
the file's bytes). Relative URLs in the iframe resolve relative
to the blob URL — '../.archive/X.html' becomes 'blob:.../.archive/
X.html', which is gibberish. The browser never sends a request to
the server, so the .archive interception never fires.
2. sandbox="" disables every iframe capability including popups,
so even <a target=_blank> is silently swallowed.
Fix per tool:
- archive (table.js): for HTML preview, use file.url (the real
server URL) directly when available; fall back to blob only for
File-System-Access-API mode where there's no server to intercept
anyway. Now relative links in archived HTMLs resolve against the
actual server origin and the .archive interception fires as
designed. Sandbox loosens to allow-same-origin + allow-popups +
allow-popups-to-escape-sandbox so resources within the iframe
load and link clicks (default target / target=_blank / middle-
click) work normally. allow-scripts is intentionally NOT set —
archived HTML still cannot run JS in the popup's origin.
- transmittal (files-preview.js) + classifier (preview.js): same
sandbox loosening for consistency. These tools' files are
typically local (FileSystemAccessAPI), so the file.url branch
doesn't apply — relative URLs that depend on a server still
won't resolve in local mode (intrinsic limitation, no server).
Tested behavior preserved:
- PDFs: unchanged (no sandbox, browser's PDF viewer handles).
- Images / docx / xlsx / tiff / zip / text: unchanged.
- HTML in zddc-server-backed archive: relative '../.archive/' links
now navigate the iframe to the correct target file.
HTML files in the file previewer (archive, transmittal, classifier
popups) were dispatched to the text renderer because 'html'/'htm'
are in shared/preview-lib.js's TEXT_EXTENSIONS (which is shared with
the syntax-highlighting code path). Result: opening an .html file in
preview showed its source as a <pre> block, not the rendered page.
Fix in each tool's popup builder + dispatcher:
- Add 'html' / 'htm' to the iframe branch (alongside pdf), so the
popup ships an <iframe src=blob:...> instead of an empty
#previewContent div. The blob's MIME type from getMimeType()
is already 'text/html', so the browser renders natively.
- Skip the text-render dispatch for html/htm (the iframe is enough).
- Add to the HTML iframe so an arbitrary archived
HTML file cannot run scripts, navigate top, submit forms, or
open popups in the popup-window's origin. PDFs don't need this
since the browser's PDF viewer is sandboxed natively.
classifier/js/preview.js uses a getPreviewType() switch instead of
chained ifs; adds 'html' as its own preview type (checked BEFORE
'text' since html is in TEXT_EXTENSIONS).
mdedit already handled HTML specially (file-tree.js has an isHtml
check); no change there.
TIFF was already rendered via the shared zddc.preview.renderTiff
canvas viewer in all four tools — no change needed for that path.
If TIFF preview appears broken on the live prod server, that's the
v0.0.9-alpha-baked-in image; the fresh stable redeploy fixes it.
Adds shared/preview-lib.js with two cross-tool renderers:
- renderTiff (UTIF.js, lazy-loaded from CDN; PDF-style toolbar with
page nav, zoom, fit-width/fit-page; multi-page TIFFs decode lazily)
- renderZipListing (JSZip; sortable name/size/modified table, sticky
header, host-grouped paths)
Wired into the four tools that have a preview surface (archive, classifier,
mdedit, transmittal). Cross-document compatible so the same renderer works
for popup-window tools (archive/classifier/transmittal) and inline tools
(mdedit). Archive previously had no image branch at all — now previews
JPG/PNG/GIF/WebP/BMP/SVG natively, plus TIFF via UTIF, plus the ZIP listing.
Adds the dark-blue rounded-square favicon to each app's header (left of
the title) and to the website navigation. Single inline SVG, sized via
.app-header__logo (in shared/base.css) for tools and .brand-logo (in
website/css/style.css) for the website. Self-contained — the SVG carries
its own background, no wrapper styling needed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Removes the codeberg.org/varasys/zddc-server registry image, which had
no remaining consumer outside this shop. The two chart Dockerfiles
(tnd-zddc-chart) now compile zddc-server from source at build time,
fetching the right tag from a Codeberg release. release-image.sh,
zddc/Containerfile, and zddc/podman-compose.yaml are gone.
Build artifacts (HTML tools + zddc-server binaries) move from
website/releases/ in this repo to Codeberg release assets attached to
git tags. The website at zddc.varasys.io serves them by reverse-
proxying /releases/<tag>/<asset> to the corresponding Codeberg URL,
so consumers (zddc-use, level-2 bootstrap stubs, the chart
Dockerfiles) only ever talk to zddc.varasys.io.
Releases page becomes server-rendered static HTML regenerated on each
build via a single Codeberg API call. A small website/releases/manifest.json
maps <tool>-<channel> → tag for runtime channel resolution by zddc-use
and the level-2 stubs.
Files added:
- shared/publish-codeberg-release.sh — POSIX-sh helper that creates a
Codeberg release for a tag (sets prerelease flag from tag suffix)
and uploads/replaces release assets idempotently. Sourced by
build-lib.sh and zddc/release.sh.
- zddc/release.sh — replaces release-image.sh. Tags + cross-compiles
binaries via native Go (no podman needed; install Go) + uploads to
Codeberg release assets. No image build, no registry push.
Files modified:
- shared/build-lib.sh — promote_release tags + uploads via the helper
for stable AND alpha/beta now (alpha/beta were untagged before).
update_alpha removed; per-tool build.sh files no longer mirror to
website/releases/<tool>_alpha.html on plain dev builds.
- build.sh — prefers native go build over the old podman-based
cross-compile (which is gone with Containerfile). build_releases_index
queries the Codeberg API once and writes static HTML + manifest.json,
with graceful fallback when the API is unreachable.
- bootstrap/level2.html.tmpl — fetches manifest.json to resolve
channel → tag, then fetches the asset from /releases/<tag>/<asset>
(Caddy proxy). Replaces the old /releases/<tool>_<channel>.html flat
URL pattern. Operators with curl'd level-2 stubs need to re-issue
them — this is a breaking change.
- AGENTS.md, CLAUDE.md — rewritten to describe the new flow.
- .gitignore — releases/ artifacts now expected to be on Codeberg, not
committed locally.
NOT in this commit (deferred until $CODEBERG_TOKEN is provisioned):
- Backfilling existing tags as Codeberg releases.
- Cleanup commit: git rm-ing the existing artifacts in website/releases/.
Until backfill happens, those files are how operators with old
bootstrap stubs still get content. Once Codeberg has the assets,
drop them.
- The Caddy reverse-proxy config on zddc.varasys.io.
Operator-side changes (not in this repo):
- tnd-zddc-chart Dockerfile.prod and Dockerfile (dev) need updating
to compile from source rather than `FROM codeberg.org/...:stable`.
Done in a separate commit on that repo.
- Caddyfile rule for the /releases/<tag>/<asset> reverse-proxy.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Bundles a stretch of in-progress work across the SPA tools so the
tree returns to a coherent shippable state ahead of cutting a new
zddc-server stable image:
- landing: substantial rework of the project picker (sortable/filterable
table, presets refactor, ?projects= filter, ?v= channel propagation,
loading/error states)
- archive: presets cleanup, source.js refactor, filtering/url-state
alignment with the landing page
- mdedit: file-system module split, resizer, file-tree improvements,
base/toc styling tweaks
- transmittal/classifier: small template touch-ups for shared chrome
- shared: build-lib.sh helpers, new favicon.svg
- bootstrap, build.sh: pick up the channel-aware install/track zip
generation
- tests: new landing.spec.js, expanded archive/mdedit/build-label specs
- docs: CLAUDE.md picks up the zddc-server section and freshens the
alpha-build exception note
- regenerated artifacts: install.zip, track-{alpha,beta,stable}.zip,
*_alpha.html — these are produced by `sh build.sh` and per project
convention are committed alongside the source changes
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Every plain `sh tool/build.sh` invocation now reasserts a relative symlink
website/releases/<tool>_alpha.html → ../../<tool>/dist/<tool>.html so the
alpha hyperlinks always serve whatever dist currently holds. Idempotent — git
sees no churn on rebuild. `--release alpha` still wins by overwriting the
symlink with a real "alpha · <date> · <sha>" file; the next plain build
re-symlinks it.
Five existing alpha files become typechanges (regular file → symlink) — the
one-time migration cost. The reassertion survives deployment because the
website is served directly from the working tree.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
ZDDC — Zero Day Document Control. A file-naming convention plus five
single-file HTML tools (archive, transmittal, classifier, mdedit,
landing) and an optional Go HTTP server (zddc-server) with ACL and a
virtual archive index. Self-contained, offline-capable, dependency-free.
See README.md for an overview, AGENTS.md and ARCHITECTURE.md for the
build/release/architecture detail, bootstrap/README.md for the
two-level deployment install pattern, and zddc/README.md for the
HTTP server.
