Ships the fix from 04b38cb: a run lifted out of the document is text, so
the selection bar now offers Edit, size, bold and italic on it, and the
size picker carries the document's own size rather than rounding an 11 pt
line up to 12 on the way out.
Launched from a shell, the main process inherits that shell's stdout. When
the shell exits the pipe breaks, and the next console.log from the add-on
host raises EPIPE — which Electron reports to the user as a fatal uncaught
exception, over a diagnostic line nobody was left to read.
Testing 0.3.0 on a real install showed the gap immediately: select a run you
had replaced and the selection bar offered duplicate and delete and nothing
else. The one mark made entirely of words was the one with no way to change
them, double-clicking it did nothing, and the size stepper and the bold and
italic buttons all stayed hidden. Everything else that holds text could be
reopened; this could not.
The cause was three places testing `kind === "text"` where the question was
really "does this mark hold words". A replacement holds words.
Reusing the dialog exposed a second, quieter fault. A run lifted out of a
document is whatever size the document set — 11 pt, 9.5 pt — while the size
picker lists round numbers. Selecting a value the list does not contain leaves
the select empty, and the size on the way out fell back to 12. Editing the
wording of an 11 pt line would silently have resized it. The dialog now adds
the document's own size as an option for as long as it is open, and falls back
to the size it started with rather than to a guess.
A PDF does not contain paragraphs. It contains glyphs with coordinates, and
there is no heading, no list, no table and no guaranteed reading order —
only runs of characters that happen to sit near each other. Converting to
Word means working out where the paragraphs were, from geometry. That
inference is the whole feature, and it is sometimes wrong, so this is called
a conversion and never an edit, and the dialog reports what it found before
anything is written.
Lines are grouped by baseline, runs joined with the spaces a PDF only implies
by leaving a gap, and paragraphs ended where the next line sits unusually far
below, is indented, or where the previous one stopped short of the measure.
Headings come from size relative to the body — which is the most common size
on the page, not the average, because a page of 11 pt under a 28 pt title
averages to something that is neither. Bold and italic come from the font's
name, the only place a PDF records them.
What it refuses to fake is as important. A page set in columns is reported,
not silently interleaved. A page with no text says so, and says why: it is an
image of writing, and reading that needs character recognition this editor
does not have. Tables become plain paragraphs rather than an invented grid,
because a wrong table is harder to repair than no table.
The .docx is written here rather than by a vendored builder: a Word file is a
zip of five XML parts, and the subset that can honestly be produced —
paragraphs of styled runs — is about two hundred lines. Vendoring a document
library would have added another megabyte on top of the four pdf.js and
pdf-lib already weigh, to generate markup we would still have to get right.
Entries are stored rather than deflated, which keeps a compressor out of the
add-on; the CRCs are the part that cannot be skipped, since Word calls the
file corrupt rather than naming the part that upset it.
Text replaced in place converts as replaced. Converting would otherwise hand
back the words the user had just edited away.
Checked by taking the output apart — every CRC verified, both XML parts run
through a real parser — and then, because that is still marking my own
homework, by opening the result in the Word editor extension, where mammoth
reads it with none of my code involved.
Aegis relaunches Theseus after staging its own update; seen twice in a
dev instance, the whole browser restarted with no warning, mid-session.
The add-on API's restartApp now asks the user in a native dialog
("Aegis Wallet wants to restart Theseus" — Restart now / Later, Later
is the default) and resolves { restarted, deferred }. Declining loses
nothing: a staged update applies on the next normal launch. The guard
lives in the host, so it covers every Aegis version on the channel and
any future add-on.
Until now "editing" a PDF here meant laying things over it. You could put a
word on top of a word, but the document underneath never changed, and the
result read like a sticker because it was one. This adds the thing the word
Edit actually promises: click a line of the document's text, type different
words, and they land where the old ones were, in the old size and the old
colour.
The position and size come from pdf.js's text layer, which has already placed
a span over every run and carries that run's size in unscaled PDF points — so
the size is right whatever the zoom, which reading it off the rendered box
would not be. The colours come from the rendered page, because nothing in the
text API reports them: the background is the average of the most common colour
bucket in the run's box, since type is a minority of the pixels even when it
is dense, and the ink is whatever sits furthest from that background. On the
test fixture it recovers the marker's red exactly.
Two things that look like details and are not. The bucket only chooses WHICH
pixels are background; the colour itself is their average, because rebuilding
it from the bucket index rounds white down to #f8f8f8 and a not-quite-white
patch on a white page is a visible seam. And the cover reaches below the
baseline by a quarter of the font size, because pdf.js sizes its spans to the
em box: cut the cover to the span and every descender in the original line
survives as a little hook under the replacement.
A replacement is a cover plus text, so it is a mark like any other — movable,
resizable, undoable, and rendered on screen from the same numbers the writer
uses, which is what makes the preview trustworthy.
Said plainly in the dialog and again in the save summary: this hides the
original, it does not remove it. The old glyphs are still in the content
stream underneath. Redact is the tool that takes text away, and it says so
too.
Drawing a rectangle left it selected with its handles showing, and then
refused to let you touch them. The handles were only live under the select
tool, so sizing the shape you were still looking at meant a trip to the
toolbar and back — for a gesture the editor had already drawn the grips for.
Handles are now grabbable whatever tool is armed. They cannot be confused
with drawing: a handle is a nine-pixel square that only exists while
something is selected, nobody lands on one by accident, and Escape drops the
selection if the space is wanted back for drawing. The original gate was
protecting against a collision that does not really happen, at the cost of
one that does.
Making them universal opened a trap in the release path, fixed here too: the
text-markup tools return early from onUp to commit a text selection, which
would have stranded a resize half-done — applied to the live mark, never
journalled, with the drag still set so the next press behaved oddly. A
handle drag is now finished first, whatever tool is armed.
Updates published after launch never showed: the add-on channel was
checked once, 30 s after boot, and staged copies only applied on the
next launch with nothing telling the user. On 2026-09-22 Aegis 0.8.3
and VPN 0.1.3 landed minutes after the app's only check and stayed
invisible through manual scans made earlier and a restart made before
they were published.
Now: the check repeats every 4 hours; every check (boot, timer, manual,
add-on-driven) reports what is staged to the chrome, which shows a chip
for staged extensions; clicking it, or the "Update to vX" button that
appears on the extension's row and detail in Settings, promotes the
staged folder over the installed one and rebuilds the add-on host, so
the new version runs without a restart. Plug-ins (Aegis) are excluded
from the chip and the hot swap — a wallet updates from its own panel
and applies on the next launch.
Also from the same review: the new-tab button follows the last tab and
parks after the scroll arrow only when the strip overflows; Tor sits
left of the Aegis chip; plug-ins no longer appear in Settings ›
Extensions (they have Plug-ins); the extension detail view has a
labelled Back button, a close button and a Check now button; the
promotion helper returns what it promoted and accepts a filter.
Addon:
- Paste any https:// URL that returns a list of vless:// (either
newline-separated or base64) and the extension fetches, decodes,
parses, and adds every server to the dropdown. The full vless URL
never leaves the panel — the addon holds it in its own storage and
passes an opaque "sub-<hash>" id back for selection.
- Subscription CRUD on the addon side (listSubscriptions,
addSubscription, refreshSubscription, removeSubscription). A refresh
is a no-op inside the 6-hour TTL to avoid pounding the provider.
- Merges subscription servers with the baked-in three and gateway
overlay by id; the dropdown groups them under one banner.
Site:
- silentmode.st/vpn landing page: three-plan grid (Free, Pro at $1/mo
BCH, Max at $4/mo BCH), how-it-works four-step block, "the three
servers" strip with per-tier availability, why-this-VPN cards, FAQ.
Priced in USD, paid in BCH via the oracle at pay-time — same pattern
as the marketplace's USD-listing covenant, no reintroduction of
fiat/card processors.
The Extensions list was a stack of tall cards — description, author,
capabilities and buttons on every one — so seven add-ons filled the
page before the user found the toggle. Rows are now one line each,
Firefox-style: icon, name, built-in badge, version, a short update
status, the on/off switch and a ⋯ menu, grouped Enabled / Disabled /
Failed to load. Clicking a row opens the detail view in place: back
arrow, description, update status, author, version, type, folder with
Show folder, and a Permissions block that explains each declared
capability in plain words. The ⋯ menu (and right-click) offers Turn
on/off, Details, Show folder and, for non-bundled extensions, Remove —
a new addons-remove IPC that deletes the folder under the extensions
directory, refuses bundled add-ons (they would only be reseeded), and
clears the dock prefs it left behind.
The lock still said 0.3.51 after the version bump; npm rewrote the two
version fields on the next install. No dependency changed — the diff is
those two lines.
Every commercial VPN client stores its server catalog as a JSON on the
backend and lets the panel pick from a dropdown; this pulls that shape
into the extension.
- server-list.json: baked-in default the tarball ships with. Three
Silent Mode slots (sm-1..sm-3), status "coming-soon" until the VLESS
URLs land — the toggle stays disabled for any entry whose status is
not "ready", so a placeholder cannot be selected by accident.
- Gateway overlay: index.js fetches
https://navigate.st/api/vpn/servers on activation (with a 6-hour TTL
and a "refresh" button in the panel) and merges by id — remote wins,
new remote entries append. Cached to per-addon storage so an offline
boot still has the last-good catalog.
- turnOn now accepts { serverId } or { vless }. Server id is resolved
through the catalog inside the addon; the panel only sees a public
view (label, flag, country, ready/coming-soon), never the raw URL.
- Panel: dropdown of servers + a "Custom vless://" option that reveals
the paste box. Selection persists per-machine, refresh button forces
a re-fetch, disabled toggle explains why in the hint area.
No behavioural change for anyone with a saved vless:// paste — that
path is now "Custom" in the dropdown and still works identically.
Since 0.1.1: Save as… with a real file dialog, PDF export through Chromium's
print pipeline, several documents open at once as tabs with their own close
buttons and menus, a page that scales to fill the window, and Ubuntu and
Fraunces bundled so the ribbon can offer fonts Windows does not have.
The description gains the parts a user would look for before installing.
The page sat marooned in the middle of a wide window with dark space either
side of it. It is now drawn at the size the document actually claims — A4
stays A4, margins come from its own sectPr — and CSS `zoom` scales that to
fit, defaulting to Fit width with a control in the footer and Ctrl +/-/0.
Scaling rather than widening is deliberate. A page stretched to the window
would break every line somewhere different from where the printed page
breaks it, and an editor whose whole claim is that it shows you the document
should not lie about where the lines end. `zoom` also beats a transform
here: it affects layout, so the board scrolls correctly and ProseMirror's
coordinate maths keeps working.
Ubuntu and Fraunces now ship in fonts/, because Windows has neither and a
font offered in the ribbon that the machine lacks is a font the user picks
and then cannot see. Fetched once by `npm run fonts` and committed, never at
runtime: an extension in a browser built around not phoning home should not
ask a font CDN what a document looks like every time one is opened.
Two things had to be worked around. On file:// Chromium registers @font-face
rules and then refuses to fetch the files — the family appears in
document.fonts and every glyph still renders in the fallback — so the add-on
reads the woff2 and hands the page a stylesheet with them inlined as data
URLs. The PDF export needed the same treatment for a different reason: its
print window runs from a temp folder, where a relative url() resolves to
nothing, which would have quietly undone the one-stylesheet-for-both promise
that lib/doc-css.js exists to keep. If either path fails, the ribbon labels
those families "(not available)" rather than implying otherwise.
About 700 KB, most of it Ubuntu's Cyrillic and Greek — kept because the
documents this is used on are not all English. Licences ship alongside.
Opening a second .docx used to mean a second browser tab: a whole ribbon,
banner and footer repeated, with one ✕ at the far end of a row that also
held the file's name. The name looked like a tab and nothing about it
behaved like one.
Now the editor holds documents the way the browser holds pages. A strip
under the toolbar carries one tab per open document — icon, name, unsaved
dot, its own ✕ — plus a + to open another. Middle-click closes, Ctrl+W
closes, Ctrl+Tab cycles, and right-click (or the caret on the tab under the
pointer) drops a menu: Duplicate, Open in the default app, Show in folder,
Close others, Close. The gestures are the browser's because that is the tab
strip every user of this editor already knows.
Under it, one ProseMirror view is handed a different state per document
rather than one view per tab, and the module-level "current document"
variables are marshalled in and out on a switch. That keeps the change out
of every function that touches the current document, at the price of one
list — DOC_FIELDS in captureActive/adoptDoc — that has to stay complete. A
variable missed there leaks one document's state into another, which would
look like the editor corrupting a file, so it is called out in a comment.
Closing the last document closes the editor tab, the way closing a
browser's last tab closes the window; an empty ribbon staring at the user
is not a state worth having.
The add-on hands a newly opened document to the editor that is already up
and fronts it, falling back to opening a tab if no editor acknowledges
within 900ms — so a crashed or closed editor degrades to exactly the old
behaviour rather than swallowing the document.
api.emit() only ever reached the sidebar, even though addon-tab-preload has
always exposed silentmode.on() — so an add-on could hear from its panel but
had no way to tell its own editor tab anything. An add-on that wants a
second document to land in the editor already open, instead of a third tab
full of ribbon, needs exactly that.
Two small pieces:
- emitToPanel now also delivers to every tab owned by the same add-on. The
sidebar keeps its existing condition; tabs are additional, not instead.
- addon-tab-focus, the counterpart of addon-tab-close: a page asking for its
own tab to be fronted. Needed because the click that hands an open page
something new usually happens somewhere else — the sidebar — and the
result would otherwise appear in a tab nobody is looking at.
Both derive the tab from the sender's webContents, the way the close handler
already does, so a page can only front or close the tab it is itself in.
The extension buttons sat in registration order with no way to change
it, hide one, or switch an add-on off without opening Settings. Buttons
are now draggable (drop side follows the pointer, same feedback as tabs
and bookmark chips) and the order is persisted per profile. Right-click
opens a native menu: open/close the panel, move left/right, hide from
the toolbar, turn the add-on off, and Manage extensions; the dock's own
right-click offers Show hidden. Main owns the prefs (dockOrder,
dockHidden) and the actions, so the chrome only renders. Add-on
rediscovery now pushes a fresh dock state to the chrome, which it never
did before — turning an add-on off or installing one from a page
updates the toolbar at once.
Opening a PDF cost four clicks across two screens: the dock icon, a dropdown
with a single entry in it, an editor with an empty drop zone, and finally the
Open button that produced a file browser. Three of those were the program
asking the user to confirm what they had already said by clicking a PDF icon.
The dock item now raises the file browser itself. The dialog is native and
raised from the add-on's Node side, because a file:// page cannot open one
without a user gesture of its own and a freshly-opened tab has none to spend.
The chosen file goes through the same scratch handoff a right-clicked link
already used, so the editor opens with the document in it rather than with an
invitation to find one. Cancelling still lands in the empty editor, which is
where drag-and-drop works and is probably still where someone who changed
their mind about the file wants to be.
Second click removed in the chrome: a toolbar menu holding exactly one item is
not a menu, it is a button wearing a dropdown, so it now dispatches directly.
Falls through to the popup if the renderer has no addonMenuSelect.
Not yet exercised end to end — TheseusNavigator/node_modules is empty while
another session reinstalls it, so the CDP suites cannot boot Electron. Both
files parse; run scratchpad/verify-pdf-editor/drive.mjs once the tree is back.
The Tor chip was the widest control in the toolbar and competed with
the address bar for space. It now sits on the favorites row beside the
Aegis chip, under the Theseus menu, at the same chip height — the
network-level switch next to the wallet, both always visible. State
tints (off / connecting / on) are unchanged.
Multiple signing keys already work — PUBKEYS_HEX is an array and
verifySignature accepts any entry that verifies, which is an allow-list
already. What it isn't is scoped: every add-on is checked against the same
array, so a key added for one extension can sign an update for any other.
The canonical message binds a signature to an extension; nothing binds an
extension to a signer.
The publisher path doesn't have that problem, because a name answers "who
may sign this" per extension and rotates on chain without an app release.
So the note now argues for generalising that rather than replacing it:
per-extension key bindings, several acceptable keys each, the chain
preferred over a compiled-in array, and co-signing considered only for the
default set — the extensions that reach every fresh profile without anyone
choosing them.
Records the open questions honestly too: where a binding can live without
being restatable by whoever controls the manifest, and that refusing to
update an extension is a much smaller decision than refusing to run it.
Aegis sat among the extension buttons and drifted as add-ons were
installed or removed. First-class Silent Mode components (manifest
category "plugin") now get a labelled chip at the right end of the
favorites row, directly below the Theseus menu — always visible, never
collapsed into the extensions overflow. The add-on host tags each
sidebar panel and toolbar menu with the category so the chrome can
split the two docks; the extensions dock keeps its behaviour.
Picks the mark from 95c93f9 over the one I drew: at the 16–18px the dock
actually renders, a solid fill reads and an outlined page does not, and
pairing with pdf-editor's red badge makes the two editors obviously a set.
It moves into icon.svg rather than living only as a data URL in the
manifest, so make-icons.mjs keeps deriving the PNGs and addon.json's copy
from one drawing. Same picture as before, one source instead of three
places to forget.
Binaries uploaded to bns/theseus/vpn-binaries/<platform>/, served
through the gateway at navigate.st/bns/theseus.x/vpn-binaries/.
Manifest hashes match a curl-fetched copy through the gateway.
The mark now shares pdf-editor's silhouette, corner geometry and weight, in
blue against its red, so the two editors read as a pair in the dock. It
carries a pilcrow rather than a format label: at 16px "DOC" is a smudge
while the paragraph mark is still a glyph, and it says word processor rather
than file extension.
Drawn as paths, not <text>. An icon that needs a particular font installed
to make sense is an icon that eventually renders wrong somewhere.
Still nothing of Microsoft's: their blue sheet carries a white W, and the
page-with-a-folded-corner is the universal document glyph rather than
anyone's property.
Also adds docs/DESIGN-one-extension-catalogue.md, which writes down the
larger point this kept bumping into — that bundled and community extensions
are two systems for one kind of thing, with two trust rules and two lists in
Settings, and that "ships by default" should be a line of configuration
rather than a separate distribution path. Proposed only; the awkward part is
keeping a fresh offline install usable, which the note answers by treating
the shipped copies as a pre-seeded cache that still verifies like everything
else.
Ships the extension small (~50 KB tarball). No binaries in it — the
platform-matched sing-box is downloaded on first "Turn on" from
bns/theseus.x/vpn-binaries/<platform>/, sha256-verified against the
manifest that ships inside this operator-signed tarball, and cached
under extensions-data/vpn/bin/. Every subsequent launch re-verifies
before spawning; a mismatch redownloads rather than trusts what is on
disk.
Config generator produces a sing-box config from a vless:// URL (the
shape a 3x-UI VLESS+Reality inbound produces), plus a SOCKS5 inbound
on 127.0.0.1:<ephemeral>. api.setSessionProxy points every Theseus
request at that port while the tunnel is up; child.on("exit") clears
it if sing-box dies. Off again clears the proxy back to whatever the
browser had.
Panel is a big on/off toggle with a status pill, a paste-and-save
endpoint box, and an Advanced disclosure with "auto-on at browser
start", "re-download binary" and "clear cache". Any user with a
vless:// URL can flip it on today; the free tier and the Silent Mode
exit inbound are the server-side half, documented under DESIGN.md.
Binary manifest ships with PENDING sha256s until the binaries are
uploaded to Sia — ensureBinary refuses to activate on a platform whose
sha256 is PENDING, so a user cannot flip it on against an unverified
download.
0.3.50's relocateProfile checked for %APPDATA%\Theseus Navigator\, but
Electron's userData path is derived from app.getName(), which reads
package.json's top-level "name" ("theseus-navigator") because there is
no top-level productName — the "productName": "Theseus Navigator" in
this file lives under "build", where electron-builder reads it for the
installer, not where Electron reads it for the runtime path. So the
folder the user's Theseus writes to is %APPDATA%\theseus-navigator\,
never %APPDATA%\Theseus Navigator\.
On 0.3.50 that meant relocateProfile found nothing at its search path,
returned the new Theseus\ location, and Electron happily created a
fresh empty profile there. The user's addons, vault, bookmarks and
settings stayed in theseus-navigator\ but the running Theseus was no
longer looking at them. Losing the vault is not something the user
can recover from.
Check both candidate names — the one the code was written for and the
one that actually exists — and migrate whichever is present. If the
new Theseus\ already exists (Windows fresh installs after 0.3.51), we
leave it alone.
Moving it out of the build left it with no way in. Settings can only install
from the community catalogue, so a first-party extension that isn't bundled
has a working update channel and no first copy for anyone to update — the
mechanism was all there and the front door was missing.
So it goes back beside screenshot, aegis and pdf-editor: seeded into every
profile by the build, listed under "Built into Theseus", and kept current
between releases by the operator-signed channel at
theseus.x/extensions/docx-editor/. That is the arrangement docs/ADDON-UPDATES.md
describes, and the one the signing script was written for.
About 400 KB compressed in the installer, most of it the vendored editor
libraries — next to the ~4 MB of pdf.js that pdf-editor already ships, the
weight argument for keeping it out didn't survive contact with the numbers.
The end-to-end driver goes back to checking that a fresh profile seeds it,
which is the property that actually matters now.
Extensions read as two kinds today, and the division is not in the engine:
they already share one directory under the profile, one discovery pass and one
enable/disable switch. The split is imposed in four places, and this writes
down where and what each costs to remove.
The load-bearing one is that removal cannot stick. Seeding copies back any
pre-installed extension whose folder is missing, so deleting one is undone on
the next launch — which is why the UI marks them unremovable, and why there is
no uninstall action for anything at all. A tombstone list fixes both.
The urgent one is that the two publishing channels have one install path
between them. Everything installs through the community catalogue; the
operator channel has no index and no install route. So the Install buttons on
the shipped extensions' cards do not work, and moving an extension off the
community channel makes it uninstallable rather than better distributed.
Also records what not to flatten: one catalogue and one lifecycle, but the
signer stays visible, because an extension signed by the operator key and one
signed by a stranger's wallet get the same access once installed.
This is our own extension, so it belongs on the channel the operator key
signs — theseus.x/extensions/docx-editor/ — alongside screenshot, aegis and
pdf-editor. The community catalogue is for extensions published by whoever
owns a BNS name, and routing ours through it would have meant asking a name
owner to vouch for code we wrote.
Drops the community packing script: with one channel there is one publish
path, and it is the one already written down in docs/ADDON-UPDATES.md.
Worth stating plainly, because it is currently true and awkward: Settings
can only *install* from the community catalogue. An extension that is
neither bundled nor catalogued has a working update channel and no way for
anyone to get the first copy. Either it goes back into the build or it needs
a first-party entry point.
Cuts the shipping trigger for the changes accumulated since 0.3.49:
translate is bundled (so a fresh install has it out of the box, and
seedBundledAddons reseeds the folder on any install that lost it),
Settings > Extensions no longer double-renders installed extensions
in a separate Community section, and the PDF/DOC data-URI icons the
dock now paints as <img>s ship alongside their addons.
Three gaps, one theme: the editor could produce a file but not decide where
it went, what format it was in, or look like anything in the dock.
**Save as…** opens a real file dialog, and the extension typed there picks
the format. Save then writes to that file instead of dropping another copy
in Downloads every time. The renderer never names a path: the dialog returns
an opaque token, and the add-on will only write to a path a dialog actually
returned. An extension page is the least trusted thing in the add-on, and
"write these bytes anywhere" is not a capability it needs.
**PDF** goes through Chromium's own print pipeline in a hidden window — the
same engine as Ctrl+P — on the paper size read out of the document's own
sectPr. For that to match what the user was looking at, the page's
typography had to stop living in editor.css, which the export window can't
reach: it moves to lib/doc-css.js and both surfaces read the one string. The
result embeds subsetted fonts, keeps images, and turns hyperlinks into real
PDF link annotations.
**The icon** is ours. Microsoft's Word mark is a trademark and borrowing it
to look official is not something a browser that talks about sovereignty
should do. icon.svg says "text document" in its own words — a turned corner,
a heading rule, body lines, a pilcrow badge in Silent Mode green — and
`npm run icons` derives the PNGs and addon.json's copy from it, so there is
one drawing rather than several that drift.
Also: the scratch folder follows the profile rename to extensions-data/ via
the api.dataDir the host now provides, instead of creating a stale
addons-data/ beside it.
The panel was doing the HTTP call itself, which meant any mirror sitting
behind a Cloudflare-style anti-bot check returned "<!doctype html>…" for
a POST from Origin: file:// and JSON.parse choked on it. Moving the
fetch into index.js gets rid of the whole class of browser-context
interceptors (CORS preflights, captive portals, anti-bot pages) and
lets the addon look at the response body before trying to parse it —
an HTML body is now reported cleanly as "server returned an HTML page
instead of JSON".
While there, chain a small mirror list — translate.disroot.org,
translate.plausibility.cloud, lingva.ml — so a single mirror being down
does not take the feature with it. A user whose saved URL points at a
mirror that stopped resolving (translate.argosopentech.com is the
notable case) now transparently gets a translation from the next mirror
in line instead of a stack trace.
Verified end-to-end from Node against both a working URL and a dead
one; the dead one falls through to disroot as expected.
Discovery of new community extensions belongs on theseus.x/extensions,
which has room to describe each one and lists them by publisher. Settings
just shows what's installed on this machine — one card per extension,
whether bundled or community. The Install button on the web page invokes
the same addons-install-community IPC, and Theseus refreshes the
installed list after each install, so the two lists were mirroring the
same state with a docx-editor card appearing on both sides.
Removes the Community h2, its list container, the Refresh catalog
button, loadCommunity() and its wiring — kept only the trust warning as
a note directly under the installed list.
Everything the editor put on a page was final. A text stamp could not be
corrected without deleting it and typing it again, nothing could be resized,
and the only way to remove a mark was a Delete key nobody had been told
about — the selection drew a dashed box and offered no action at all. Placing
a stamp also left its tool armed, so the next click stamped a second copy.
Marks are now editable objects. Selecting one gives it grab handles and a
small bar pinned above it: delete and duplicate for anything, and for text an
edit button, a size stepper and bold and italic. Double-clicking text reopens
it for rewriting in place rather than adding a second one. Placing a text
stamp or a signature drops straight back to the select tool with the new mark
live, which is both what people expect and what puts it immediately within
reach of a nudge.
Resizing is one function over every mark type rather than a special case per
kind: a handle drag produces a new bounding box, and the mark is mapped from
its old box into that one. Text scales by font size instead of stretching its
glyphs, signatures keep their aspect on a corner, and lines offer their two
endpoints instead of a box that would let you stretch them in ways you never
aimed at. A whole gesture lands on the undo stack as one step.
Selecting a thin mark used to mean clicking its outline exactly — about one
screen pixel. Each stroked mark now carries an invisible fat copy of itself
purely to catch the pointer.
New marks to go with it: underline and strike-through, which share the
highlight's text-selection geometry and differ only in where the rule sits; a
plain line; and a fill toggle for rectangles and ellipses. Bold and italic
mean three more Helvetica variants embedded at save time, since a PDF treats
them as separate fonts rather than as a style.
Double-click is detected from the pointer stream rather than from a dblclick
listener, because selecting a mark calls preventDefault() on the pointerdown
and that suppresses the compatibility mouse events the browser would have
synthesised the dblclick from.
- Settings > Extensions: the Community list showed an entry even for
extensions already listed above under Installed, so docx-editor
appeared twice as soon as it landed in the gateway catalog. The
render now filters out any catalog item whose id is installed —
the Installed list handles updates via the poll + "Check for
updates" button, so the Community list only needs to surface
things you don't have. Empty-state copy branches on whether the
catalog is empty vs "everything is installed".
- theseus.x/extensions catalog renderer accepts data:image/... icons
in addition to emoji, so an extension whose manifest.icon is an
SVG data URI (pdf-editor, docx-editor) renders as an <img> instead
of the raw URI text.
- Static PDF Editor entry on that page now uses the same inline SVG
badge the browser dock renders.
- pdf-editor 0.1.1 → 0.1.2: manifest.icon is a data:image/svg+xml red PDF
document badge (dock renders it as <img>). Same on toolbar-menu.icon.
Item-level icons dropped — those go through the native OS menu that
doesn't render data URIs.
- translate 0.1.0 → 0.1.1: default LibreTranslate mirror was
translate.argosopentech.com, which is now a dead domain — panel just
said "Failed" on every request. Switched default to
translate.disroot.org (currently up), added a datalist of known
mirrors, and a one-shot migration off the dead default so existing
installs recover on next load.
- docx-editor 0.1.0 → 0.1.1: manifest.icon is a data:image/svg+xml blue
DOC document badge, index.js no longer overrides it with 📝, and the
panel header uses the same inline SVG. Same rationale as pdf-editor —
every extension was rendering as either 📄 or 📝, so PDF and Word
were visually identical to the Notepad.
The extensions page could only hand out tarballs; installing meant going
to Settings › Extensions › Community and finding the entry again. Pages
now get window.bcnr.installExtension(id) and Theseus intercepts
theseus://extensions/install/<id> links (page clicks, target=_blank and
the address bar). The page only names a catalog id: Theseus fetches the
catalog and package itself, asks in a native dialog the page cannot draw
over, verifies the publisher signature against the name's current owner
and activates the add-on — the same path a Settings install takes. One
prompt at a time; an already-installed version says so instead of
offering a no-op update.
The site shows the button inside Theseus (feature-detected on the
bridge), a "update Theseus" hint on older builds and a download hint in
other browsers.
The profile folder was Electron's default from the product name
("Theseus Navigator") and add-ons lived in addons\ under it. Now:
%APPDATA%\Theseus\extensions\ installed extensions
%APPDATA%\Theseus\extensions-data\ per-extension storage + scratch
%APPDATA%\Theseus\extensions-backups\ replaced copies
%APPDATA%\Theseus\extensions-staged\ staged updates
Both moves are one-time migrations on the first start that finds the old
layout: the profile folder is renamed (same volume, instant) or copied
when a rename is refused, with the old folder left in place in that case;
the four sub-folders are renamed before the extension host first reads
them. Nothing is deleted. THESEUS_USER_DATA still overrides everything.
The host now hands each extension its data folder as api.dataDir; the
Screenshot and PDF editor add-ons used to rebuild the old path from their
own folder for scratch files (so they recreated addons-data\ after the
move) and now use the field, with versions bumped so the bundles reseed.
A .docx editor is a megabyte of vendored library. Bundling it would charge
that to everyone who wanted a browser, including the people who will never
open a Word document in it.
So it leaves the build: out of bundled-addons/, out of extraResources, absent
from a fresh profile. It arrives the way anyone else's extension does —
Settings › Extensions › Community, from the catalogue the gateway builds, and
listed on theseus.x/extensions alongside everything else published there.
That also means it is signed by the owner of a BNS name rather than by the
operator key, which is the right trust story for something that isn't part of
the browser.
`npm run pack` produces the tarball the publish page takes; the signature
needs the publisher name's wallet, so it isn't something the repo can do.
The end-to-end test now installs the extension into a throwaway profile the
way the community installer would, and asserts up front that a fresh profile
doesn't already have it — the bundling is what was being removed, so it is
worth a test that would notice it coming back.
A PDF that needs a signature, a highlight or a page removed currently sends
the user out to a desktop application or, worse, to a web service that wants
the document uploaded first. Both are poor answers for a browser whose point
is that nothing has to leave the machine. This is a full-tab editor that opens
a PDF, marks it up, fills its forms and saves a new copy, entirely locally.
Two engines, vendored rather than installed, because an add-on ships as a
self-contained folder over the signed update channel and nothing runs a
package manager on the way: pdf.js reads and renders, pdf-lib writes. They
share no state. Everything in between lives in PDF user space — points,
origin bottom-left — which is the one coordinate vocabulary both speak, so a
mark survives zooming, rotating and reordering with no conversion table and
save-time needs to know nothing about how a page happened to be displayed.
The page strip is built from pdf.js's PDFPageView components rather than its
PDFViewer, which renders pages in the file's own order and cannot hide,
reorder or individually rotate one — three of the features here. Text layers
are ours and stay attached for every page, drawn or not, because Theseus's
find bar is Chromium's findInPage over the live DOM and a torn-down text layer
is a page Ctrl+F cannot see. Canvases are virtualised; a letter page at 100%
is 3.4 MB of bitmap.
Redaction is the part worth being careful about. A black box over text hides
nothing — the text stays in the content stream and comes straight out of a
copy-paste — so the editor says so in a modal before the tool can be used,
and on save rebuilds each redacted page as an image, which genuinely removes
it. Pages that were not redacted are untouched. Form widgets and links are
kept, since they were never the leak.
Saving never writes over the original: every save reloads the source bytes and
replays the session onto a fresh copy, so a botched save cannot poison the
next one.
Out of scope for this first version: editing the text that is already in the
document, and writing XFA forms back (pdf-lib cannot, so those are fill-and-
print only, and the editor says so on open).
A .docx editor is easy to write badly: read the file into HTML, let someone
edit it, write a fresh document back, and hand them a file that lost its
headers, its page size and half its formatting without ever saying so.
Three things keep this one honest.
The reader doesn't use mammoth's HTML. mammoth's converter is deliberately
semantic, and HTML has nowhere to put a run's colour or a paragraph's line
spacing, so it drops them — and those are controls this editor puts in the
ribbon. Taking its parsed document model instead means what the ribbon offers
is what the file can actually carry. Six properties mammoth's model didn't
keep are added by build-time patches, each asserting its anchor so an upgrade
that moves the code fails the build rather than shipping a lossy reader.
The writer rebuilds the body but carries the rest of the package across:
headers, footers, footnotes, endnotes, the document's own style catalogue,
its theme and its page setup, with relationship ids and content types
re-wired. Word features the editor can't model are still lost, so they are
detected when the file opens and named in a banner before anyone edits.
Tracked changes get their own gate. mammoth renders insertions as ordinary
text and drops deletions, so saving would accept every pending revision
without Word ever asking. Such a document opens read-only until the user
says that is what they want.
Verified over 66 real documents: 65 round-trip with an identical model and a
structurally valid package, the one exception being a 7 MB WMF picture, which
no browser can display and the writer cannot emit. Also driven end to end
through a real Theseus over CDP — sidebar, ribbon, typing, save, reopen.
Adds a new "context-menu-item" capability. Add-ons declare a
"context-menu-items" array in their manifest:
{
"capabilities": ["context-menu-item", ...],
"context-menu-items": [
{ "id": "translate-selection", "label": "Translate selection",
"when": "selectionText", "icon": "🌐" }
]
}
The `when` filter is one of selectionText | linkURL | editable | image
| always. Right-click on a page, and items whose `when` matches the
current context get merged into the native menu after the built-in
Search-for entry, before Back/Forward/Reload. Both context-menu
handlers (main tab area + detached link windows) share the same
merging logic.
Picking an item dispatches "context-menu" to the add-on's onMessage
handler with the full context (selectionText, linkURL, mediaType,
srcURL, pageURL, host). The add-on decides what to do — the
translate add-on stashes the selection to storage and calls
api.revealSidebar("main") which surfaces its own sidebar panel.
api.revealSidebar(panelId) is the paired hook. Ownership is enforced
by the host — an add-on can only reveal panels it registered —
before routing to main's setSidebar path.
Unknown capabilities were already silently dropped by
validateManifest, so older Theseus builds that don't understand
"context-menu-item" just ignore it, and the manifest still loads.
Add-ons that also declare "sidebar-panel" keep working; the new
capability doesn't require it.
This is the wiring that pairs with the translate/ add-on landed in
4498fbb — right-click "Translate selection" is live once this ships.
Adds a bundled add-on `translate` with a sidebar panel + a right-click
"Translate selection" menu item. Two swappable backends:
- LibreTranslate (default) — free MIT engine; the panel's Settings tab
lets the user point at any instance (public or self-hosted) and drop
in an API key if one's required.
- Google (unofficial free endpoint at translate.googleapis.com/
translate_a/single) — no key, wide coverage, but unofficial and
Google can break it any time. Opt-in fallback.
Flow: user selects text on a page, right-clicks -> "Translate
selection". Add-on's context-menu handler stashes the selection under
storage.__pending and calls api.revealSidebar("main"); the panel
loads, drains __pending on first paint, and translates. Ctrl/Cmd+Enter
in the input textarea also translates. Source + target language
choices, browser-language default target, swap button, copy-to-
clipboard on the output, settings gear.
Depends on a new "context-menu-item" capability + api.revealSidebar
hook in addons-host.js / main.js. Those wiring changes are prepared
but not committed here — a parallel session is refactoring the same
functions concurrently, so the safe path is to land translate/ first
and let the wiring go in alongside the next host-facing commit. Until
the wiring lands, the manifest's "context-menu-item" cap is silently
dropped (per validateManifest's unknown-caps policy) and the sidebar
panel + the panel's translation UI still work standalone — the
right-click entry point is what's gated.
Tabs keep a readable minimum width (76px). Once they overflow, the row
scrolls: the earliest tabs slide out on the left and an arrow at each end
moves the row by 60% of its width, disabled at its end of travel and
hidden while everything fits. The wheel scrolls the row too, the selected
tab is brought into view when the selection changes (never while the user
is scrolling), and the + button stays outside the row. Moves are immediate
rather than animated: the chrome view has no smooth scrolling, and frame
callbacks stop while the window is occluded, which stranded a frame-driven
slide at its start.
With many tabs open, the last ones slid under the minimise / maximise /
close overlay: the strip reserved that space as padding, and padding does
not stop overflowing flex items. The tabs now live in their own row that
clips at its own edge, shrink down to icon + close before anything is
hidden, and the row scrolls sideways with the wheel once they hit that
minimum, keeping the selected tab in view. The + button sits outside the
row so it stays reachable no matter how many tabs are open.
