Repository navigation
Releases
omapass 0.2.1
Fixed
- Theming broke after updating to a Quickshell that ships its own built-in
Colorobject — its color-math helpers (rgba,blend,tint, ...) share
the name with theqs.Commons.Colortheming singleton, and the Quickshell
one won. Every bareColor.menu.background,Color.urgent, etc. silently
resolved toundefined, dropping every theme color from the overlay and bar
widget. Importsqs.Commons as Commonsalongside the existing unqualified
import and references the singleton asCommons.Color.*, the same pattern
Omarchy's own first-party plugins already use for this. ([#52])
Install
omarchy plugin add https://github.com/cschaba/omapass.git --enable --yesOr from the tarball below:
tar -xzf omapass-0.2.1.tar.gz
cd omapass-0.2.1 && ./install.shsha256: 5ae7eab06fc8043ad0d5066af8e4a6a7a2747e71a858a9704cc22ce6803ef140
Under development and largely untested — see the warning at the top
of the README before trusting it with anything.
omapass 0.2.0
Changed
-
Version 0.2.0. This is a marker rather than a feature release: nothing in the
code changed between 0.1.50 and here. ([#23])0.1.0 was an overlay that could list a
passstore and copy out of it. Fifty
patch releases later the shape has stopped moving. Two surfaces that stay in
step — a bar pulldown for finding a password, a full manager for changing one
— and entries created, edited, generated, renamed and deleted without leaving
the keyboard. One-time codes, an optional fingerprint gate in front of the
store, a git sync that now says so when it fails, and a debug log that records
what happened without recording what it happened to. Listed and verified on
the Omarchy plugin marketplace since 4 September.What the number claims is narrow: the config file, the CLI surface and the
entry format have settled enough that breaking one of them would be worth a
major bump. It is not a claim that the work is finished. ([#23])
Install
omarchy plugin add https://github.com/cschaba/omapass.git --enable --yesOr from the tarball below:
tar -xzf omapass-0.2.0.tar.gz
cd omapass-0.2.0 && ./install.shsha256: 80b964b19581811349767e0218887693d8619a398b1eec0dc3f4e0a15b1a9bb6
Under development and largely untested — see the warning at the top
of the README before trusting it with anything.
omapass 0.1.50
Removed
lockVault(), which nothing could reach. No key was bound to it and nothing
called it, so the grace window was already the only re-lock — it just did not
look that way, because a smoke test asserted on the function's body and made
it read as live code.Ctrl+Lis fill login, in both the pulldown and the
manager, and always was. Binding it to a key was the alternative and was
decided against: the grace window is the re-lock. ([#51])
Install
omarchy plugin add https://github.com/cschaba/omapass.git --enable --yesOr from the tarball below:
tar -xzf omapass-0.1.50.tar.gz
cd omapass-0.1.50 && ./install.shsha256: 859027204c607bb5221f33182168250b2ddc8c95da36df3a9b1fba6a2797cca1
Under development and largely untested — see the warning at the top
of the README before trusting it with anything.
omapass 0.1.49
Fixed
-
Authenticating at the bar pulldown now counts at the manager too. The two
are separate surfaces and each kept its own record of the scan, so passing
the gate in the pulldown left the manager locked — andCtrl+Ethere does
not open an editor in the pulldown, it summons the manager. Being asked a
second time, a second later, for the same store on the same desktop, reads
as the first scan having failed. ([#50])The unlock now lives in one place for the whole plugin, and the grace window
runs from whenever the last surface closed rather than per window. Sharing it
is not a weakening: the grace already allowed a surface to be closed and
reopened without a second scan, so "recently authenticated counts" was
already the policy — it was only ever scoped to whichever window happened to
ask.fingerprint-grace = 0still means scan every time. The state is held
in memory in the shell process and is never written to disk. ([#50]) -
The release tarball now includes the QML module directory the plugin imports.
It is assembled from an explicit list, and a directory missing from that list
works perfectly from a git checkout while failing for everyone who installs
the release. A test now holds that list to what the QML actually imports.
([#50])
Install
omarchy plugin add https://github.com/cschaba/omapass.git --enable --yesOr from the tarball below:
tar -xzf omapass-0.1.49.tar.gz
cd omapass-0.1.49 && ./install.shsha256: c379fad721e1c75250231e9cd6894eb6b03e9074bdbc5f2ac571b11f72d7dbf6
Under development and largely untested — see the warning at the top
of the README before trusting it with anything.
omapass 0.1.48
Added
- New
auto-syncsetting (defaultoff): whenon, OmaPass runs a sync
right after every insert, edit or delete, instead of only onCtrl+S. Off
by default because it adds a network round trip — and its possible failure
— to every single edit. (#46)
Fixed
-
Ctrl+S (sync) no longer leaves the entry list stale once the pull/push
finishes.syncwas fired through the same detached helper used for
clipboard/type actions, so nothing ever reloaded the list or surfaced a
failure when it was done; it is now a tracked process like every other
write, and reloads the list (or reports the error) on exit. (#45) -
syncno longer reports success when git failed.pass(1)ends inexit 0
and its git handler never checks the git it just ran, sopass git push || diecould not fire however badly the push went — an unreachable remote, a
rejected update and a store with no remote at all all printed "Store synced"
and exited 0, with2>/dev/nulldiscarding the only remaining evidence.syncnow runs git directly in the store, which is the same invocation
pass gitmakes minus the discarded status, and each call is guarded. A
store with no remote configured says so and names the command that adds one,
rather than failing as though the network were at fault. The wording of each
failure stays ours: git's own stderr names the file it tripped over, and an
entry name is disclosure in its own right. ([#49])This is the only place it bit —
pass show,mv,rmandotpall fail
through pass's owndie, which exits 1 before that finalexit 0is
reached, so their guards work. ([#49])
Install
omarchy plugin add https://github.com/cschaba/omapass.git --enable --yesOr from the tarball below:
tar -xzf omapass-0.1.48.tar.gz
cd omapass-0.1.48 && ./install.shsha256: d5bdfe2f08e9c98dc2f752b5d199791a63bc14829b2f09db35e4446edd5abea3
Under development and largely untested — see the warning at the top
of the README before trusting it with anything.
omapass 0.1.47
Fixed
-
install.shanduninstall.shno longer delete another plugin that happens
to use the bareomapassid. That id was ours until 0.1.12 namespaced it to
cschaba.omapass, so~/.config/omarchy/plugins/omapasswas treated as an
old install of ours and cleared away — and it is another plugin's address
now: tonyrumans/omapass declaresomapassas its plugin id and
installs exactly there. Anyone with both would have lost theirs, silently, to
a message claiming OmaPass had tidied up after itself.Both scripts now read the manifest in that directory before touching it. Only
a manifest that is recognisably ours — the namespaced id, our homepage, or
Omapass.qmlas the overlay entry point, which every version we have
shipped declares — is removed or disabled. Anything else is named and left
alone, and OmaPass installs alongside it under its own id. A directory with
no manifest at all is still ours: it predates the plugin contract.The rule fails closed, because the costs are not symmetric. Leaving a stale
directory of ours behind is untidy; deleting a working install of somebody
else's is not something a password manager gets to do twice. ([#44])
Install
omarchy plugin add https://github.com/cschaba/omapass.git --enable --yesOr from the tarball below:
tar -xzf omapass-0.1.47.tar.gz
cd omapass-0.1.47 && ./install.shsha256: 5b07debbee61f03f3600b2875032d864c0fc989061707433cda0703d9ad98d62
Under development and largely untested — see the warning at the top
of the README before trusting it with anything.
omapass 0.1.46
Added
-
The
?that opens the shortcut sheet is on every surface now, not only the
manager's list. The two places you are most likely to be stuck — halfway
through a form, or in the pulldown — were the two with no way to the sheet
but a key you had to already know.F1works in the form as well. The
pulldown has no room for the sheet, so both the key and the button open the
manager with it already up. ([#32]) -
The
?sits in the top-right corner in the manager and the editor, rather
than on the hint row at the bottom. That row is the busiest line in the app,
and in the editor a third thing to aim at beside Cancel and Save is one too
many. The entry count moves left to make room. The pulldown keeps its?at
the bottom, next toManage…, because its top row is the search field and
its bottom row is short. ([#32])Escis the reflex that closes a help window, and in the editor it is also
the key that discards the draft. It only means the first of those while the
sheet is up. ([#32])
Changed
scripts/release.shasksgit ls-remotewheremainpoints instead of
runninggit fetch. It only ever wanted the answer to "is this checkout
behind?", and it now gets it without pulling anything into the repository.
Thatgit fetchwas the marketplace scanner's one finding against omapass
(remote-git-execution-unpinned), and a finding is fail-closed there — no
maintainer can accept one, which is why the submission sat for a month.
Fixed
-
Three rows on the shortcut sheet were not true.
Ctrl+Q(scan a QR code) and
Ctrl+S(sync) are bound in the manager and not in the bar pulldown, but
neither said manager only, so the sheet sent you to the pulldown to press
a key that was never there. And the sheet's own fallback for the open hotkey
still readSuper + Shift + K, the chord that changed in 0.1.45. Two tests
now hold the sheet to the code: every key the pulldown does not bind has to
be marked, and the fallback chord is derived fromlib/config.shrather than
typed. ([#32], [#41]) -
The shortcut sheet is readable again. It had no background of its own, over a
list that nothing had told to get out of the way, so the two sets of text sat
on top of each other. It paints its own background now and covers the card
corner to corner — a view on top of the others, not a panel floating in one —
and everything underneath goes dark while it is up. Those panes are hidden in
place rather than removed, because aColumnlays out only its visible
children and dropping them would have pulled the footer to the top of the
card. ([#32])
Install
omarchy plugin add https://github.com/cschaba/omapass.git --enable --yesOr from the tarball below:
tar -xzf omapass-0.1.46.tar.gz
cd omapass-0.1.46 && ./install.shsha256: 7f83e416bb0eacfe201922096003df13aeff101e735b414bc084fdc432df5d9f
Under development and largely untested — see the warning at the top
of the README before trusting it with anything.
omapass 0.1.45
Changed
- The default hotkey is
SUPER + ALT + P, wasSUPER + SHIFT + K. Checked
against a live Omarchy install rather than picked by feel:SUPER + P,
SUPER + SHIFT + PandSUPER + CTRL + Pare all taken, andSUPER + ALT
is busy with 25 window and group bindings but not withP. Nothing breaks
for anyone already installed — OmaPass has never written tobindings.lua,
it prints the line for you to paste, so an existing binding keeps working.
Re-runninginstall.shnow says which chord it actually found and offers the
new line rather than reporting "already there" and hiding the change. ([#41])
Install
omarchy plugin add https://github.com/cschaba/omapass.git --enable --yesOr from the tarball below:
tar -xzf omapass-0.1.45.tar.gz
cd omapass-0.1.45 && ./install.shsha256: 42ed7b4b84fc36d768adbddb1219137d4a831475b20fd49f473d1262268f0e1d
Under development and largely untested — see the warning at the top
of the README before trusting it with anything.
omapass 0.1.44
Added
- A shortcut sheet, on
F1or the?in the bottom-right corner. OmaPass is
keyboard-first, which is only a virtue while you can find out what the keys
are — and the footer hints can only ever carry the handful that fit. The
sheet covers both surfaces at once, says which rows are manager-only, and
lists only what your machine can do: no one-time-code rows without
pass-otp, no sync row without a git store. It also spells your configured
hotkey the way a person says it rather than the waykeybindwrites it.
F1in the bar pulldown opens the manager showing it. ([#32])
Changed
F1is the shortcut sheet rather than the About screen, which is the
convention everywhere else. About moved onto the sheet, so it is one keypress
further away and no harder to find. ([#32])
Install
omarchy plugin add https://github.com/cschaba/omapass.git --enable --yesOr from the tarball below:
tar -xzf omapass-0.1.44.tar.gz
cd omapass-0.1.44 && ./install.shsha256: 1478ccc8ecc2207940df24f88b26ebd43efa21d426f9eaa54aba0459e21cccb8
Under development and largely untested — see the warning at the top
of the README before trusting it with anything.
omapass 0.1.43
Fixed
- The OTP field accepts what sites and other password managers actually hand
out. A bare secret is enough — spaces and lower case included — and a URI
missing its label, such asotpauth://totp?secret=…, is repaired rather than
refused. Both are then stored as a completeotpauth://line and shown in
the field before saving. This is not just a looser check:pass-otp's parser
makes the label optional to match and then dies on the missing account name,
so the URI in the report would have been unreadable by the very command that
has to read it.Ctrl+QQR enrolment refuses such a code up front instead of
storing one that cannot be used. A word shorter than 16 characters is never
taken for a secret, so a password pasted into the wrong field is not silently
turned into a one-time code. ([#40])
Install
omarchy plugin add https://github.com/cschaba/omapass.git --enable --yesOr from the tarball below:
tar -xzf omapass-0.1.43.tar.gz
cd omapass-0.1.43 && ./install.shsha256: 04ffc4fd92a04a4ab8e018044745db36eb3785a7e89c485f46a796f885c19c9c
Under development and largely untested — see the warning at the top
of the README before trusting it with anything.