Skip to content

Give a clear diagnosis when libfprint has no driver for the detected fingerprint reader - #11842

Open
busbyjon wants to merge 4 commits into
omacom:quattrofrom
busbyjon:fingerprint-no-driver-diagnosis
Open

busbyjon wants to merge 4 commits into
omacom:quattrofrom
busbyjon:fingerprint-no-driver-diagnosis

Conversation

@busbyjon

@busbyjon busbyjon commented Sep 14, 2026 •

Copy link
Copy Markdown

Summary

  • Fixes Fingerprint setup gives no diagnosis when libfprint has no driver for the detected reader #11841: omarchy setup security fingerprint sends readers with no libfprint driver at all (e.g. Goodix TLS-encrypted sensors like 27c6:55b4 on the ThinkPad L13 Yoga Gen 3) through fprintd-enroll and tells the user to "try again" on failure — advice that can never work, since libfprint has nothing to enroll against.
  • Adds an fprintd-list "$USER" check before enrollment. It reliably exits 1 with "No devices available" when libfprint has no driver for the detected reader (verified against fprintd's own source). On that path, print a clear message pointing at the supported-devices list and the Discord instead of attempting enrollment.
  • This generalizes a case the script already anticipated in a comment for Elan MOC sensors outside the elanmoc table — same failure mode, different vendor.

Test plan

  • Added fprintd-list stub and a new case to test/shell.d/fingerprint-package-test.sh covering the zero-driver path
  • bash test/shell.d/fingerprint-package-test.sh — all 6 cases pass
  • bash test/shell (full shell suite) — no new failures; the 5 pre-existing failures (config-test.sh, snapper-test.sh, unowned-system-paths-test.sh, locate-test.sh, bar-widget-contract-test.sh) reproduce identically on a clean checkout with this change stashed out, and are caused by environment setup unrelated to this change (missing OMARCHY_PKGS_PATH checkout, a stray non-UTF8 file, and one flaky run)
  • Reproduced and fixed on real hardware: ThinkPad L13 Yoga Gen 3, Goodix 27c6:55b4

🤖 Generated with Claude Code

https://claude.ai/code/session_01SMHmPqFQQroLFtP5t8Athe

Fixes #10892

omarchy-hw-fingerprint only proves a reader is present, not that
libfprint can drive it. Readers like Goodix TLS-encrypted sensors
(27c6:55b4, found on the ThinkPad L13 Yoga Gen 3) or an Elan MOC sensor
outside the elanmoc table enumerate on USB but have no libfprint driver
at all, so enrollment fails no matter how many times it's retried.

Check fprintd-list before attempting enrollment and bail with a clear
message when it reports zero devices, instead of sending the user
through fprintd-enroll and telling them to try again.

Fixes omacom#11841

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SMHmPqFQQroLFtP5t8Athe
@llstrk

llstrk commented Sep 30, 2026

Copy link
Copy Markdown

Automated AI review

Community review: Independent automated community review, unaffiliated with the Omarchy team, intended to help prepare PRs for their review.

Outcome: The new check reports every fprintd-list failure as "libfprint has no driver", not only the zero-device case.

Any fprintd-list failure is reported as a missing driver

fprintd-list exits 1 for zero devices (No devices available), but also when fprintd cannot be reached (not activatable, crashed, timed out) or polkit denies ListEnrolledFingers outside an active local session. Impact: users hitting those failures (for example over SSH, or when fprintd fails to start) are told the sensor is unsupported and retrying won't help, while &>/dev/null hides the real error.

Suggested change: capture the output and show the missing-driver message only when it contains No devices available (an untranslated literal in list.c); otherwise print that output.

Reproducer (fails at 097eed0)

Reproducer: every fprintd-list failure is reported as "libfprint has no driver"

Test case added to the PR's own test/shell.d/fingerprint-package-test.sh (reviewed head 097eed0). The fprintd-list stub now prints a reason before exiting 1, as the real client does, and a second case feeds it the error a supported reader gets when polkit refuses ListEnrolledFingers:

--- a/test/shell.d/fingerprint-package-test.sh
+++ b/test/shell.d/fingerprint-package-test.sh
@@ -43,6 +43,8 @@
 cat > "$scratch/bin/fprintd-list" <<'STUB'
 #!/bin/bash
 echo list >> "$CALL_LOG"
+# fprintd-list prints its reason on stdout before exiting 1 (utils/list.c).
+[[ ${LIST_STATUS:-0} == 0 ]] || printf '%s\n' "${LIST_OUTPUT:-No devices available}"
 exit "${LIST_STATUS:-0}"
 STUB
 cat > "$scratch/bin/fprintd-enroll" <<'STUB'
@@ -105,3 +107,11 @@
   fail "a reader with no libfprint driver does not attempt enrollment"
 fi
 pass "a reader with no libfprint driver stops before enrollment with a clear message"
+
+# fprintd-list also exits 1 when the reader is supported but the query fails,
+# e.g. polkit refusing ListEnrolledFingers outside an active local session.
+LIST_STATUS=1 LIST_OUTPUT='ListEnrolledFingers failed: GDBus.Error:net.reactivated.Fprint.Error.PermissionDenied: Not Authorized: net.reactivated.fprint.device.verify' run_setup
+if grep -q 'no driver' "$scratch/output"; then
+  fail "a failed fprintd-list query on a supported reader is not reported as a missing driver" "$(cat "$scratch/output")"
+fi
+pass "a failed fprintd-list query on a supported reader is not reported as a missing driver"

Output at the reviewed head (bash test/shell.d/fingerprint-package-test.sh, exit 1; ANSI colour codes removed):

ok - a fresh machine installs libfprint-git and reaches enrollment
ok - installed stock libfprint is replaced without a removal step
ok - a rerun with everything installed goes straight to enrollment
ok - a failed installation stops before enrollment
ok - missing hardware performs no package operations
ok - a reader with no libfprint driver stops before enrollment with a clear message
Setting up fingerprint scanner for authentication.

Installing required packages...
A fingerprint reader was detected, but libfprint has no driver for it.
Retrying enrollment won't help — this exact sensor isn't supported yet.
Check https://fprint.freedesktop.org/supported-devices.html, or ask in the Omarchy Discord (https://omarchy.org/discord).
not ok - a failed fprintd-list query on a supported reader is not reported as a missing driver

On the head merged into quattro 8b4eae6 (keeping quattro's ---aware pacman stub) it fails the same way (exit 1, same not ok line). With the check changed to give the missing-driver diagnosis only when the output says No devices available, the same test file passes (exit 0), and the PR's unmodified test file also still passes:

--- a/bin/omarchy-setup-security-fingerprint
+++ b/bin/omarchy-setup-security-fingerprint
@@ -90,10 +90,15 @@
 # here with a diagnosis instead of sending the user into fprintd-enroll, which
 # will just hang waiting on a reader nothing will ever answer for, then tell
 # them to "try again" when no amount of retrying can succeed.
-if ! fprintd-list "$USER" &>/dev/null; then
-  echo -e "\e[31mA fingerprint reader was detected, but libfprint has no driver for it.\e[0m"
-  echo "Retrying enrollment won't help — this exact sensor isn't supported yet."
-  echo "Check https://fprint.freedesktop.org/supported-devices.html, or ask in the Omarchy Discord (https://omarchy.org/discord)."
+if ! fprintd_list_output=$(fprintd-list "$USER" 2>&1); then
+  if [[ $fprintd_list_output == *"No devices available"* ]]; then
+    echo -e "\e[31mA fingerprint reader was detected, but libfprint has no driver for it.\e[0m"
+    echo "Retrying enrollment won't help — this exact sensor isn't supported yet."
+    echo "Check https://fprint.freedesktop.org/supported-devices.html, or ask in the Omarchy Discord (https://omarchy.org/discord)."
+  else
+    echo -e "\e[31mCould not query the fingerprint service:\e[0m"
+    echo "$fprintd_list_output"
+  fi
   exit 1
 fi
 
ok - a fresh machine installs libfprint-git and reaches enrollment
ok - installed stock libfprint is replaced without a removal step
ok - a rerun with everything installed goes straight to enrollment
ok - a failed installation stops before enrollment
ok - missing hardware performs no package operations
ok - a reader with no libfprint driver stops before enrollment with a clear message
ok - a failed fprintd-list query on a supported reader is not reported as a missing driver

What the real fprintd-list prints in each state

Real /usr/bin/fprintd-list and fprintd-enroll (outputs match the fprintd 1.94.5 strings in utils/list.c and utils/enroll.c), talking over a private bus to a small stand-in that returns the replies fprintd 1.94.5 gives in each state (src/manager.c 538-604, src/device.c 2329-2336 and 846-851). denied is the reply for a non-root caller outside an active local session under the shipped polkit defaults (net.reactivated.fprint.device.verify: allow_any=no, allow_inactive=no). noservice is a bus with no fprintd at all.

nodev     fprintd-list alice  exit=1  No devices available / 
noprints  fprintd-list alice  exit=0  found 1 devices / Device at /net/reactivated/Fprint/Device/0 / Using device /net/reactivated/Fprint/Device/0 / User alice has no fingers enrolled for Mock reader. / 
prints    fprintd-list alice  exit=0  found 1 devices / Device at /net/reactivated/Fprint/Device/0 / Using device /net/reactivated/Fprint/Device/0 / Fingerprints for user alice on Mock reader (press): /  - #0: right-index-finger / 
denied    fprintd-list alice  exit=1  found 1 devices / Device at /net/reactivated/Fprint/Device/0 / Using device /net/reactivated/Fprint/Device/0 / ListEnrolledFingers failed: GDBus.Error:net.reactivated.Fprint.Error.PermissionDenied: Not Authorized: net.reactivated.fprint.device.verify / 
noservice fprintd-list alice  exit=1  Impossible to get devices: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name net.reactivated.Fprint was not provided by any .service files / 
More findings

Code comment says fprintd-enroll "will just hang". With zero devices, GetDefaultDevice returns NoSuchDevice and fprintd-enroll exits 1 at once; issue #11841 also says enrollment "fails immediately". The older comment at lines 110-114 (an Elan MOC sensor "gets this far and then fails to enroll") is now stale too, since such a sensor stops at the new check. Impact: comments only; they misdescribe current behaviour for later readers of the script. Suggested change: reword both comments.

Reproducer (comment: fprintd-enroll exits at once)

Reproducer: fprintd-enroll does not hang when libfprint has no driver

The new code comment says fprintd-enroll "will just hang waiting on a reader nothing will ever answer for". With zero usable readers fprintd answers GetDefaultDevice with NoSuchDevice (fprintd 1.94.5 src/manager.c 575-603) and fprintd-enroll exits 1 at once (utils/enroll.c 73-79).

Script (starts a private bus and a private fprintd; run where no reader is visible, here inside a bubblewrap sandbox with no /sys and a minimal /dev):

#!/bin/bash
# What fprintd-list and fprintd-enroll do when fprintd has zero usable readers
# (the state when libfprint has no driver for the reader). Starts a private
# bus and a private fprintd instance; run it where no reader is visible, for
# example inside bubblewrap with --dev /dev --tmpfs /sys --unshare-all.
set -u
user=${USER:-alice}
dir=$(mktemp -d)
cat > "$dir/bus.conf" <<CONF
<!DOCTYPE busconfig PUBLIC "-//freedesktop//DTD D-BUS Bus Configuration 1.0//EN"
 "http://www.freedesktop.org/standards/dbus/1.0/busconfig.dtd">
<busconfig>
  <listen>unix:path=$dir/socket</listen>
  <auth>EXTERNAL</auth>
  <policy context="default">
    <allow user="*"/>
    <allow own="*"/>
    <allow send_destination="*" eavesdrop="true"/>
    <allow eavesdrop="true"/>
  </policy>
</busconfig>
CONF
dbus-daemon --config-file="$dir/bus.conf" --nofork --nopidfile & bus=$!
for _ in $(seq 100); do [[ -S $dir/socket ]] && break; sleep 0.05; done
export DBUS_SYSTEM_BUS_ADDRESS="unix:path=$dir/socket"
/usr/lib/fprintd 2>/dev/null & daemon=$!
for _ in $(seq 100); do
  gdbus call --system --dest org.freedesktop.DBus --object-path /org/freedesktop/DBus \
    --method org.freedesktop.DBus.NameHasOwner net.reactivated.Fprint 2>/dev/null | grep -q true && break
  sleep 0.05
done
for cmd in fprintd-list fprintd-enroll; do
  echo "\$ $cmd $user"
  start=$EPOCHREALTIME
  timeout 30 "$cmd" "$user"
  status=$?
  printf '[exit %s after %.2fs]\n' "$status" "$(awk -v a="$start" -v b="$EPOCHREALTIME" 'BEGIN{print b-a}')"
done
kill "$daemon" "$bus"
wait 2>/dev/null
rm -rf "$dir"

Actual output (USER=alice, installed fprintd, libfprint 1.94.100 per the daemon log):

$ fprintd-list alice
No devices available
[exit 1 after 0.00s]
$ fprintd-enroll alice
Impossible to enroll: GDBus.Error:net.reactivated.Fprint.Error.NoSuchDevice: No devices available
[exit 1 after 0.00s]

What the unpatched setup script prints on this path today (script at the PR's base commit 31bd80d, real fprintd-enroll against a private-bus stand-in answering NoSuchDevice; exit 1 after 0.01 s; ANSI colour codes removed):

Setting up fingerprint scanner for authentication.


Let's setup your right index finger as the first fingerprint.
Keep moving the finger around on sensor until the process completes.

Impossible to enroll: GDBus.Error:net.reactivated.Fprint.Error.NoSuchDevice: No devices available

Enrollment failed. Please try again.
Details

The exit paths are from fprintd 1.94.5 utils/list.c: bus or Manager proxy failure, GetDevices failure, zero devices, and ListEnrolledFingers failing with anything other than NoEnrolledPrints (polkit action net.reactivated.fprint.device.verify: allow_inactive=no, allow_active=yes).

Overlapping open PRs (maintainer information): #12480 adds a competing busctl ... GetDefaultDevice readiness guard at the same place in the script and a stub at the same place in fingerprint-package-test.sh. Merged head to head they conflict in both files; merged one after the other through quattro, only the test file conflicts and the script silently ends up with both guards back to back. #12833 rewrites the adjacent package-install lines; the script merges cleanly, but both PRs append cases at the end of fingerprint-package-test.sh, so that file conflicts.

Verified:

  • With zero devices (the state libfprint leaves when no driver matches), fprintd-list "$USER" exits 1 with No devices available (installed fprintd on a private bus, no devices visible), and the new block then stops before enrollment (private-bus stand-in and the PR's test).
  • Supported readers, with or without enrolled prints, pass the check and reach enrollment (private-bus stand-in).
  • 27c6:55b4 appears only in the known-unsupported table of the pinned libfprint-git (fprint-list-udev-hwdb.c).
  • The PR merges cleanly into the current quattro tip, and fingerprint-package-test.sh passes 6/6 at the head and on that merge.

Optional test improvement: the new case asserts neither the message nor the "$USER" argument; mutants that delete the message or drop "$USER" both pass all six cases, and the real client without a username exits 1 for every reader. Have the stub log $* and print a reason, then assert both.

Mutation runs (optional test improvement)

Reproducer: the new test case does not check the message or the username argument

Two mutations of the reviewed head script. The PR's own test/shell.d/fingerprint-package-test.sh still passes (exit 0, all six cases) for each.

Message removed

--- a/bin/omarchy-setup-security-fingerprint
+++ b/bin/omarchy-setup-security-fingerprint
@@ -91,9 +91,6 @@
 # will just hang waiting on a reader nothing will ever answer for, then tell
 # them to "try again" when no amount of retrying can succeed.
 if ! fprintd-list "$USER" &>/dev/null; then
-  echo -e "\e[31mA fingerprint reader was detected, but libfprint has no driver for it.\e[0m"
-  echo "Retrying enrollment won't help — this exact sensor isn't supported yet."
-  echo "Check https://fprint.freedesktop.org/supported-devices.html, or ask in the Omarchy Discord (https://omarchy.org/discord)."
   exit 1
 fi
 
ok - a fresh machine installs libfprint-git and reaches enrollment
ok - installed stock libfprint is replaced without a removal step
ok - a rerun with everything installed goes straight to enrollment
ok - a failed installation stops before enrollment
ok - missing hardware performs no package operations
ok - a reader with no libfprint driver stops before enrollment with a clear message

"$USER" dropped from the fprintd-list call

--- a/bin/omarchy-setup-security-fingerprint
+++ b/bin/omarchy-setup-security-fingerprint
@@ -90,7 +90,7 @@
 # here with a diagnosis instead of sending the user into fprintd-enroll, which
 # will just hang waiting on a reader nothing will ever answer for, then tell
 # them to "try again" when no amount of retrying can succeed.
-if ! fprintd-list "$USER" &>/dev/null; then
+if ! fprintd-list &>/dev/null; then
   echo -e "\e[31mA fingerprint reader was detected, but libfprint has no driver for it.\e[0m"
   echo "Retrying enrollment won't help — this exact sensor isn't supported yet."
   echo "Check https://fprint.freedesktop.org/supported-devices.html, or ask in the Omarchy Discord (https://omarchy.org/discord)."
ok - a fresh machine installs libfprint-git and reaches enrollment
ok - installed stock libfprint is replaced without a removal step
ok - a rerun with everything installed goes straight to enrollment
ok - a failed installation stops before enrollment
ok - missing hardware performs no package operations
ok - a reader with no libfprint driver stops before enrollment with a clear message

The real client without a username prints its usage and exits 1 whatever the reader state (fprintd 1.94.5 utils/list.c 156-160), so this mutation would send every user, including those with supported readers, to the "no driver" message:

nodev     fprintd-list  exit=1  Usage: /usr/bin/fprintd-list <username> [usernames...]
noprints  fprintd-list  exit=1  Usage: /usr/bin/fprintd-list <username> [usernames...]
prints    fprintd-list  exit=1  Usage: /usr/bin/fprintd-list <username> [usernames...]
denied    fprintd-list  exit=1  Usage: /usr/bin/fprintd-list <username> [usernames...]
noservice fprintd-list  exit=1  Usage: /usr/bin/fprintd-list <username> [usernames...]

(nodev, noprints, prints, denied: private-bus stand-in in that state; noservice: no fprintd on the bus.)


Review information

Test scope: Source and sandbox tests at 097eed0 (real fprintd clients, private D-Bus); no fingerprint hardware or real polkit session.

AI process: Opus 5.5 Medium coordination and synthesis, Opus 5.5 Xhigh technical review and final fact check, GPT 6 Sol Xhigh search for related issues, Opus 5.5 Medium editorial check.

Opt out: To stop receiving these reviews, reply to this comment saying so.

fprintd-list exits 1 for more than an empty device list: a daemon or D-Bus failure, or a denied verify permission (any session polkit does not treat as local and active, such as SSH). On a supported reader those were all reported as "libfprint has no driver for it". Match the "No devices available" line fprintd prints for an empty list, and leave every other failure to enrollment as before.

fprintd-enroll does not hang without a device: GetDefaultDevice returns NoSuchDevice and it exits at once, so the comment now says that.

The test now checks the message is printed, and that another fprintd-list error still reaches enrollment.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Codex Medium <noreply@openai.com>
@greptile-apps

greptile-apps Bot commented Oct 1, 2026 •

Copy link
Copy Markdown
Contributor

RetriggerView in GreptileConfidence Score: 5/5

[Medium impact] The PR appears safe to merge; the earlier diagnosis concern is addressed.

Summary

Adds a check with fprintd-list before fingerprint enrollment.

  • Fingerprint setup checks for a usable device before enrollment.

No new actionable issues were found.

Reviews (3) · Last reviewed commit: "Say fprintd has no usable device, and th..." · Reviewed by Greptile

Comment thread bin/omarchy-setup-security-fingerprint Resolved
@omarchybot omarchybot added the verified Omarchy Triage has verified that this issue is ready for final review label Oct 1, 2026
@omarchybot

Copy link
Copy Markdown
Collaborator

Reviewed and reproduced, with one fix pushed to your branch.

Reproduction. I ran this on a disposable Omarchy VM with real libfprint-git and fprintd, no reader attached, and only USB detection stubbed. On quattro (8b4eae6), setup goes into sudo fprintd-enroll, which fails at once with NoSuchDevice: No devices available, followed by "Enrollment failed. Please try again." With this branch merged into quattro, setup stops before enrollment with your diagnosis. test/shell.d/fingerprint-package-test.sh passes 7 of 7 on the merge and fails with the fix reverted. ./test/cli has one failure (vscode generated theme references current theme file), and quattro has the same failure without this change.

Pushed: fc38b33. fprintd-list exits 1 for more than an empty device list. It also exits 1 when fprintd or D-Bus fails, and when polkit denies the verify permission that listing enrolled prints needs. With libfprint's virtual device standing in for a working reader, an SSH session got ListEnrolledFingers failed: ... PermissionDenied, so the original check would have reported that reader as having no driver. The check now matches the exact No devices available line fprintd prints for an empty list (untranslated, utils/list.c), and leaves every other failure to enrollment as before. From inside the desktop session, which is how the menu and the first-run notification launch it, the virtual reader passes the check and setup goes on to enrollment. The comment no longer says fprintd-enroll hangs: with no device it exits immediately. The test now checks that the message is printed, and that a different fprintd-list error still reaches enrollment.

Second opinion. Codex at medium effort reviewed it twice. It raised the daemon and D-Bus failures behind the broad exit-status check and the untested message, and agreed with the hang correction. Its independence from this review is not guaranteed. On the fixed head it found one remaining point, left for the maintainer rather than changed here: an empty device list proves fprintd sees no reader, not that the sensor is unsupported. A supported reader that disconnects after detection would get the same "isn't supported yet, retrying won't help" message. That wording is your call and the maintainer's.

Related. #12480 fixes the same bug with a GetDefaultDevice probe over busctl. Asked to choose between the two without being told mine, Codex chose this one, and so do I. That probe stops setup on any D-Bus or daemon failure, not only an empty device list, and this one does the same job with less. I also linked #10892, the same symptom on a Validity VFS491, which asks for exactly this check.

This is waiting on the maintainer: whether to keep the message as worded, and which of this and #12480 to take.

@omarchybot omarchybot removed verified Omarchy Triage has verified that this issue is ready for final review labels Oct 4, 2026
Keeps both sets of new cases in fingerprint-package-test.sh: the no-driver diagnosis cases from this branch and quattro's apply-lock setup cases from omacom#7158.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@omarchybot omarchybot added the verified Omarchy Triage has verified that this issue is ready for final review label Oct 7, 2026
@omarchybot

Copy link
Copy Markdown
Collaborator

Merged quattro into your branch to clear the conflict, and verified it again: cc721dc.

The conflict. #7158 landed on quattro and reworked test/shell.d/fingerprint-package-test.sh around omarchy-apply-lock. The merge keeps both sets of cases: your two no-driver cases now run under its new harness, followed by its two apply-lock cases. The setup script merged without conflict, and the fprintd-list check still runs before enrollment, setup_pam_config and omarchy-apply-lock.

Verification. I ran this on a disposable Omarchy VM with real fprintd 1.94.5 and libfprint-git, no reader attached, and only USB detection stubbed. On quattro (d9b970d), setup installs the packages, goes into fprintd-enroll, fails with NoSuchDevice: No devices available, and says "Enrollment failed. Please try again." On this head, it stops before enrollment with your diagnosis. fprintd-list $USER printed exactly No devices available and exited 1. fingerprint-package-test.sh passes 9 of 9 here, and fails at the no-driver case with the script change reverted. ./test/cli passes in full.

Second opinion. Codex at medium effort reviewed the merge. It found the conflict resolution correct, found no interaction with #7158's apply-lock changes, and judged the fix proportionate. It agreed with this review, and its independence from it is not guaranteed. It raised one point again: an empty device list shows that fprintd cannot use the reader, not that libfprint lacks a driver for it. So "this exact sensor isn't supported yet, retrying won't help" claims more than the check proves. Its suggestion was "A fingerprint reader was detected, but fprintd cannot use it. This may mean the installed libfprint does not support this sensor." I have not changed it, because user-facing wording is the maintainer's call and yours.

This is waiting on the maintainer: whether to keep the message as worded.

@jandrusk

Copy link
Copy Markdown

Verified approach for #11841. Same diagnosis class as #11047 and #15080 — please pick one to merge and close the others as dups.

…kely cause

An empty fprintd device list shows fprintd cannot use the reader, not that libfprint has no driver for it: a driver that failed to initialise leaves the same empty list. So the message no longer says the sensor isn't supported or that retrying won't help; it says fprintd has no usable device, that libfprint most likely has no driver, and to look the USB ID up with lsusb. A new case checks that a failed scan on a reader fprintd can see still asks for a retry. Both come from omacom#15080, which made the same fix after enrollment and is folded in here.

Co-Authored-By: Omarchybot <317366263+omarchybot@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@omarchybot

Copy link
Copy Markdown
Collaborator

Pushed one commit to your branch, 6f83c77, which folds in the parts of #15080 that this pull request did not have, and verified it again.

What changed and why. fprintd lists no device when libfprint has no driver for the reader, but also when a driver exists and failed to initialise. So "libfprint has no driver for it" and "this exact sensor isn't supported yet, retrying won't help" claimed more than the check shows. The message now reads "Your fingerprint reader was detected, but fprintd has no usable device. libfprint most likely has no driver for this reader." The second line now says to look up the reader's USB ID (shown by lsusb) on the supported-devices list, and still mentions the Discord. That wording, and a test case checking that a failed scan on a reader fprintd can see still asks for a retry, come from #15080. #15080 made the same fix but only after enrollment had failed, so this one, which was opened first, keeps the fix. The check itself is unchanged.

Verification. I ran this on a disposable Omarchy VM with real fprintd 1.94.5 and libfprint-git, no reader attached, and only USB detection stubbed. On quattro (33783b1), and on this head with the check removed, setup goes into fprintd-enroll, which fails with NoSuchDevice: No devices available, and then says "Enrollment failed. Please try again." On this head, setup stops before enrollment with the new message. fingerprint-package-test.sh passes 10 of 10 and fails at the no-driver case with the check reverted. hw-fingerprint-test.sh passes 11 of 11, and ./test/cli passes in full.

Second opinion. Codex at medium effort reviewed the head and found no defects. It confirmed that the tests reject both a version that drops the message and one that treats every fprintd-list failure as no device. It agreed with this review, and its independence from it is not guaranteed.

Related. #11047 adds a python-validity hint for four Synaptics/Validity readers on the path after a failed enrollment, and it also keys on No devices available. If both land, this check exits first and that hint is never shown. Codex raised this as well and suggested moving the hint into this early branch. Whichever lands second should do that.

This is waiting on CI for the new head, and then on the maintainer.

@omarchybot omarchybot removed the verified Omarchy Triage has verified that this issue is ready for final review label Oct 17, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

bug Something isn't working

Projects

None yet

4 participants