IJHack / QtPass / 34888458418

93%

push

github

Backport: ignore a configured GPG home that does not exist (#1711) (#1741)

* Ignore a configured GPG home that does not exist (#1711)

The 1.7.0 test suite wrote its temporary GNUPGHOME into the live
QtPass.conf as gpgHome and never restored it, so anyone who ran `make check`
as themselves (a distribution package build, #1711) had every gpg call fail
with "keyblock resource ... No such file" and "No secret key" from then on,
in every QtPass version, with nothing in the settings dialog to explain it.

Pass::init() now exports GNUPGHOME only when the configured directory
exists; otherwise it logs a warning, shows a status message naming the
path, and leaves whatever the process environment already had. Clearing
gpgHome at runtime restores the inherited value too instead of keeping the
previous path for the rest of the session.

Tests: gpgHomeExportedWhenItExists, gpgHomeMissingFallsBackAndWarns,
gpgHomeClearedRestoresInheritedValue (the last two fail on the previous
code). FAQ entry for the symptom.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JuQsrHonihp1nARE7bzstc
(cherry picked from commit 8d79f748f)

* Address review: QVERIFY2 with invariants on the gpgHome test setup and cleanup checks

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JuQsrHonihp1nARE7bzstc

* Address review: say which GNUPGHOME the missing-gpgHome fallback uses

With GNUPGHOME already in the process environment the fallback uses that,
not the default keyring; the status message and warning now say so, and the
test checks the message for both cases.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JuQsrHonihp1nARE7bzstc

---------

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>

16 of 22 new or added lines in 1 file covered. (72.73%)

5032 of 7375 relevant lines covered (68.23%)

67.24 hits per line