autogoo: replace AC_HELP_STRING with AS_HELP_STRING

See merge request !72

Niels De Graef authored Mar 23, 2024 and Niels De Graef committed Mar 25, 2024

Autoconf made the `AC_HELP_STRING` macro obsolete many moons ago in
Autoconf. It should be replaced with `AS_HELP_STRING`, so let's do that.

Default "automatically unlock" checkbox to false

Closes #7

See merge request !38

Ryan Hendrickson authored May 20, 2021 and Niels De Graef committed Mar 18, 2024

The GCR prompt used to unlock keyrings and other secure objects has a
checkbox for automatically unlocking the object when the user logs in.
Previously, this checkbox was checked by default if the object in
question had the CKA_GNOME_TRANSIENT attribute set. This linkage frankly
makes no sense; the change can be traced back to a nine-year-old commit
(084ce5e0) which appears to have mistakenly altered code that used
CKA_GNOME_TRANSIENT to determine the *nature* of the choice into code
that used CKA_GNOME_TRANSIENT to set the default *value* of the choice.

Defaulting this checkbox to true, which was the practical effect of
using CKA_GNOME_TRANSIENT in this way, has obvious drawbacks for
security, as the reporter of #7 and the various reporters in bugs linked
from there have attested.

Removing this linkage is a small change, but it results in a larger
amount of code being no longer needed, which this commit cleans up.

Fixes #7.

build: Fix distcheck with systemd units

See merge request !70

Dan Nicholson authored Mar 14, 2024 and Niels De Graef committed Mar 18, 2024

The user unit directory gathered from pkg-config is an absolute path and
doesn't follow the configured prefix. distcheck doesn't like that, so
set `systemduserunitdir` to follow the prefix even though that wouldn't
make sense on a real build.

rpc-layer: Fix undefined reference to g_strlcpy

See merge request !66

fixes build with -Wl,-z,defs

Signed-off-by: Dhanuka Warusadura <dhanuka@gnome.org>

ssh-agent: avoid deadlock when agent process dies before we connect to it

See merge request !65

If the ssh-agent process we launch dies before it becomes ready to take
input, self->pid will be set to 0 by on_child_watch().  If that happens,
there's no point in continuing to wait for the process to become ready.
This should avoid an unnecessary five-second wait in cases like [1] or
[2].

[1] https://bugzilla.gnome.org/show_bug.cgi?id=794848
[2] https://bugzilla.redhat.com/show_bug.cgi?id=1841855 (migrated to
https://issues.redhat.com/browse/RHEL-9302)

gkd_ssh_agent_process_connect() waits for the ssh-agent process to
become ready to accept input by entering the main loop while holding
self->lock.  However, if the ssh-agent process dies before becoming
ready, the main loop will call on_child_watch(), which needs to take
self->lock, causing a deadlock.  Fix this by releasing the lock before
entering the main loop.

This should prevent a busyloop that's been reported multiple times [1]
[2] from lasting forever.

[1] https://bugzilla.gnome.org/show_bug.cgi?id=794848
[2] https://bugzilla.redhat.com/show_bug.cgi?id=1841855 (migrated to
https://issues.redhat.com/browse/RHEL-9302)

gkm: fix test-sexp with GCC 14

See merge request !62

pkcs11/gkm/test-sexp.c: In function ‘test_sign_verify’:
pkcs11/gkm/test-sexp.c:231:92: error: passing argument 6 of ‘gkm_crypto_sign_xsa’ from incompatible pointer type [-Wincompatible-pointer-types]
In file included from pkcs11/gkm/test-sexp.c:31:
./pkcs11/gkm/gkm-crypto.h:92:86: note: expected ‘CK_ULONG_PTR’ {aka ‘long unsigned int *’} but argument is of type ‘gsize *’ {aka ‘unsigned int *’}
pkcs11/gkm/test-sexp.c:251:87: error: passing argument 6 of ‘gkm_crypto_sign_xsa’ from incompatible pointer type [-Wincompatible-pointer-types]
./pkcs11/gkm/gkm-crypto.h:92:86: note: expected ‘CK_ULONG_PTR’ {aka ‘long unsigned int *’} but argument is of type ‘gsize *’ {aka ‘unsigned int *’}
pkcs11/gkm/test-sexp.c:271:91: error: passing argument 6 of ‘gkm_crypto_sign_xsa’ from incompatible pointer type [-Wincompatible-pointer-types]
./pkcs11/gkm/gkm-crypto.h:92:86: note: expected ‘CK_ULONG_PTR’ {aka ‘long unsigned int *’} but argument is of type ‘gsize *’ {aka ‘unsigned int *’}

Signed-off-by: Yaakov Selkowitz <yselkowi@redhat.com>

Signed-off-by: Dhanuka Warusadura <dhanuka@gnome.org>

build: Remove build with ssh component from default build instructions

See merge request GNOME/gnome-keyring!60

The SSH functionality of gnome-keyring has been moved to gcr [1] as
part of the modular gnome-keyring Fedora change [2] proposal. And
currently the SSH functionality is provided by `gcr-ssh-agent`. So, it
is recommended to build gnome-keyring without the `ssh` component.

[1] gcr!67
[2] https://fedoraproject.org/wiki/Changes/ModularGnomeKeyring



Signed-off-by: Dhanuka Warusadura <dhanuka@gnome.org>

main: Reduce scope of ssh_started variable

See merge request GNOME/gnome-keyring!59

This change fixes the "‘ssh_started’ defined but not
used [-Wunused-variable]" warning, during build with
`--disable-ssh-agent` option.

Signed-off-by: Dhanuka Warusadura <dhanuka@gnome.org>

desktop: Expand variables

Closes #131

See merge request GNOME/gnome-keyring!58

2f202ab6 removed expansion of
variables, but `@bindir@` still needed to be expanded.

This partially reverts that commit, without re-introducing Bugzilla
related data.

Also the makefile rule for `$(desktop_DATA)` was using the same .in
file as input for all the desktop outputs. Resulting in the same file
instead of having the correct `--components`.

desktop: Remove obsolete Bugzilla entries

See merge request GNOME/gnome-keyring!55

The `X-GNOME-Bugzilla-*` entries were for use by bug-buddy, a GNOME 2
technology that's been gone for over a decade. These entries are
obsolete and can be removed from the desktop files.

The desktop files then have no variables so do not need to be
configured. They are renamed from `*.in.in` to `*.in` to reflect that
and build files are updated for this change.