Add a test-case for combining receive_type and send_destination

Until the previous commit, this would have worked. Now it correctly fails
with "send and receive attributes cannot be combined".

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Thiago Macieira <thiago@kde.org>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=92853

config-parser: Clarify how <allow>, <deny> attributes work

The giant conditionals used to check policy attributes are increasingly
unwieldy, so let's try something else. Bundle together the send_
attributes, the receive_ attributes, the eavesdrop attribute
(which can go on either send or receive rules) and the other attributes
into equivalence classes, and write the conditionals in terms of those
equivalence classes.

In particular, this correctly forbids
    <allow receive_type="..." send_destination="..."/>
which was previously allowed but nonsensical (the send part took
precedence and the receive part was ignored).

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Reviewed-by: Thiago Macieira <thiago@kde.org>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=92853

Merge branch 'dbus-1.10'

NEWS for #101858

config-loader-expat: Tell Expat not to defend against hash collisions

By default, Expat uses cryptographic-quality random numbers as a salt for
its hash algorithm, and since 2.2.1 it gets them from the getrandom
syscall on Linux. That syscall refuses to return any entropy until the
kernel's CSPRNG (random pool) has been initialized. Unfortunately, this
can take as long as 40 seconds on embedded devices with few entropy
sources, which is too long: if the system dbus-daemon blocks for that
length of time, important D-Bus clients like systemd and systemd-logind
time out and fail to connect to it.

We're parsing small configuration files here, and we trust them
completely, so we don't need to defend against hash collisions: nobody
is going to be crafting them to cause pathological performance.

Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101858
Tested-by: Christopher Hewitt <hewitt@ieee.org>
[smcv: Adjust build-system changes for 1.11.x]
Signed-off-by: Simon McVittie <smcv@debian.org>
Reviewed-by: Philip Withnall <withnall@endlessm.com>

config-loader-expat: Tell Expat not to defend against hash collisions

By default, Expat uses cryptographic-quality random numbers as a salt for
its hash algorithm, and since 2.2.1 it gets them from the getrandom
syscall on Linux. That syscall refuses to return any entropy until the
kernel's CSPRNG (random pool) has been initialized. Unfortunately, this
can take as long as 40 seconds on embedded devices with few entropy
sources, which is too long: if the system dbus-daemon blocks for that
length of time, important D-Bus clients like systemd and systemd-logind
time out and fail to connect to it.

We're parsing small configuration files here, and we trust them
completely, so we don't need to defend against hash collisions: nobody
is going to be crafting them to cause pathological performance.

Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101858
Signed-off-by: Simon McVittie <smcv@debian.org>
Tested-by: Christopher Hewitt <hewitt@ieee.org>
Reviewed-by: Philip Withnall <withnall@endlessm.com>

Start 1.10.24

Signed-off-by: Simon McVittie <smcv@collabora.com>

Belatedly bump the libtool versioning

Signed-off-by: Simon McVittie <smcv@collabora.com>

Start towards 1.11.18

Signed-off-by: Simon McVittie <smcv@collabora.com>

1.11.16

1.10.22

NEWS

Signed-off-by: Simon McVittie <smcv@collabora.com>

DBusMainLoop: ensure all required timeouts are restarted

This is a followup of 529600397bcab47b9bed5da9208c2df05c8b86b4. We can't
shortcut the timeouts iteration in order not to miss any timeouts that
might require timestamp restart.

Bug: https://bugs.freedesktop.org/show_bug.cgi?id=95619
Reviewed-by: Simon McVittie <smcv@collabora.com>

Merge branch 'dbus-1.10'

NEWS for fd.o#101716

Signed-off-by: Simon McVittie <smcv@collabora.com>

test/name-test: Be compatible with Python 3

configure.ac will detect PYTHON=python3 if there is no python
executable in the PATH.

Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101716
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Signed-off-by: Simon McVittie <smcv@collabora.com>

Merge branch 'dbus-1.10'

NEWS for #101698

Signed-off-by: Simon McVittie <smcv@collabora.com>

tests: Make tests fail if they try to connect to the real session bus

It is too easy for a developer working in an environment that has a
session bus to write tests that pass locally, but fail in minimal
environments. This is also risky because the tests might do
destructive things on the developer's real session bus. We can avoid
connecting to the session bus by consistently removing its address
from the environment, and replacing it with something that will
always fail.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101698

test/dbus-daemon: Unset DBUS_SESSION_BUS_ADDRESS

When we intend to exercise the default behaviour in the absence of
DBUS_SESSION_BUS_ADDRESS (but with an XDG_RUNTIME_DIR present), it would
help if we unset DBUS_SESSION_BUS_ADDRESS. Otherwise we'll just connect
to the real session bus, if there is one.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101698

Update NEWS for fd.o#101700

Signed-off-by: Simon McVittie <smcv@collabora.com>

Update NEWS for 1.10 branch

Signed-off-by: Simon McVittie <smcv@collabora.com>

name-test: Backport dbus-run-session wrapper from git master

test-pending-call-disconnected relies on being run under a session bus.
On master, the TESTS in this directory all get that treatment, but
in dbus-1.10 they do not. This caused test-pending-call-disconnected
to fail in minimal environments like travis-ci where there is no
developer-initiated session bus.

Backport part of commit ec6b220 "name-test: run most C tests directly,
not via run-test.sh" to wrap it in dbus-run-session. This is better
than putting it in run-test.sh because this way, its TAP output is
parsed directly by Automake.

It also has the side benefit of exercising dbus-run-session in the
automated tests.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101698

system.conf: Allow unprivileged users to get dbus-daemon properties

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101700

Update NEWS for the various things that ended up in #101568

Signed-off-by: Simon McVittie <smcv@collabora.com>

test-variant: Add a regression test for DBusVariant

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

message: Add DBusVariant, a way to copy a single message item

For #100344, we will need a way to store the metadata from the
original method call, and copy them back into arbitrarily many
messages later. This would be easy in GDBus, which has GVariant
as a first-class object. However, libdbus doesn't have an object for
message items, only messages.

We could copy the message's content, but it will carry file descriptors,
which we don't want to copy. Instead, introduce an internal object
representing a message item in a small buffer. It is stored as a variant
(D-Bus type 'v') so that it naturally carries its own type.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

Test dbus_message_iter_abandon_container_if_open under OOM conditions

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

DBusMessageIter: Add a function to abandon possibly-zero-filled iterators

See the doc-comment of the new
dbus_message_iter_abandon_container_if_open() function for details.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

dbus_message_iter_open_container: Don't leak signature on failure

If we run out of memory while calling _dbus_type_writer_recurse()
(which is impossible for most contained types, but can happen for
structs and dict-entries), then the memory we allocated in the call to
_dbus_message_iter_open_signature() will still be allocated, and we
have to free it in order to return to the state of the world prior to
calling open_container().

One might reasonably worry that this change can break callers that use
this (incorrect) pattern:

    if (!dbus_message_iter_open_container (outer, ..., inner))
      {
        dbus_message_iter_abandon_container (outer, inner);
        goto fail;
      }
    /* now we know inner is open, and we must close it later */

However, testing that pattern with _dbus_test_oom_handling()
demonstrates that it already dies with a DBusString assertion failure
even before this commit.

This is all concerningly fragile, and I think the next step should be
to zero out DBusMessageIter instances when they are invalidated, so
that a "double-free" is always detected.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568
(cherry picked from commit 031aa2ceb3dfff373e7b398dfc5d020d77262512)

dbus_message_iter_append_basic: Don't leak signature if appending fd fails

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568
(cherry picked from commit 8384e795516066960bb9fcfbfe138f569420edb9)

dbus_message_append_args_valist: Don't leak memory on inappropriate type

Found by source code inspection while trying to debug an unrelated
leak.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568
(cherry picked from commit 6b7bdb105b120b3db312de93af94af1bb6a2a474)

DBusMessageIter: Zero out the iterator on failure

This ensures that callers won't accidentally use it for something
in a way that is considered to be programmer error.

In _dbus_message_iter_check(), insert a specific check for this before
dereferencing iter->message, so that we get a nice assertion failure
(potentially non-fatal) instead of a segfault.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

DBusMessageIter: Clarify the API

Having opened a container for appending, the container must be closed
exactly once.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

test/message: Add a targeted test for recently-fixed leaks

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

internals: Make a minimal _dbus_test_oom_handling() universally available

Previously, it was only available under DBUS_ENABLE_EMBEDDED_TESTS,
because the infrastructure to pretend malloc had failed is only
compiled then. However, I'd like to use it in more modular tests, to
avoid test-dbus continuing to grow. To facilitate that, inline a
trivial version of it when DBUS_ENABLE_EMBEDDED_TESTS is disabled:
it just calls the function, once, without doing any strange things to
the malloc interface.

Similarly, amend the stub implementation of
_dbus_get_malloc_blocks_outstanding() so that references to it are
syntactically valid, and move the DBusTestMemoryFunction typedef so
that it can be used with or without DBUS_ENABLE_EMBEDDED_TESTS.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

dbus_message_iter_open_container: Don't leak signature on failure

If we run out of memory while calling _dbus_type_writer_recurse()
(which is impossible for most contained types, but can happen for
structs and dict-entries), then the memory we allocated in the call to
_dbus_message_iter_open_signature() will still be allocated, and we
have to free it in order to return to the state of the world prior to
calling open_container().

One might reasonably worry that this change can break callers that use
this (incorrect) pattern:

    if (!dbus_message_iter_open_container (outer, ..., inner))
      {
        dbus_message_iter_abandon_container (outer, inner);
        goto fail;
      }
    /* now we know inner is open, and we must close it later */

However, testing that pattern with _dbus_test_oom_handling()
demonstrates that it already dies with a DBusString assertion failure
even before this commit.

This is all concerningly fragile, and I think the next step should be
to zero out DBusMessageIter instances when they are invalidated, so
that a "double-free" is always detected.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

dbus_message_iter_append_basic: Don't leak signature if appending fd fails

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

_dbus_message_iter_open_signature: Clarify why this is not leaky

The same assertion appears closer to the top of the function, and there
is no opportunity for it to have become false here.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

dbus_message_append_args_valist: Don't leak memory on inappropriate type

Found by source code inspection while trying to debug an unrelated
leak.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

_dbus_marshal_validate_test: Uncomment commented-out test coverage

This was added around 12½ years ago, in a commented-out state, and has
remained commented out ever since. It turns out these test vectors
do pass, although perhaps they didn't at the time.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

_dbus_marshal_validate_test: Merge two sets of signature validity checks

The deleted lines used to be a test for _dbus_validate_signature(),
until I deleted that function. We also had a completely separate
test for _dbus_validate_signature_with_reason() which remains present.
Some of the test vectors were tested in both places.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

internals: Decouple logging an error from exiting unsuccessfully

This lets _dbus_warn() and _dbus_warn_check_failed() fall through
to flushing stderr and calling _dbus_abort(), meaning that failed
checks and warnings can result in a core dump as intended.
By renaming the FATAL severity to ERROR, we ensure that any code
contributions that assumed the old semantics will fail to compile.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

Remove now-unused _dbus_validate_signature()

All callers should use _dbus_validate_signature_with_reason() directly.
The only remaining callers were this function's own tests.

As a side benefit, this commit removes a TODO pointing out that this
function did not follow normal DBusString conventions, by considering
a length outside the bounds of the DBusString to be an ordinary
lack of validity rather than a fatal programming error.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

Remove now-unused _dbus_check_is_valid_signature()

As noted in the previous commit, it's a trap.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

DBusMessage: Stop using _dbus_check_is_valid_signature()

This function looks appealing, but it is a trap, particularly in
_dbus_return_val_if_fail() checks. It returns a boolean result, which
cannot distinguish between "failed because we ran out of memory" and
"failed because the string is actually invalid"; but
_dbus_validate_signature_with_reason() allocates memory. Use the
over-complicated version directly, so libdbus can continue to
bend over backwards to support the (possibly mythical) operating systems
that limit memory consumption and do not overcommit, such that malloc()
can genuinely return NULL.

Bug detected by running the DBusVariant unit test (fd.o #101568) under
dbus' failing-malloc() instrumentation.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

_dbus_message_set_signature: Delete unused function

If this is reinstated it will need some checks. In particular, it
was using _dbus_check_is_valid_signature() in an unsafe way:
_dbus_check_is_valid_signature() cannot be used in a
_dbus_return_val_if_fail() check because it does not distinguish
between error by the caller, and out-of-memory conditions.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568

NEWS: Reference #101569

Signed-off-by: Simon McVittie <smcv@collabora.com>

build: Introduce ${runstatedir} and use it for the pid file

By default ${runstatedir} is the same as ${localstatedir}/run, but many
Linux distributions configure it to be /run and mount a tmpfs in that
location. All other factors being equal, it is preferable to use /run
where available because it is guaranteed to be local, whereas traversing
/var might involve automounting a networked filesystem (even though
/var/run itself is very likely to be a tmpfs).

/run or /var/run is currently only used in a few places in dbus, but
I plan to make more use of it during the development of
<https://bugs.freedesktop.org/show_bug.cgi?id=100344>.

The pid file is not part of the API between dbus and other software
(other than distribution init scripts for dbus itself), so we do not
need to keep it strictly compatible; so it is OK to move it.

We do not yet use /run for the system bus socket, because that is
part of the API between D-Bus clients and servers, and has always been
"officially" /var/run/dbus/system_bus_socket.
<https://bugs.freedesktop.org/show_bug.cgi?id=101628> tracks the
possibility of changing that.

Similarly, we do not replace /var/run/console with /run/console, because
that path is part of the API between dbus-daemon and the obsolete PAM
modules pam_console and pam_foreground that used /var/run/console.
<https://bugs.freedesktop.org/show_bug.cgi?id=101629> tracks the possible
future removal of that code path.

In the CMake build system, the equivalent of ${runstatedir} remains
hard-coded to the equivalent of ${localstatedir}/run for simplicity. For
the sort of system-wide installations that would consider redefining
${runstatedir} to /run, the Autotools build system is strongly
recommended: in particular this is what Linux distributions are expected
to use.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101569

Start on 1.10.22

Signed-off-by: Simon McVittie <smcv@debian.org>

spec: Update my email address

Signed-off-by: Simon McVittie <smcv@collabora.com>

Start 1.11.16 development

Signed-off-by: Simon McVittie <smcv@collabora.com>

1.10.20

Signed-off-by: Simon McVittie <smcv@collabora.com>

1.11.14

Signed-off-by: Simon McVittie <smcv@collabora.com>

NEWS for spec changes from #101567

Signed-off-by: Simon McVittie <smcv@collabora.com>

spec: Document versioning of eavesdrop='true'

The wording and formatting used here is consistent with other
semi-recently-added match keys.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101567

spec: Formally deprecate eavesdropping

Reviewed-by: Philip Withnall <withnall@endlessm.com>
[smcv: Wrap BecomeMonitor in <literal> as per Philip's review]
Signed-off-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101567

spec: Do not promise match rules with eavesdrop='true' can be added

This is no longer true, and it seems less misleading to raise an
error than to obey the letter of the spec by quietly ignoring calls
from an inappropriate caller.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101567

NEWS: Mention the user-visible part of #101567

Signed-off-by: Simon McVittie <smcv@collabora.com>

bus_driver_send_ack_reply: Make available to other modules

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101567

driver: Use a data-driven approach to marking methods as privileged

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101567

driver: Share bus_driver_get_conn_helper with other modules

Now that we're starting to implement methods in more places, it makes
sense to share this code. The Stats interface can already benefit.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101567

driver: Make eavesdropping a privileged action

Eavesdropping on unicast messages to other processes is not something
that should be done by processes in containers, or on the system bus
by users other than root or the bus owner. bus/system.conf.in
does not enable eavesdropping, but adding inadvisable configuration
could. This brings it into line with Monitoring.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101567

NEWS: Document #101567

Signed-off-by: Simon McVittie <smcv@collabora.com>

test/loopback: Exercise unix:dir addresses

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101567

test/loopback: Assert that the connectable address makes sense

When we listen on a tcp: address we should get a connectable tcp:
address, and so on.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101567

test/loopback: Don't leak address strings

dbus_server_get_address() returns a copy. It isn't clear why.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101567

Add unix:dir=/something addresses

These are like unix:tmpdir=/something, except that the resulting
socket is always path-based, never abstract.

This is desirable for two reasons:

* If a Linux container manager wants to expose a path-based socket
  into the container, it can do so by bind-mounting it in the
  container's filesystem namespace. That cannot work for abstract
  sockets because they are not files.

* Conversely, if a Linux container manager does not want to expose
  a path-based socket in the container, it can avoid bind-mounting it,
  or bind-mount some harmless object like /dev/null over it.
  That cannot work for abstract sockets because access to abstract
  sockets is part of the network namespace, which is all-or-nothing.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101567

Update NEWS for fd.o #101570

Signed-off-by: Simon McVittie <smcv@collabora.com>

Install a systemd tmpfiles.d snippet to create /var/lib/dbus/machine-id

On systemd systems, /etc/machine-id is guaranteed to exist and has
the same format as the D-Bus machine ID. The major D-Bus implementations
read /etc/machine-id if it exists, but some less up-to-date
implementations still only read /var/lib/dbus/machine-id. We can be
nice to those implementations by ensuring /var/lib/dbus/machine-id
is a symlink; this way, the two files can never get out of sync.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101570

NEWS: Re-word

Signed-off-by: Simon McVittie <smcv@collabora.com>

Merge branch 'dbus-1.10'

Update NEWS for 1.10.x branch

Signed-off-by: Simon McVittie <smcv@collabora.com>

transport: Don't pile up errors for semicolon-separated components

If we somehow get an autolaunch address with multiple
semicolon-separated components, and one of them fails, then we will
hit an assertion failure when we try the next one.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101257
(cherry picked from commit ecdcb86bff42d2bb9cac617bf79f0aa3d47676d9)

Doxyfile.in: do not put timestamps in HTML

The build timestamp is not particularly useful (the version number of
the package is already present in the HTML), and it prevents the build
from being reproducible. See <https://reproducible-builds.org/> for more
information.

Signed-off-by: Simon McVittie <smcv@debian.org>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=100692
(cherry picked from commit 0310ead0022b3537392869cc2ed3296ba1a7c17d)

Fix missing dbus_message_unref() in error reply path

The error message was leaked when blocking on a pending call after
the connection was disconnected.

Reviewed-by: Philip Withnall <withnall@endlessm.com>
[smcv: re-word commit message]
Reviewed-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101481

Fix missing dbus_message_unref() in error reply path

The error message was leaked when blocking on a pending call after
the connection was disconnected.

Reviewed-by: Philip Withnall <withnall@endlessm.com>
[smcv: re-word commit message]
Reviewed-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101481

test-pending-call-dispatch: Add copyright and license grant

According to git history, this test was written in 2006 by Red Hat
employee John Palmieri and has received only trivial changes since
then. Red Hat gave permission in 2007 for their contributions to
be relicensed under the MIT/X11 license. We cannot take advantage
of that permission to relicense the core library or the dbus-daemon
from GPL-2+|AFL-2.0 to MIT/X11, because one early copyright holder
(CodeFactory AB) could not be traced, but we might as well use a
permissive license for simple test code that has not had
CodeFactory AB contributions.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101481

Update NEWS

Signed-off-by: Simon McVittie <smcv@collabora.com>

Return the dbus-daemon SELinux context when asking about org.freedesktop.DBus

Currently when asked the SELinux context of the owner of
org.freedesktop.DBus, the dbus-daemon is returning an error.

In the same situation when asked about the Unix user or the PID, the
daemon would return its own user or pid. Do the same for the SELinux
context by returning the daemon one.

In particular this avoids an issue seen with systemd --user, where
dbus-daemon responds to UpdateActivationEnvironment() by passing on the
new environment to systemd with o.fd.systemd1.Manager.SetEnvironment(),
but systemd cannot get the caller's SELinux context and so rejects the
SetEnvironment() call.

Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101315
[smcv: Extend commit message to describe the symptom this fixes]
Reviewed-by: Simon McVittie <smcv@collabora.com>

NEWS

Signed-off-by: Simon McVittie <smcv@collabora.com>

Opt-in to deprecation warnings for GLib 2.38 and 2.40

Now that g_test_trap_fork() has gone, we no longer have any calls to
GLib functions deprecated in or before 2.40.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101362

test/internals/syslog: Switch to g_test_trap_subprocess()

This re-executes the same binary with special command-line options
instead of forking. As a result, it can work on Windows, and is less
dependent on Unix libraries continuing to work across a fork().

(This has been confirmed to work in Windows binaries running under Wine.)

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101362

Bump GLib dependency for tests to 2.40

This is quite old (it's the version in Ubuntu 14.04 LTS, and older
than the version in Debian 8) but gives us g_test_skip(),
g_test_trap_subprocess() and GVariantDict, all of which will be
useful in the regression tests.

Remove workarounds for old versions.

After this commit we are still using the deprecated g_test_trap_fork(),
which will be removed in a subsequent commit. Don't opt-in to the new
deprecation warnings from 2.38 and 2.40 yet, because under our recommended
settings for dbus developers (-Werror) they would break the build.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101362

dbus-launch: Use dbus_try_get_local_machine_id()

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=13194

test-autolaunch: Make recently-added message similar to the others

Signed-off-by: Simon McVittie <smcv@collabora.com>

NEWS

Signed-off-by: Simon McVittie <smcv@collabora.com>

Refer to dbus_try_get_local_machine_id() in doc-comments

Signed-off-by: Simon McVittie <smcv@collabora.com>

Add dbus_try_get_local_machine_id()

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=13194

Make _dbus_get_local_machine_uuid_encoded() properly failable

This function already raised an error, and all callers handled that
error as gracefully as they could (because _dbus_generate_uuid() is
failable, since 2015). Given that, it seems unnecessarily hostile
to do a _dbus_warn_check_failed() unless we have no better alternative:
yes, it indicates that dbus has not been installed correctly, but
during build-time tests it's entirely reasonable that dbus has not
yet been installed.

Callers are:

* DBusConnection, to implement Peer.GetMachineId()
* The bus driver, to implement Peer.GetMachineId()
* X11 autolaunching
* dbus_get_local_machine_id()

Of those, only the last one is not in a position to return an error
gracefully, so move the _dbus_warn_check_failed() to there.

Migrate the text about the D-Bus library being incorrectly set up
into the error emitted by the Unix implementation, and to make it
less misleading, include separate error messages for both the
files we try to read:

$ bwrap --ro-bind / / --dev /dev --tmpfs /etc --tmpfs /var \
  ./tools/dbus-uuidgen --get
D-Bus library appears to be incorrectly set up: see the manual
page for dbus-uuidgen to correct this issue. (Failed to open
"/var/lib/dbus/machine-id": No such file or directory; Failed to open
"/etc/machine-id": No such file or directory)

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=13194

tests: Don't exercise GetMachineId() or autolaunch if no machine ID

At the moment there is a hack in the implementation of GetMachineId()
to stop tests from failing during "make check" on a system where
dbus has never been installed, by silently generating a new unique
fake "machine ID" for each process. I'm about to change that
behaviour to report errors properly; skip affected test-cases if we
can't read the real machine ID.

The shell scripts to test dbus-launch are run both as "make check"
tests (for which it is valid for dbus to be not correctly installed)
and as installed-tests (for which that is not valid), so make them
pass during "make check" but fail during installed testing.
The tests in bus/ and test/name-test/ are only run during "make check"
so they only have the code path where they are skipped.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=13194

uuidgen: Remove unimplemented declaration

As far as I can tell from git history, this function never existed.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=13194

NEWS

Signed-off-by: Simon McVittie <smcv@collabora.com>

test/dbus-daemon: Exercise Properties, Features and Interfaces

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101257

test/dbus-daemon: Exercise the Peer interface

We have to skip the GetMachineId() part during build-time testing
if it wouldn't work - there is no guarantee that dbus has ever been
installed on the build system. However, we can insist on it during
installed-tests, if we make sure to complete the installation for the
Travis-CI build by running dbus-uuidgen.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101257

Unix sysdeps: Only copy /etc/machine-id to ${sysconfdir} in "ensure" mode

System integration scripts use dbus-uuidgen --ensure, so they are
unaffected by this, and in particular the solution to Bug #77941
is still valid.

The shared library is typically loaded by unprivileged users, so
trying to write out the machine-id file is not going to work anyway.
However, if we *can* write to our ${sysconfdir} - notably during
`make distcheck` - then it is unexpected that merely reading the
machine ID has the side-effect of writing to ${sysconfdir},
and in particular it will make the check for a complete uninstall
fail. We definitely must not delete the machine ID during
`make uninstall`.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101257

spec: Document the Features and Interfaces properties on o.fd.DBus

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101257

spec: Document the Peer and Properties interfaces for the message bus

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101257

driver: Implement the Peer interface, for completeness

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101257

driver: Implement Properties, with Features and Interfaces properties

We recommend using Properties for this sort of thing when designing
D-Bus APIs, so it's a bit hypocritical that the reference message bus
didn't. The Features and Interfaces properties can be used for
feature-discovery as we add new larger features, while the Properties
support can be used for finer-grained properties, for example in the
interface planned for #100344.

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101257