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

Rhosydd only uses the SDK-API library canterbury-0.

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

Differential Revision: https://phabricator.apertis.org/D6591

Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>

Differential Revision: https://phabricator.apertis.org/D6005

This test won't get far without them.

Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>

Differential Revision: https://phabricator.apertis.org/D6004

This removes some Makefile.am rules and simplifies the build system a
bit.

Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Reviewed-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4912

Mathieu Duponchelle authored 8 years ago and Andrej Shadura committed 8 years ago

The following issue types were raised by hotdoc:

* Invalid plural links: hotdoc does not know how to pluralize
  symbols such as `#RsdSubscription`.

* Invalid inter-documentation links: hotdoc does not yet install
  symbol databases, links to other documentation symbols are only
  possible when the link target is documented by gtk-doc.

* Invalid function links, for example when referring to a
  user-provided `main()` function.

Signed-off-by: Mathieu Duponchelle <mathieu.duponchelle@opencreed.com>
Signed-off-by: Andrew Shadura <andrew.shadura@collabora.co.uk>
Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4490

Mathieu Duponchelle authored 8 years ago and Andrej Shadura committed 8 years ago

And reduce our extensions requirements to the newly-consolidated
hotdoc packages.

+ Update depends in debian/control

Signed-off-by: Mathieu Duponchelle <mathieu.duponchelle@opencreed.com>
Signed-off-by: Andrew Shadura <andrew.shadura@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4491

 • Add apparmor rules to enable Apertis applications to talk to rhosydd.

Signed-off-by: Andrew Shadura <andrew.shadura@collabora.co.uk>

Madgula Rohan authored 8 years ago and Emanuele Aina committed 8 years ago

Any store or built-in applications should be able to communicate
with rhosydd service.

Signed-off-by: Rohan Madgula <Rohan.Madgula@in.bosch.com>
Differential Revision: https://phabricator.apertis.org/D4296

Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>

Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>

Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>

+ Use the hotdoc autotools helpers

This requires using -Wno-portability in the call to AM_INIT_AUTOMAKE,
as discussed with Simon it's a reasonable requirement for using the
hotdoc.mk fragment Makefile.

The produced html is now cleanly installable and uninstallable.

+ Make documentation build parallel for croesor and rhosydd

The issue preventing this is fixed upstream, although in a
different manner than the linked github issue proposed.

+ Remove distribution of the generated documentation

Distributing build products in the source tarball is a bad practice,
and there's no compelling reason for doing this in Apertis libraries.

+ Use a sitemap

This means subpages do not need to be linked and updated in markdown
files.

+ Enable ‘smart index’

+ Enable devhelp books generation

+ Update the debian files

+ remove build-depends on hotdoc-dbus-extension, not sure why it
  was added in the first place.

Using this new feature of hotdoc means we no longer need to maintain
potentially huge lists of symbols in the markdown files.

As desired by Philip, we now pass the C sources to hotdoc in the
Makefile, which means the source list does not need updating.

However, combined with smart-index, this means every public symbol
is now exposed in the documentation.

Subsequent patches should filter out potentially undesirable files
or symbols, this requires more in-depth knowledge of the project
than I currently possess.

Signed-off-by: Mathieu Duponchelle <mathieu.duponchelle@opencreed.com>

Differential Revision: https://phabricator.apertis.org/D3784

This ignores:

+ The hotdoc depdir folder
+ The stamp files generated by the hotdoc fragment
+ The private folder for each hotdoc project
+ The output directory for each hotdoc project

Signed-off-by: Mathieu Duponchelle <mathieu.duponchelle@opencreed.com>

Differential Revision: https://phabricator.apertis.org/D3873

Signed-off-by: Mathieu Duponchelle <mathieu.duponchelle@opencreed.com>

Differential Revision: https://phabricator.apertis.org/D3872

Links to symbols use the following syntax: [](symbol-name), not
[symbol-name]().

Signed-off-by: Mathieu Duponchelle <mathieu.duponchelle@opencreed.com>

Differential Revision: https://phabricator.apertis.org/D3783

This is not yet supported by hotdoc.

Signed-off-by: Mathieu Duponchelle <mathieu.duponchelle@opencreed.com>

Differential Revision: https://phabricator.apertis.org/D3793

The syntax for linking to a recognized symbol is [](symbol-name),
not [symbol-name]().

Signed-off-by: Mathieu Duponchelle <mathieu.duponchelle@opencreed.com>

Differential Revision: https://phabricator.apertis.org/D3782

This updates it to the latest recommended style. See the commit for
rationale about the change.

Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3927

 • Replace UNRELEASED with 0.5.0 in documentation
 • Update symbols files to include all 0.5.0 symbols
 • Bump library SONAME, rename librhosydd package accordingly
 • libcroesor has not changed API at all
 • Update NEWS

Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>

If a client subscribes to property changes from an aggregate vehicle
before its cache is populated, they would previously have been sent out
with a clock delta of zero. If the vehicle’s cache was then populated,
this clock delta would have been updated, potentially resulting in time
going backwards. As we all know from popular culture, this rarely ends
well. Although the UK government is currently having a good attempt to
make time go backwards.

This was previously partially masked by the fact that the timestamps for
attributes forwarded by VddAggregateVehicle::attributes-changed were not
updated to be in the aggregate clock domain. Do that too.

Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3727

Reviewed-by: Mathieu Duponchelle <mathieu.duponchelle@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3726

If rhosydd-client is running on the same machine as isolation.py is run
on, its network and mount namespaces will cause the test to fail, as it
uses the standard namespaces, and hence makes the test think that an
important part of Rhosydd is (erroneously) using the standard
namespaces.

rhosydd-client itself doesn’t need to use mount namespaces as it doesn’t
have access to any more data than any other consumer of the sensors and
actuators SDK API.

rhosydd-client could be running on the machine while the tests are
running if a developer is doing multiple things at once.

Reviewed-by: Mathieu Duponchelle <mathieu.duponchelle@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3725

Reviewed-by: Mathieu Duponchelle <mathieu.duponchelle@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3724

Otherwise monitor-vehicle will sit there forever doing nothing.

Reviewed-by: Mathieu Duponchelle <mathieu.duponchelle@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3723

If any code is going to use this generically, it needs to be on the
interface, not on one of its implementations. It’s a sufficiently
general signal that having it on the interface is OK.

This is an API addition (RsdVehicle::invalidated), but not a break, as
RsdProxyVehicle still has the same signal (just provided by an interface
that it implements now, rather than by the class itself).

Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3722

We depend on 3.7.0 //or later//. The fact that the dependency is (by
default) explicitly equal, was not clear before.

Reviewed-by: Mathieu Duponchelle <mathieu.duponchelle@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3721

In order for the right .pc files for dependencies to be installed, we
need to list those dependencies for our *-dev packages.

https://phabricator.apertis.org/T1297



Reviewed-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3716

This should not have been there.

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3547

If a VddAggregateVehicle has a fully-valid cache, but no subscriptions
to any attributes on backend vehicles (because no clients have
subscriptions to attributes on the aggregate vehicle), an incoming
request for an attribute will return a potentially stale value from the
cache, rather than querying the backends.

Fix this by invalidating the cache on incoming requests if suitable
subscriptions are not in place.

New API:
 • rsd_subscription_manager_has_subscription()

This adds an integration test for the failure seen in T2042.

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3546

This will make an upcoming change a bit simpler. This commit introduces
no functional changes.

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3545

• Replace UNRELEASED with 0.4.0 in documentation
 • Update symbols files to include all 0.4.0 symbols
 • Bump library SONAME, rename packages accordingly
 • Update NEWS

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3534

This one will accumulate forever, so is fairly important to fix.

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3532

Also fix the return type on one of the error paths to return a NULL
pointer instead of FALSE.

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3521

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3520

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3519

They all need to be cleared before the GArray memory block itself can be
freed.

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3518

Otherwise there is no way to free the memory inside an RsdAttribute
which is on the stack.

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3517

Document that clients can expect to receive notifications about some
attributes which they have not subscribed to, if it is more efficient
for the server to emit those alongside attributes the client //has//
subscribed to.

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D3500