DBus description: Add annotations for Qt support

[travier]

Add an initial set of simple annotations to help with interface code generation for Qt base applications.

[openshift-ci]

Skipping CI for Draft Pull Request.
If you want CI signal for your change, please convert it to an actual PR.
You can still manually trigger a test run with /test all

[travier]

The objective is to drop the copy we have in https://invent.kde.org/plasma/discover/-/blob/master/libdiscover/backends/RpmOstreeBackend/org.projectatomic.rpmostree1.xml with the annotations.

[travier]

Looks like I have to figure out how to make it work for gdbus-codegen first:

[2021-09-06T18:10:02.103Z] gdbus-codegen \
[2021-09-06T18:10:02.103Z] 	--interface-prefix org.projectatomic.rpmostree1 \
[2021-09-06T18:10:02.103Z] 	--c-namespace RPMOSTree \
[2021-09-06T18:10:02.103Z] 	--c-generate-object-manager \
[2021-09-06T18:10:02.103Z] 	--generate-c-code rpm-ostreed-generated \
[2021-09-06T18:10:02.103Z] 	--generate-docbook rpm-ostreed-generated \
[2021-09-06T18:10:02.103Z] 	./src/daemon/org.projectatomic.rpmostree1.xml \
[2021-09-06T18:10:02.103Z] 	
[2021-09-06T18:10:02.103Z] Traceback (most recent call last):
[2021-09-06T18:10:02.103Z]   File "/usr/bin/gdbus-codegen", line 55, in <module>
[2021-09-06T18:10:02.103Z]     sys.exit(codegen_main.codegen_main())
[2021-09-06T18:10:02.103Z]   File "/usr/share/glib-2.0/codegen/codegen_main.py", line 406, in codegen_main
[2021-09-06T18:10:02.103Z]     parsed_ifaces = parser.parse_dbus_xml(
[2021-09-06T18:10:02.103Z]   File "/usr/share/glib-2.0/codegen/parser.py", line 302, in parse_dbus_xml
[2021-09-06T18:10:02.103Z]     parser = DBusXMLParser(xml_data, h_type_implies_unix_fd)
[2021-09-06T18:10:02.103Z]   File "/usr/share/glib-2.0/codegen/parser.py", line 58, in __init__
[2021-09-06T18:10:02.103Z]     self._parser.Parse(xml_data)
[2021-09-06T18:10:02.103Z]   File "/builddir/build/BUILD/Python-3.9.6/Modules/pyexpat.c", line 407, in StartElement
[2021-09-06T18:10:02.103Z]   File "/usr/share/glib-2.0/codegen/parser.py", line 151, in handle_start_element
[2021-09-06T18:10:02.103Z]     self._cur_object.annotations.append(anno)
[2021-09-06T18:10:02.103Z] AttributeError: 'NoneType' object has no attribute 'annotations'

[cgwalters]

Looks like it's this stray annotation:

diff --git a/src/daemon/org.projectatomic.rpmostree1.xml b/src/daemon/org.projectatomic.rpmostree1.xml
index 49775a2c..9ae49a6d 100644
--- a/src/daemon/org.projectatomic.rpmostree1.xml
+++ b/src/daemon/org.projectatomic.rpmostree1.xml
@@ -74,8 +74,6 @@
     </property>
   </interface>
 
-  <annotation name="org.qtproject.QtDBus.QtTypeName" value="QVariantMap"/>
-
   <interface name="org.projectatomic.rpmostree1.OS">
     <property name="BootedDeployment" type="a{sv}" access="read">
         <annotation name="org.qtproject.QtDBus.QtTypeName" value="QVariantMap"/>

[cgwalters]

Seems OK to me; a bit surprising that the a{sv} case requires annotation.

[travier]

I'll most likely do one small rename before we go with this (finding something better for: GetDeploymentsRpmDiffResult).

[cgwalters]

I'll most likely do one small rename before we go with this (finding something better for: GetDeploymentsRpmDiffResult).

Rename of what?

/test all

[travier]

I'll most likely do one small rename before we go with this (finding something better for: GetDeploymentsRpmDiffResult).

Rename of what?

There was a custom class specified in the annotation for the "result" received for some DBus calls. I switched it to a standard Qt class instead and it looks OK so far. I'm keeping this as a draft for now as I'm still working on the code using this and may find other improvements.

[jlebon]

Is the < expected here? Should > also be changed to >?

[travier]

I've tried using < and it breaks somewhere in the code generation. Using > is currently working. I've not tried using >.

[travier]

Rebased to trigger retest. I have not made any new changes on that one so far so I don't think there will be any soon even though I'm not done with rpm-ostree support in Discover. Let's get this in and we will see if more is needed later.

[cgwalters]

Sorry about missing this one in the review queue!

I think we really need to set up a periodic group "PR review backlog" thing.

[travier]

All good. Thanks!