[PATCH weston-ivi-shell v3 02/10] ivi application protocol:

Pekka Paalanen ppaalanen at gmail.com
Mon Mar 17 00:15:17 PDT 2014


On Mon, 17 Mar 2014 13:48:45 +0900
Nobuhiko Tanibata <nobuhiko_tanibata at xddp.denso.co.jp> wrote:

> 2014-03-17 10:24 に Nobuhiko Tanibata さんは書きました:
> > 2014-03-15 15:58 に Nobuhiko Tanibata さんは書きました:
> >> 2014-03-14 23:16 に Pekka Paalanen さんは書きました:
> >>> On Wed, 12 Mar 2014 23:59:33 +0900
> >>> Nobuhiko Tanibata <NOBUHIKO_TANIBATA at xddp.denso.co.jp> wrote:
> >>> 
> >>>> Add interface ivi_application, which creates ivi_surface objects 
> >>>> tied
> >>>> to a given wl_surface with a given id. The given id can be used in a
> >>>> shell to identify which application is assigned to a wl_surface and
> >>>> layout the surface wherever the shell wants. ivi_surface objects can
> >>>> be used to receive status of wl_surface in the scenegraph of the
> >>>> compositor.
> >>>> 
> >>>> Signed-off-by: Nobuhiko Tanibata 
> >>>> <NOBUHIKO_TANIBATA at xddp.denso.co.jp>
> >>>> ---
> >>>> 
> >>>> Changes for v2:
> >>>>    - Rename "error" to "warning" because meaning of "error" in 
> >>>> wayland is fatal.
> >>>> 
> >>>> Changes for v3:
> >>>>    - Move "warning" from ivi_application to ivi_surface.
> >>>>    - Squash Makefile.
> >>>>    - Add description to ivi_surface:destroy.
> >>>>    - Update description of ivi_application:surface_create.
> >>>> 
> >>>>  protocol/Makefile.am         |  3 +-
> >>>>  protocol/ivi-application.xml | 96 
> >>>> ++++++++++++++++++++++++++++++++++++++++++++
> >>>>  2 files changed, 98 insertions(+), 1 deletion(-)
> >>>>  create mode 100644 protocol/ivi-application.xml
> >>>> 
> >>>> diff --git a/protocol/Makefile.am b/protocol/Makefile.am
> >>>> index 5e331a7..9913f16 100644
> >>>> --- a/protocol/Makefile.am
> >>>> +++ b/protocol/Makefile.am
> >>>> @@ -8,7 +8,8 @@ protocol_sources =				\
> >>>>  	text-cursor-position.xml		\
> >>>>  	wayland-test.xml			\
> >>>>  	xdg-shell.xml				\
> >>>> -	scaler.xml
> >>>> +	scaler.xml                              \
> >>>> +	ivi-application.xml
> >>>> 
> >>>>  if HAVE_XMLLINT
> >>>>  .PHONY: validate
> >>>> diff --git a/protocol/ivi-application.xml 
> >>>> b/protocol/ivi-application.xml
> >>>> new file mode 100644
> >>>> index 0000000..8f5c23d
> >>>> --- /dev/null
> >>>> +++ b/protocol/ivi-application.xml
> >>>> @@ -0,0 +1,96 @@
> >>>> +<?xml version="1.0" encoding="UTF-8"?>
> >>>> +<protocol name="ivi_application">
> >>>> +
> >>>> +    <copyright>
> >>>> +    Copyright (C) 2013 DENSO CORPORATION
> >>>> +    Copyright (c) 2013 BMW Car IT GmbH
> >>>> +
> >>>> +    Permission is hereby granted, free of charge, to any person 
> >>>> obtaining a copy
> >>>> +    of this software and associated documentation files (the 
> >>>> "Software"), to deal
> >>>> +    in the Software without restriction, including without 
> >>>> limitation the rights
> >>>> +    to use, copy, modify, merge, publish, distribute, sublicense, 
> >>>> and/or sell
> >>>> +    copies of the Software, and to permit persons to whom the 
> >>>> Software is
> >>>> +    furnished to do so, subject to the following conditions:
> >>>> +
> >>>> +    The above copyright notice and this permission notice shall be 
> >>>> included in
> >>>> +    all copies or substantial portions of the Software.
> >>>> +
> >>>> +    THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, 
> >>>> EXPRESS OR
> >>>> +    IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF 
> >>>> MERCHANTABILITY,
> >>>> +    FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO 
> >>>> EVENT SHALL THE
> >>>> +    AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES 
> >>>> OR OTHER
> >>>> +    LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, 
> >>>> ARISING FROM,
> >>>> +    OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER 
> >>>> DEALINGS IN
> >>>> +    THE SOFTWARE.
> >>>> +    </copyright>
> >>>> +
> >>>> +    <interface name="ivi_surface" version="1">
> >>>> +        <description summary="application interface to surface in 
> >>>> ivi compositor"/>
> >>>> +
> >>>> +        <request name="destroy" type="destructor">
> >>>> +            <description summary="destroy ivi_surface">
> >>>> +                This removes link from surface_id to wl_surface. 
> >>>> However it doesn't
> >>>> +                remove internal properties, e.g. position, 
> >>>> visibility, and so on, which
> >>>> +                is set to the surface_id. This means when some 
> >>>> issues happen on clients
> >>>> +                and a ivi_surface is destroyed, it can use previous 
> >>>> properties immediately
> >>>> +                without setting it again if it restarts and 
> >>>> attaches new wl_surface to
> >>>> +                the same surface_id.
> >>> 
> >>> This keeping of properties... why even mention it here? They are
> >>> nothing this client can affect, are they?
> >>> 
> >>> Otherwise I would ask, how would the client know if the properties 
> >>> are
> >>> already set, or if it needs to set them again, or if the client
> >>> actually needs to know some of them to function correctly.
> >>> 
> >>> Not a big deal.
> >> 
> >> Hi pq,
> >> 
> >> Yes, it doesn't affect client behavior so I will remove description
> >> after "However...".
> >> The decision shall be done in server side. So as you say, it should
> >> not be mentioned here.
> >> 
> >>> 
> >>>> +            </description>
> >>>> +        </request>
> >>>> +
> >>>> +        <event name="visibility">
> >>>> +            <description summary="visibility of surface in ivi 
> >>>> compositor has changed">
> >>>> +                The new visibility state is provided in argument 
> >>>> visibility.
> >>>> +                If visibility is 0, the surface has become 
> >>>> invisible.
> >>>> +                If visibility is not 0, the surface has become 
> >>>> visible.
> >>>> +            </description>
> >>>> +            <arg name="visibility" type="int"/>
> >>>> +        </event>
> >>>> +
> >>>> +        <enum name="warning_code">
> >>>> +            <description summary="possible warning codes returned 
> >>>> by ivi compositor">
> >>>> +                These warning codes define all possible warning 
> >>>> codes returned by ivi compositor
> >>>> +                on server-side warnings.
> >>>> +                invalid_wl_surface: invalid wl_surface is set. This 
> >>>> happen if wl_surface is destroy before this.
> >>> 
> >>> I guess you mean, that invalid_wl_surface is emitted, if the 
> >>> wl_surface
> >>> is destroyed before the ivi_surface, right?
> >> 
> >> Yes, you are right.
> >>> 
> >>> Usually we deal with these things by saying the ivi_surface would 
> >>> just
> >>> become inert, but I assume you really want the notification here.
> >>> 
> >>> Is invalid_wl_surface also for the case when the wl_surface already 
> >>> has
> >>> an ivi_surface associated?
> >> 
> >> In the above case, a code "surface_id_in_use" is used. But, I shall
> >> add "wl_surface_in_use" to the warnings.
> >> 
> >>> 
> >>>> +                surface_id_in_use: surface_id is already assigned 
> >>>> by another application.
> >>>> +            </description>
> >>>> +            <entry name="invalid_wl_surface" value="1" 
> >>>> summary="wl_surface is invalid"/>
> >>>> +            <entry name="surface_id_in_use" value="2" 
> >>>> summary="surface_id is in use and can not be shared"/>
> >>>> +        </enum>
> >>>> +
> >>>> +        <event name="warning">
> >>>> +            <description summary="server-side warning detected">
> >>>> +                The ivi compositor encountered warning while 
> >>>> processing a request by this
> >>>> +                application. The warning is defined by argument 
> >>>> warning_code and optional
> >>>> +                warning_text. If the warning is detected, client 
> >>>> shall destroy the ivi_surface
> >>>> +                object.
> >>> 
> >>> You still need to specify how the server handles this ivi_surface
> >>> object after sending the warning, but before it is destroyed. Does 
> >>> the
> >>> server ignore all requests referring to this ivi_surface, or the ID? 
> >>> Is
> >>> the ID immediately free again, or does the ivi_surface need to be
> >>> destroyed before the ID becomes available again?
> >> 
> >> Yes, the server ignore all requests.
> >> In case of "surface_id_in_use", ivi_surface, e.g. "A" doesn't have to
> >> be destroyed to use surface_id. the surface_id is tied only to the
> >> other ivi_surface, e.g. "B" already. If the other ivi_surface "B" is
> >> destroyed, client can use surface_id without destruction of
> >> ivi_surface "A" because "A" doesn't have any link to the surface_ID.
> >> 
> >> I will add more description of correct behavior here.
> >> 
> >>> 
> >>>> +            </description>
> >>>> +            <arg name="warning_code" type="int"/>
> >>>> +            <arg name="warning_text" type="string" 
> >>>> allow-null="true"/>
> >>>> +        </event>
> >>>> +
> >>>> +    </interface>
> >>>> +
> >>>> +    <interface name="ivi_application" version="1">
> >>>> +        <description summary="interface for ivi applications to use 
> >>>> ivi compositor features"/>
> >>>> +
> >>>> +        <request name="surface_create">
> >>>> +            <description summary="create ivi_surface with numeric 
> >>>> ID in ivi compositor">
> >>>> +                surface_create will create a interface:ivi_surface 
> >>>> with numeric ID; surface_id in
> >>>> +                ivi compositor. These surface_ids are defined as 
> >>>> unique in the system to identify
> >>>> +                it inside of ivi compositor. The ivi compositor 
> >>>> implements business logic how to
> >>>> +                set properties of the surface with surface_id 
> >>>> according to status of the system.
> >>>> +                E.g. a unique ID for Car Navigation application is 
> >>>> used for implementing special
> >>>> +                logic of the application about where it shall be 
> >>>> located. Created ivi_surface
> >>>> +                notifies warnings when following situation happens,
> >>>> +                - Invalid wl_surface is set. This happen if 
> >>>> wl_surface is destroy before this.
> >>>> +                - surface_id is already assigned by another 
> >>>> application.
> >>>> +            </description>
> >>>> +            <arg name="id_surface" type="uint"/>
> >>>> +            <arg name="surface" type="object" 
> >>>> interface="wl_surface"/>
> >>>> +            <arg name="id" type="new_id" interface="ivi_surface"/>
> >>> 
> >>> Does this give the surface a role? E.g. what should happen if a 
> >>> client
> >>> registers the same wl_surface as both an ivi_surface and a pointer
> >>> image (wl_pointer.set_cursor).
> >>> 
> >>> If your weston implementation sets weston_surface::configure, it is a
> >>> strong indication that this gives a role, which excludes all other
> >>> roles. IOW, this request should fail, if weston_surface::configure is
> >>> already set, so you need a warning code for it.
> >>> 
> >>> You should probably look at all interfaces, where a wl_surface can be
> >>> an argument for a request, and check if those interfaces can exist in
> >>> an IVI environment, and if they can, how they interoperate with a
> >>> wl_surface that has a ivi_surface.
> >> 
> >> Good comments. I will check them and come back immediately.
> >> Hi pq,
> > 
> > I am investigating the above point in wayland.xml.
> > 
> > -wl_data_device::start_drag
> >   This may be applied for ivi system as well. I think this is also
> > used with wl_pointer.set_cursor.
> > Please give me your suggestion how to use the above two interfaces for
> > one wl_surface.
> > I think it is avoided by calling them sequentially from client. E.g.
> > after one configure of set_sursor is done and then another configure
> > of start_drag shall be called.
> > I think similarly configure of ivi_application is one called, it can
> > be set to another callback from by another request. Shall I write them
> > in protocol as manner?
> 
> Hi pq,
> 
> I mis-understand this case. My clarification is that if client want to 
> use a wl_surface for wl_subsurface and wl_data_device::start_drag??

It cannot.

Sub-surface, drag icon, pointer cursor, and a shell surface are all
surface roles and so they are exclusive; A surface can only be at
most one of those at a time, and quite often also for the whole
wl_surface lifetime as it does not make much sense to change the role of
a surface after the initial assignment.

A role gives wl_surface a meaning. Without a role, the compositor will
have no idea how to present the wl_surface, and hence wl_surfaces
without a role will never be visible.


Thanks,
pq

> > 
> > -wl_shell::get_shell_surface
> > -wl_shell_surface::set_transient
> > -wl_shell_surface::set_popup
> >   wl_shell is for desktop style shell so above three request is not
> > used with ivi_application.xml.
> > 
> > -wl_pointer::set_cursor
> >   This may be applied for ivi system as well.
> > 
> > -wl_subcompositor::get_subsurface
> > -wl_subsurface::place_above
> > -wl_subsurface::place_below
> >   wl_surface shall be set to child surface of a wl_surface which is
> > set to ivi_surface. If client set a wl_surface which is already set to
> > wl_subsurface, it should fail. I will take core of it.
> > 
> > BTW, thank you for many comment, this is very useful for me.
> > 
> > Thanks,
> > Nobuhiko
> > 
> >> Thanks,
> >> Nobuhiko
> >>> 
> >>>> +        </request>
> >>>> +
> >>>> +    </interface>
> >>>> +
> >>>> +</protocol>
> >>> 
> >>> This is a very simple interface, but there are many details to
> >>> get right anyway. You are doing good progress. :-)
> >>> 
> >>> 
> >>> Thanks,
> >>> pq



More information about the wayland-devel mailing list