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

Nobuhiko Tanibata nobuhiko_tanibata at xddp.denso.co.jp
Fri Mar 14 23:58:30 PDT 2014


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.

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