Incorrect documentation for C code attaching to events

Submitted by snrkiwi on Fri, 2009-03-13 13:18

The online doc's indicate that this should work

 Handle h
    = a_task.events()->setupConnection("Event")
                 .callback( &r, &Reactor::react_callback )
                 .handle();
  assert( h.ready() );

but it *ONLY* works if you do

 
events()->addEvent(&evt, "Event");

but *NOT* if you do

 
events()->addEvent(&evt);

Why is this different? The first case adds to "this->mhooks[]" however the second case does not.

If this is not a bug, then we should update section "3.13.1. Adding Event" of "orocos-components-manual.html" at the least.
S

‹ [Bug 631] New: KDL does not support OmniORB Orocos.org server down ›
»

Incorrect documentation for C code attaching to events

Submitted by sspr on Mon, 2009-03-16 09:50.

On Friday 13 March 2009 14:22:41 kiwi [dot] net [..] ... wrote:
> The online doc's indicate that this should work
> 

>  Handle h
>     = a_task.events()->setupConnection("move")
>                  .callback( &r, &Reactor::react_callback )
>                  .handle();
>   assert( h.ready() );
>

>
> but it *ONLY* works if you do
> 
> events()->addEvent(&evt, "Event");
>

>
> but *NOT* if you do
> 
> events()->addEvent(&evt);
>

>
> Why is this different? The first case adds to "this->mhooks[]" however the
> second case does not.
>
> If this is not a bug, then we should update section "3.13.1. Adding Event"
> of "orocos-components-manual.html" at the least. S

It's the infamous 'adding with a description' is 'adding with extra features'
behaviour (there was a thread about this some week ago, wrt rtt 2.0 & data
flow). The 'setupConnection' function uses the same infrastructure as scripting
to setup an event connection. So if you added an event with addEvent( &evt ),
it's not available to scripting, and not available to that part of the
interface.

This needs to be documented clearly indeed.

Peter

»

Incorrect documentation for C code attaching to events

Submitted by snrkiwi on Mon, 2009-03-16 16:10.

On Mar 16, 2009, at 05:44 , Peter Soetens wrote:
> On Friday 13 March 2009 14:22:41 kiwi [dot] net [..] ... wrote:
>> The online doc's indicate that this should work
>> 

>> Handle h
>>    = a_task.events()->setupConnection("move")
>>                 .callback( &r, &Reactor::react_callback )
>>                 .handle();
>>  assert( h.ready() );
>>

>>
>> but it *ONLY* works if you do
>> 
>> events()->addEvent(&evt, "Event");
>>

>>
>> but *NOT* if you do
>> 
>> events()->addEvent(&evt);
>>

>>
>> Why is this different? The first case adds to "this->mhooks[]"
>> however the
>> second case does not.
>>
>> If this is not a bug, then we should update section "3.13.1. Adding
>> Event"
>> of "orocos-components-manual.html" at the least. S
>
> It's the infamous 'adding with a description' is 'adding with extra
> features'
> behaviour (there was a thread about this some week ago, wrt rtt 2.0
> & data
> flow). The 'setupConnection' function uses the same infrastructure
> as scripting
> to setup an event connection. So if you added an event with
> addEvent( &evt ),
> it's not available to scripting, and not available to that part of the
> interface.

But this simply wasn't available to C++ code. I wasn't trying to use
it within a script.

> This needs to be documented clearly indeed.

Aye.

Cheers
S

»

Incorrect documentation for C code attaching to events

Submitted by snrkiwi on Fri, 2009-03-13 13:25.

The online doc's indicate that this should work

 Handle h
    = a_task.events()->setupConnection("move")
                 .callback( &r, &Reactor::react_callback )
                 .handle();
  assert( h.ready() );

but it *ONLY* works if you do

 
events()->addEvent(&evt, "Event");

but *NOT* if you do

 
events()->addEvent(&evt);

Why is this different? The first case adds to "this->mhooks[]" however the second case does not.

If this is not a bug, then we should update section "3.13.1. Adding Event" of "orocos-components-manual.html" at the least.
S

»