Re: [Opensync-devel] Preparing for OpenSync API 0.40 soft-freeze

[Chris F. <cd...@fo...>]

On Thu, Sep 17, 2009 at 09:17:03PM -0400, Chris Frey wrote:
>  OSyncXMLField *convert_vcal_rrule_to_xml(OSyncXMLFormat *xmlformat, VFormatAttribute *attr, const char *rulename, OSyncError **error)
> @@ -305,6 +312,7 @@ static char *convert_rrule_vcal_until(const char *until_utc)
>  {
>  	int offset = 0; 
>  	char *until = NULL;
> +	OSyncError *error = NULL;
>  
>  

[...]

> +	until = osync_time_vtime2localtime(until_utc, offset, &error);
> +
> +	if( error != NULL ) {
> +		osync_trace(TRACE_ERROR, "%s: %s" , __func__, osync_error_print(&error));
> +		osync_error_unref(&error);
> +	}


Just to refresh my memory, I know that in the usage above, I'm responsible
for unref'ing the error, since I'm the only one that sees it.

But if an OSyncError **error is passed in to my function, am I still
responsible for unref'ing it?  This almost seems like the responsibility
of the calling code, but in the sample plugin, the error is unref'd
even inside get_sync_info().  I don't know why the pointer is being passed
in if I'm responsible for unref'ing it.

Thanks,
- Chris