Re: [Discuss-gnuradio] Error using block UHD AMsg Source

[Jose Torres Diaz]

Hi Josh, 

You mentioned that:

"the code to add a key/value actually returns a new dictionary and doesnt
change the old ones.

I know this pmt stuff is crazy. Why didnt we just use std::map? :-)"

I'm confused with this, because I created the dictionary and added the values. However, when I've checked the type, it shows that it is a dictionary (as I expected), but also it is always empty.

      pmt::pmt_t value_dict = pmt::pmt_make_dict();
      std::cout << std::endl << "Making a dictionary" << std::endl;
      pmt::pmt_dict_add(value_dict, TIME_KEY,  pmt::pmt_from_uint64(3200));
            this->post_msg(0, AMSG_KEY, value_dict, _id);
      std::cout << std::endl << "Posting downstream" << std::endl;

  std::cout << std::endl << "checking if null:" <<pmt::pmt_is_null(value_dict)<< std::endl; //Return true
  std::cout << std::endl << "checking if dict:" <<pmt::pmt_is_dict(value_dict)<< std::endl; //Return true
 
So, does it mean that I cannot access the values that are added into the dictionary "value_dict"?. If it is possible, how should I access them?.

Sort of related question:

Since this is dictionary, in order to debug my code I printed out the items, keys and values. I expected to see at least the items inside the dictionary, but again everything is empty. 

std::cout << std::endl << "Checking items ;" <<  pmt::pmt_dict_items(value_dict)<<std::endl;
std::cout << std::endl << "Checking keys ;" << pmt::pmt_dict_keys(value_dict)<<std::endl;
std::cout << std::endl << "Checking values;" << pmt::pmt_dict_values(value_dict)<<std::endl;

Thanks a lot for your help, 

Kind Regards, 

Jose

On Thu, Oct 11, 2012 at 2:43 PM, Josh Blum <address@hidden> wrote:

On 10/10/2012 07:16 PM, Jose Torres Diaz wrote:
> Hi Josh,
>

> I am very keen on this block (UHD_AMSG_SOURCE), because I saw that you post
> a dictionary downstream in this example. Thus, I would like to analyze how
> it works. I've tried to replicate your code in a different block, but I
> have some issues with the dictionary.
>
> Here is my summary:
>

> *Issue (1):* According to the wiki:

> http://gnuradio.org/redmine/projects/gnuradio/wiki/TypePMT#Dictionaries. If
> I want to set values in the dictionary, I should use
>
> pmt_dict_set(%parameters here)

Doesnt exist. I think it should be added. This version could add the
key/value to the existing dictionary. vs the pmt_dict_add which creates
a new one

>
> However, this command it doesn't work for me (as it is not part of pmt).
>
> The command that works without problem is:
>
> pmt_dict_add(%parameters here)
>
>

> *Issue (2):* I've set a dictionary with one value on it. (See the code

> below). However, it is always empty when I tried to check the items, keys
> or values:
>
> pmt::pmt_t value_dict=pmt::pmt_make_dict();
>     std::cout << std::endl << "Making a dictionary : " << value_dict
> <<std::endl;
>     pmt::pmt_dict_add (value_dict,blob_key,pmt::pmt_from_uint64(10));
>     std::cout << std::endl << "Adding the value, items : " <<
> pmt::pmt_dict_items(value_dict) <<std::endl;
>     std::cout << std::endl << "Adding the value, keys : " <<
> pmt::pmt_dict_keys(value_dict) <<std::endl;
>     std::cout << std::endl << "Adding the value, values : " <<
> pmt::pmt_dict_values(value_dict) <<std::endl;
>
> Results (always empty!):

the code to add a key/value actually returns a new dictionary and doesnt
change the old ones.

I know this pmt stuff is crazy. Why didnt we just use std::map? :-)

>
> Adding the value, items : ()
> Adding the value, keys : ()
> Adding the value, values : ()
>
>

> *Issue (3):* I post the dictionary downstream similar to the code in
> "UHD_AMSG_SOURCE":
>
> this->post_msg(0,value_dict,_msg.value,_id);

The key has to the a pmt that holds a string, the value does not have
the restriction.

>
> However, similarly to my previous post, the flow graph complains that this
> is a wrong type:
>
> thread[thread-per-block[8]: <gr_block msg_sourcer (9)>]:
> gr_block_detail::add_item_tag key: wrong_type : ()
>
> For these reasons, I would like to see how your code is working with these
> dictionaries.
>
> Thanks again,
>
> Jose
>
> On Thu, Oct 11, 2012 at 12:05 PM, Jose Torres Diaz <
> address@hidden> wrote:
>
>> Hi Josh,
>>
>> I've changed the .xml file. Also, I've checked the new top_block.py file
>> and it is generated as follows:
>>
>> self.extras_uhd_amsg_source_0 =
>> gr_extras.uhd_amsg_source(uhd.device_addr="172.17.3.140"
>>
>> However, it is still complaining when I run the flow graph:
>>
>>   File "/home/asrpuser/Desktop/Jose Torres Material /top_block.py", line 33
>>     self.extras_uhd_amsg_source_0 =
>> gr_extras.uhd_amsg_source(uhd.device_addr="add=172.17.3.140")
>> SyntaxError: keyword can't be an _expression_
>>
>> Thanks,
>>
>> Jose.
>>
>>
>>
>> On Thu, Oct 11, 2012 at 11:44 AM, Josh Blum <address@hidden> wrote:
>>
>>>
>>>
>>> On 10/10/2012 06:09 PM, Jose Torres Diaz wrote:
>>>> Hi Josh,
>>>>
>>>> Do you mean to change in the top_block.py?. I did this, but when I run
>>>> again the top_block.py, it changes back to:
>>>>
>>>> gr_extras.uhd_amsg_source(device_addr="add=172.17.3.140")
>>>>
>>>> even if I try:
>>>>
>>>> gr_extras.uhd_amsg_source(uhd.device_addr="add=172.17.3.140")
>>>>
>>>> Thanks again Josh,
>>>>
>>>
>>> I see. Its the xml definition. Too bad python/swig does get the
>>> constructor overloads that c++ sees.
>>>
>>> I pushed this change to the repo:
>>>> diff --git a/grc/extras_uhd_amsg_source.xml
>>> b/grc/extras_uhd_amsg_source.xml
>>>> index 5f8c261..5bb59f0 100644
>>>> --- a/grc/extras_uhd_amsg_source.xml
>>>> +++ b/grc/extras_uhd_amsg_source.xml
>>>> @@ -4,7 +4,7 @@
>>>>      <key>extras_uhd_amsg_source</key>
>>>>      <import>from gnuradio import uhd</import>
>>>>      <import>import gnuradio.extras as gr_extras</import>
>>>> -    <make>gr_extras.uhd_amsg_source(device_addr=$dev_addr)</make>
>>>> +    <make>gr_extras.uhd_amsg_source(uhd.device_addr($dev_addr))</make>
>>>>      <param>
>>>>          <name>Device Addr</name>
>>>>          <key>dev_addr</key>
>>>
>>> -josh
>>>
>>>> Regards,
>>>>
>>>> Jose
>>>>
>>>> On Thu, Oct 11, 2012 at 11:24 AM, Josh Blum <address@hidden> wrote:
>>>>
>>>>>
>>>>>
>>>>> On 10/10/2012 05:51 PM, Jose Torres Diaz wrote:
>>>>>> Hi All,
>>>>>>
>>>>>> I'm checking the functionality of the block "UHD AMsg Source" using
>>> the
>>>>>> USRP N210. The device is correctly detected as you can see:
>>>>>>
>>>>>> uhd_find_devices
>>>>>> linux; GNU C++ version 4.6.1; Boost_104601;
>>> UHD_003.004.003-221-g9d6f9492
>>>>>>
>>>>>> --------------------------------------------------
>>>>>> -- UHD Device 0
>>>>>> --------------------------------------------------
>>>>>> Device Address:
>>>>>>     type: usrp2
>>>>>>     addr: 172.17.3.140
>>>>>>     name:
>>>>>>     serial: E0R17TCUP
>>>>>>
>>>>>> I've attached a screen-shot of the GRC file. When I run the flow
>>> graph, I
>>>>>> get the following error:
>>>>>>
>>>>>> Traceback (most recent call last):
>>>>>>   File "/home/asrpuser/Desktop/Jose Torres Material /top_block.py",
>>> line
>>>>>> 54, in <module>
>>>>>>     tb = top_block()
>>>>>>   File "/home/asrpuser/Desktop/Jose Torres Material /top_block.py",
>>> line
>>>>>> 33, in __init__
>>>>>>     self.extras_uhd_amsg_source_0 =
>>>>>> gr_extras.uhd_amsg_source(device_addr="add=172.17.3.140")
>>>>>
>>>>> try gr_extras.uhd_amsg_source(uhd.device_addr("addr=172.17.3.140"))
>>>>>
>>>>>>   File
>>>>>>
>>> "/usr/local/lib/python2.7/dist-packages/gnuradio/extras/extras_swig.py",
>>>>>> line 5384, in make
>>>>>>     return _extras_swig.uhd_amsg_source_make(*args, **kwargs)
>>>>>> TypeError: in method 'uhd_amsg_source_make', argument 1 of type
>>>>>> 'uhd::device_addr_t const &'
>>>>>>
>>>>>> Anyone has seen an error like this before?, How can  I
>>> solve/workaround
>>>>>> this error?
>>>>>>
>>>>>> Thanks a lot for your help,
>>>>>>
>>>>>> Jose.
>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Discuss-gnuradio mailing list
>>>>>> address@hidden
>>>>>> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Discuss-gnuradio mailing list
>>>>> address@hidden
>>>>> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>>>>>
>>>>
>>>
>>
>>
>>
>>
>