[ros-users] rospkg API review (REP 114)

[Tully Foote]

+1

Tully

On Tue, Oct 11, 2011 at 10:04 AM, Ken Conley <kwc at willowgarage.com> wrote:

>
>
> On Mon, Oct 10, 2011 at 2:40 PM, Ken Conley <kwc at willowgarage.com> wrote:
>
>> Thanks to all who provided comments on the rospkg API.  Now that the
>> comment deadline has passed, I have added a summary to the review page of
>> proposed/adopted changes:
>>
>> http://www.ros.org/wiki/rospkg/Reviews/2011-09%20rospkg#Meeting_agenda
>>
>> I don't think it will be necessary to hold a formal meeting given that the
>> changes are fairly simple.
>>
>> Most of the changes are minor or additions of missing APIs.   The main,
>> breaking change being proposed is:
>>
>> rospkg.RosPack(ros_root='foo', ros_package_path='path1:path2')
>>
>> to
>>
>> rospkg.RosPack(path=['foo', 'path1:path2'])
>>
>
> Sorry for the bad cut-and-paste, this should be:
>
> rospkg.RosPack(path=['foo', 'path1', 'path2'])
>
>
>>
>> As well a similar changes to remove explicit knowledge of ROS_ROOT and
>> ROS_PACKAGE_PATH from the RosPack/RosStack APIs.  The motivation for this
>> change is forwards compatibility and versatility (see below).
>>
>> This change was proposed by Tully and I'm inclined to accept it barring
>> any objections (which you may not in the normal "+1", "-1" style).
>>
>>  - Ken
>>
>>
>> Forwards compatibility:
>>
>> ROS_ROOT, as opposed to just a generic ROS_PACKAGE_PATH, is mainly
>> necessary for constructing a system PATH; future changes will make this no
>> longer necessary, and thus make ROS_ROOT vestigial.
>>
>> Versatility:
>>
>> By accepting a generic 'path' argument, the RosPack class is decoupled
>> from any particular idea of environment variables, etc... It just knows
>> there is an ordered set of paths it operates on.
>>
>>
>>
>> On Thu, Oct 6, 2011 at 11:27 AM, Ken Conley <kwc at willowgarage.com> wrote:
>>
>>> On Thu, Oct 6, 2011 at 3:18 AM, Dirk Thomas <mail at dirk-thomas.net>wrote:
>>>
>>>>  Good suggestion.  As this is the behavior of rospack, I changed rospkg
>>>>> to match it by default.  You can get the raw export by adding a
>>>>> "convert=False" arg to the get_export() call instead.
>>>>>
>>>>
>>>> great, thank you.
>>>>
>>>>
>>>>  A minor cosmetic issue is the printing of "READ_CACHE manifest.xml" in
>>>>>> rospkg/rospack.py:188.
>>>>>> I guest this output will be removed in the next release?
>>>>>>
>>>>>
>>>>> Yes, this has already been removed.  I pushed 0.2.3 with the fix as
>>>>> well as updated get_export() API.
>>>>>
>>>>
>>>> It would be useful if the version number of rospkg could be determined.
>>>> This is not yet possible, right?
>>>> So could you also add a function which returns the version number of
>>>> rospkg?
>>>>
>>>
>>> Great idea.  Added in the next version:
>>>
>>> $ python -c "import rospkg; print rospkg.__version__"
>>> 0.2.4
>>>
>>>
>>>
>>>>
>>>> Thank you,
>>>>
>>>> Dirk
>>>> ______________________________**_________________
>>>> ros-users mailing list
>>>> ros-users at code.ros.org
>>>> https://code.ros.org/mailman/**listinfo/ros-users<https://code.ros.org/mailman/listinfo/ros-users>
>>>>
>>>
>>>
>>
>
> _______________________________________________
> ros-users mailing list
> ros-users at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>
>


-- 
Tully Foote
Systems Engineer
Willow Garage, Inc.
tfoote at willowgarage.com
(650) 475-2827
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ros.org/pipermail/ros-users/attachments/20111011/d8a8eaa9/attachment-0004.html>