Open ngr-francesco opened 11 months ago
Thank you for pointing that inconsistencies out. I guess nothing speaks against adding a name parameter to the RingBuffer and setting it for the AbstractProcess which always has it anyways. If you want that feature, feel free to start a PR.
Okay, I'll get to it in the next days
User story
I am currently working on the Brian2Lava project to bring together the UX of Brian with the backend of Lava. For debugging purposes we name every process we create with the same naming convention used in Brian. While implementing a new feature which requires the initialization of a
RingBuffer
, I noticed that theRingBuffer
__init__
method does not accept aname
argument. I suspect this might be the case for otherAbstractProcesses
, too. At the same time by printingring_buffer.name
returns a name using the conventionProcess_N
, which implies that as expected the attribute is present, but the user is simply not allowed to set it at initialization. Even though a workaround is easily implemented, I believe adding this feature should be relatively straightforward, and it would be useful for our team. As a separate small question/comment, why are there two versions ofRingBuffer
(in the io.source and io.sink modules) with different use cases and not two classes with different names, since they have different jobs?Conditions of satisfaction
name
argument.Acceptance tests