So, I was reading someone's python block that was coded as a general block, and wanted to work in blocks of X elements as input. If it got less than X, it just said I consumed nothing, and just waited to be called back repeatedly till eventually it was given X or more elements. Note X wasn't a particularly big number (I think 128?).
Seemed to work, but seemed perhaps risky to assume you'd eventually get X elements? Yet this is a streaming environment, sooner or later more data comes?
I suppose another approach would be to have a local buffer in your block and just append to that buffer till you get X or more elements in the buffer, but then you have to do more copying and bookkeeping instead of just putting that burden onto the run-time like my previous example did.
Another approach might have used set_history() to X so that the previous X elements would always be presented so you would not have to buffer locally (?), but you would have to do the bookkeeping to fetch blocks of size X.
I think I'm going to experiment a bunch with these different approaches, but would appreciate some guidance on best practices...
Regards,
Dave