Attention is currently required from: pespin.
kirr has posted comments on this change by kirr. ( https://gerrit.osmocom.org/c/osmocom-bb/+/39537?usp=email )
Change subject: trx_toolkit/*: Try to avoid copying burst data where possible ......................................................................
Patch Set 2:
(1 comment)
File src/target/trx_toolkit/data_msg.py:
https://gerrit.osmocom.org/c/osmocom-bb/+/39537/comment/729d63f9_41a26f6c?us... : PS2, Line 336: burst = memoryview(burst)[:EDGE_BURST_LEN]
Just speculating here, but with the new version you posted you are still potentially creating a copy […]
Could you please clarify? Do you mean here that we create a copy via `self.burst = bytearray(burst)` later? Or do you mean that we create a copy somewhere else?
The copy in `self.burst = bytearray(burst)` is necessarry in current data model where we want .burst to be a bytearray or bytes. In theory we could keep it as memoryview as well, but memoryview does not have .translate method and so then conversion from unpacked bits to soft bits will become much slower again.