Fix setting start_time having no effect in Python bindings #665
Conversation
|
CLA Assistant Lite bot All contributors have signed the CLA ✍️ ✅ |
|
I have read the CLA Document and I hereby sign the CLA |
|
recheck |
|
Many thanks for the find! Can you please use an email address that matches what you have in your account? |
|
recheck |
1 similar comment
|
recheck |
…pec and stream_now when using Python bindings.
zolveuran
force-pushed
the
uran/fix_python_start_transmit_start_time
branch
from
280ae01 to
e0c5b6e
Compare
|
recheck |
|
@mbr0wn I tried to add the other e-mail address as a secondary, but that didn't work. Tried to amend the original commit to update the e-mail address, but still no luck. Maybe I'm not able to trigger a recheck? Can you have another try? Would also like to comment that the CLA is creating a somewhat of a barrier towards contributing - even for minor issues. Also, as the CONTRIBUTING.md document points out the CLA should only be required for "non-trivial contribution". I'd say this commit is fairly trivial and should not require the CLA to be signed. |
mbr0wn
added
the
Fixed Internally
|
This is percolating our internal CI and review pipelines. It might get closed before the changeset goes live on public master. |
| @@ -81,6 +81,7 @@ def _start_stream(streamer): | |||
| if not stream_cmd.stream_now: | |||
| if start_time is not None: | |||
| stream_cmd.time_spec = start_time | |||
| stream_cmd.stream_now = False | |||
There was a problem hiding this comment.
This line is not necessary (see the if clause 3 lines above, it's a requirement for coming here in the first place).
Pull Request Details
Description
Transmission starts immediately even though a
start_timeis provided when usingMultiUSRP.send_waveform()in the Python bindings. Caused bymetadata.has_time_specnot being set to True whenmetadata.time_specis set.Suspect the same issue is present for
recv_num_samps, wherestream_cmd.stream_now = Falseis missing whenstream_cmd.time_spec = start_time.Related Issue
Fixes #664
Which devices/areas does this affect?
Tested on USRP 2954.
Testing Done
Tested by setting
time_spec = usrp.get_time_now().get_real_secs() + 10and pass it toMultiUSRP.send_waveform().Am currently not able to test
recv_num_samps()due to no physical access to the hardware at the moment, but according to the documentationstream_cmd.stream_nowshould be set toFalse. Would be great if someone could give that a quick test as well.Checklist
MPM compat, noc_shell, specific RFNoC block, ...)