Python3 only

[SamuelDeleglise]

Uses coroutine to replace the low-level futures/callbacks previously used for asynchronous acquisition. Pro: improved code readbility/Cons: incompatible with python<3.5.

There might be some places where callbacks are still there (for instance the LoopModule still exists and is implemented in the old fashion, or the PyrplWidget status_bar which changes color after some time was not touched).

Lockbox still needs more unittests and probably an API refresh (see https://github.com/lneuhaus/pyrpl/issues/349#issuecomment-445477673)

[codecov-io]

Codecov Report

Merging #355 into master will decrease coverage by 64.81%. The diff coverage is 7.62%.

@@             Coverage Diff             @@
##           master     #355       +/-   ##
===========================================
- Coverage      77%   12.19%   -64.82%     
===========================================
  Files          97       93        -4     
  Lines       11279    10566      -713     
===========================================
- Hits         8685     1288     -7397     
- Misses       2594     9278     +6684

Impacted Files	Coverage Δ
pyrpl/test/test_validate_and_normalize.py	0% <ø> (-100%)	:arrow_down:
pyrpl/widgets/spinbox.py	22.67% <ø> (-50.19%)	:arrow_down:
pyrpl/hardware_modules/iir/iir_theory.py	0% <ø> (-70%)	:arrow_down:
pyrpl/widgets/module_widgets/base_module_widget.py	25.4% <ø> (-43.25%)	:arrow_down:
pyrpl/widgets/module_widgets/spec_an_widget.py	16.44% <ø> (-51.68%)	:arrow_down:
pyrpl/widgets/module_widgets/scope_widget.py	12.65% <ø> (-62.71%)	:arrow_down:
pyrpl/widgets/startup_widget.py	0% <ø> (-89.12%)	:arrow_down:
pyrpl/hardware_modules/asg.py	0% <ø> (-83.23%)	:arrow_down:
pyrpl/test/test_widgets/test_startup_widget.py	0% <ø> (-100%)	:arrow_down:
pyrpl/hardware_modules/iir/__init__.py	0% <ø> (-100%)	:arrow_down:
... and 134 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 0ee0f44...9144d6c. Read the comment docs.

[lneuhaus]

Just merged "develop-0.9.3" into master. To do before merging this branch:

• merge master into this branch and fix conflicts
• bump version from 0.9.5 to 1.0.0
• add list of changes to changelog.md
• pass tests on Python 3.7 and 3.8 (Python 3.6 and 3.9 are optional I guess)

By when do you think this can be done @SamuelDeleglise? Depending on the timescale I could skip building binaries for the current master.

[lneuhaus]

pinging @SamuelDeleglise again - otherwise I might go ahead and clean this up for merging.

[lneuhaus]

Hi @SamuelDeleglise, if you want let's try to merge this branch into master. Am I fine to click "update branch" here?

[lneuhaus]

Also noting that we should ideally remove all the loop=LOOP constructs from async_utils.py, since they are deprecated and break this version on Python>=3.10. Newer versions of Python should figure out automatically which loop to use.

[SamuelDeleglise]

In principle, you can click to merge: we are all using this branch in the lab. However, I am not 100 % sure what's the status of the unit-tests on this branch ... Let me know if you think this is super important to have the unit-tests checked, I wil ltry to do that ASAP, but there are a few other stuffs I would like to push forward regarding pyrpl soon:

• The final version of the paper is almost fully drafted, you'll get a version to re-read soon (it's now on a different overleaf account).
• Tristan was able to compile the verilog code on Vivado 2022... It would be nice to upgrade towards this as well at some point

--> Could you either add Tristan to the pyrpl repos (tristan-briant) or give me admin rights ? --> Also I have been having a hard time reaching you via e-mail lately, are you still checking this readthedoc e-mail adress or should I use another one ?

[lneuhaus]

Sorry for my late reply, missed your message. I tried this yesterday and found this version to work better than master, so am happy to merge. For that same reason, I think it's better to merge this rather sooner and fix the tests later (master is more broken than this).

I tried to make you an admin but seems like that doesn't work with free github accounts? What's Tristan's username? I sent you an email so you have mine.

Thanks for taking time to write the paper.

[lneuhaus]

I kept this branch, but I suggest we keep it around in case anyone is so used to it after so many years. In that sense, it would also be fine to continue working on it, and opening a new PR with the next changes from it. Thanks for all the work that went into this @SamuelDeleglise .

I'll do my best to try adding other PRs with

• fixed tests (I think I'll stop assuming a redpitaya is connected to the internet for tests, and instead rely on doing tests on a local setup before merging), maybe move the pure software tests to github actions
• Python 3.10+ support. This is a tricky subject, as in my experience we'll have to modify lots of dependencies (the main thing hard to get right is support for event loops running both in pure python and from jupyter). What Python are you running on, and are you planning to upgrade at some point @SamuelDeleglise ?