Speed up Signals when there are no receivers.

[pfreixes]

What do these changes do?

Reduce the footprint of the send() method by a 10% when there are
no receivers connected.

Before

$ python ../test_aiohttp_signals.py
Signals per second: 2296186.5549501474

After

$ python ../test_aiohttp_signals.py
Signals per second: 2759791.7854124783

Script used [1]

[1] https://gist.github.com/pfreixes/baea4ecc7b5380d534d60a8c1e8da32f

Are there changes in behavior for the user?

No

Checklist

• I think the code is well written
• Unit tests for the changes exist
• Add a new news fragment into the changes folder
  • name it <issue_id>.<type> for example (588.bug)
  • if you don't have an issue_id change it to the pr id after creating the pr
  • ensure type is one of the following:
    • .feature: Signifying a new feature.
    • .bugfix: Signifying a bug fix.
    • .doc: Signifying a documentation improvement.
    • .removal: Signifying a deprecation or removal of public API.
    • .misc: A ticket has been closed, but it is not of interest to users.
  • Make sure to use full sentences with correct case and punctuation, for example: "Fix issue with non-ascii contents in doctest text files."

[kxepal]

Nice! Since this is a FrozenList, his length is constant. What if you compute length once at __init__ and return that static value from __len__ ?

Update: I always forgot that it's not frozen like frozenset. Ok, what if do the same - compute length once on freeze while keep it dynamic while it's not frozen?

[pfreixes]

From what I've seen the freeze is done explicitly by the developer at some moment, would be needed calculate it at the moment of the freeze() call. But, seeing the Signal implementation seems that this characteristic is not really used to determine if a Signal can be used or not, looks like that the freeze only avoids modify the items saved from that moment. Why ? no fucking idea :(

[codecov-io]

Codecov Report

Merging #2229 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #2229   +/-   ##
=======================================
  Coverage   97.15%   97.15%           
=======================================
  Files          39       39           
  Lines        7899     7899           
  Branches     1369     1369           
=======================================
  Hits         7674     7674           
  Misses        100      100           
  Partials      125      125

---

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 d7658d4...cc3eb8d. Read the comment docs.

[asvetlov]

The function is used only once, please inline it -- the speed should be the same.

[pfreixes]

done

[asvetlov]

You've raised two separate questions:

1. Why raising non-frozen signals are allowed?

2. Does length calculation on .freeze() makes sense?

3. It's implementation detail. I just forgot to add check for .frozen in Signal.send().
   I see no harm in behavior changing, we could start from raising a warning.

4. Technically we could add cdef int _len attribute but I pretty sure PyList_Size() is pretty fast too. Extra field adds unnecessary complexity.

[pfreixes]

Yes, the cost of PyList_Size() vs doing explicitly with an attribute is negligible. And makes the code more complicated.

[asvetlov]

Thanks!

[lock]

This thread has been automatically locked since there has not been
any recent activity after it was closed. Please open a new issue for
related bugs.

If you feel like there's important points made in this discussion,
please include those exceprts into that new issue.