Fix async generator issue and prefer opcode comparison to name comparison everywhere

[MatthieuDartiailh]

Closes #178

[codecov-commenter]

Codecov Report

❌ Patch coverage is 96.96970% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 95.16%. Comparing base (5229b3e) to head (2d9a0fd).

Files with missing lines	Patch %	Lines
src/bytecode/concrete.py	75.00%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #179   +/-   ##
=======================================
  Coverage   95.16%   95.16%           
=======================================
  Files           7        7           
  Lines        2025     2028    +3     
  Branches      452      449    -3     
=======================================
+ Hits         1927     1930    +3     
  Misses         54       54           
  Partials       44       44           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[MatthieuDartiailh]

Running the benchmark found in #172 does not show a significant improvement (on one run it looks faster on average but the variance is large).

[P403n1x87]

@MatthieuDartiailh thanks for putting this together and apologies for the late reply but I've not had much time to look into this. It sounds like there are no measurable performance improvements from this, but I think there still is the advantage of cleaner coding anyway. May I suggest splitting this PR into 3 separate ones to address the 3 individual goals and keep the git history cleaner?

[MatthieuDartiailh]

Sure I can split the work between dropping 3.9 and 3.10 (and subsequent simplifications) and the rest. But what is the third part you would have seen ? @P403n1x87

[P403n1x87]

what is the third part you would have seen ?

Based on the PR title I thought this PR had 3 parts:

• Drop 3.9 and 3.10
• simplify code base
• fix async gen issue

[MatthieuDartiailh]

Yes but dumping 3.9 and 3.10 without taking advantage of that to remove the related version checks feels like unfinished work to me. I will split in 2.

[amazingandyyy]

when will this be in the next release?

[amazingandyyy]

we are upgrading ddtrace, and it depends on this fix to support python 3.14

[MatthieuDartiailh]

Thanks for the ping. I do not have a date yet but I will try to come back to you ASAP

[P403n1x87]

we are upgrading ddtrace, and it depends on this fix to support python 3.14

@amazingandyyy have you opened an issue on dd-trace-py too by any chance? If not, I'd be curious to see what problem this is causing and if we can work around it there in the meantime 🙏

[MatthieuDartiailh]

0.18.0 is released