rewriting: dead code removal in GreedyRewritePatternApplier

[jumerckx]

No description provided.

[jumerckx]

Many filecheck tests assume dead code is not removed.
What are thoughts on this pr?
Should all those tests be updated by adding a user of the dead results?

[superlopuh]

Yeah I think so

[superlopuh]

I wonder if agents these days can do this easily

[superlopuh]

Ah no sorry, I think remove the use of the dead code, or add "test.op" user dependening on the test

[jumerckx]

Marking this as draft until riscv fixes have been merged.

[superlopuh]

One of the riscv things was much easier to fix by avoiding the use of the greedy rewriter, I'll take a look at the rest of the backend things over the next week.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 86.15%. Comparing base (54c86a4) to head (20a0524).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5485      +/-   ##
==========================================
- Coverage   86.16%   86.15%   -0.02%     
==========================================
  Files         395      395              
  Lines       56464    56473       +9     
  Branches     6504     6506       +2     
==========================================
+ Hits        48655    48657       +2     
- Misses       6276     6281       +5     
- Partials     1533     1535       +2     

☔ 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.

[superlopuh]

this seems like an actual optimisation

[superlopuh]

@math-fehr, are you happy to merge this with the temporary flag to disable dce in the greedy rewriter? I think it would be good to deprecate that at some point but for now seems like a good way to move forward.

[math-fehr]

Sorry for missing this, this is good for me!