Refactor messy functions — AI Prompts for C++

The four heights

The same task, four distances: today's deadline, the next reviewer, the stuck moment, the pattern.

Execute — do the immediate task

Refactor the input-event handler that accumulated 1,200 lines in player_input.cpp. Extract clear…

Refactor the input-event handler that accumulated 1,200 lines in player_input.cpp. Extract clear subfunctions, keep behavior identical, and send the cleaned file to Noor in QA and to me for integration testing. Do the refactor in small commits and verify unit tests pass; aim to finish by Wednesday.

Improve — make it easier to accept

Before I hand the refactor to Noor in QA, make it easy to approve — preserve the public function…

Before I hand the refactor to Noor in QA, make it easy to approve — preserve the public function signatures, add a brief header that explains responsibilities of each new helper, surface the single behavioral change (if any) up front, and mark any edge cases that would make a reviewer hesitate to approve.

Decide — diagnose the stuck moment

I opened player_input.cpp and found 1,200 lines of intertwined logic handling game, UI, and debug…

The input handler has become a fragile 1,200-line monster

I opened player_input.cpp and found 1,200 lines of intertwined logic handling game, UI, and debug input. I’m worried that extracting helpers will change timing and break exotic key combos players use. I don’t know which pieces are safe to pull out first. What should I extract first to lower risk and how do I validate I didn’t alter behavior?

Become — change the pattern

We keep ending up with huge, fragile functions after every feature sprint because developers patch…

Large functions keep returning after quick fixes

We keep ending up with huge, fragile functions after every feature sprint because developers patch behavior quickly for deadlines. I spend time repeatedly refactoring the same file. Which habit should I change across the team to keep code decomposed, and what simple rules for PRs and code reviews will stop monolithic growth?

Next to this one

Other programming language work people do in C++.

Parse input files Profile CPU hotspots Profile GPU or SIMD code Profile memory usage Read and write binary files Resolve ABI compatibility issues Resolve linker errors Run unit tests locally Serialize and deserialize data Set up build system

Every task here came from the work, not from a feature list — which is why the prompts name what you want done and never the button that does it. The tool changes; the work does not.

The rest of the map

Same library, five ways in.

Work Atlasthe whole map, interactive Careers1,949 jobs and their real tasks Soft skillsthe human part of the work Activities500 things people do Prompt Librarybrowse every prompt