Re:Show HN: FixBugs – Reproduce production bugs and verify f...
by kirtivr on Tuesday July 14, 2026 @01:07AM
(#48902521)
On how does the mocking work, that's a really interesting question.
We do a lot of AST parsing - for both code and build configuration languages. Even then, we still have to rely on the LLM to figure out a lot of the details.
Making this work reliably for non-frontier models and codebases that don't have existing test harnesses is where a lot of the design work goes in.
We do a lot of AST parsing - for both code and build configuration languages. Even then, we still have to rely on the LLM to figure out a lot of the details.
Making this work reliably for non-frontier models and codebases that don't have existing test harnesses is where a lot of the design work goes in.