Question 1

Is Re better than PCRE for OCaml projects?

Accepted Answer

Re offers predictable linear-time matching with DFA, avoiding PCRE's exponential backtracking, making it better for performance-sensitive apps. However, it lacks features like back-references, so for complex patterns, PCRE might still be necessary. The Re.Pcre module eases migration if you don't need those features.

Question 2

How to port existing PCRE code to Re in OCaml?

Accepted Answer

Use the Re.Pcre module, which provides a subset of the PCRE interface. Replace your PCRE function calls with Re.Pcre equivalents, and adjust for any missing features. The README notes this allows minimal code changes for easier porting.

Question 3

Does Re support look-ahead or look-behind assertions?

Accepted Answer

No, Re explicitly does not support look-ahead or look-behind assertions, as mentioned in the README under missing features. This limits its use for certain advanced regex tasks that require these capabilities.

Question 4

Why is Re slower when using bytecode compilation?

Accepted Answer

Re is written entirely in OCaml, so when compiled to bytecode, matching performance degrades significantly compared to native code. The README states this is a consequence of the pure OCaml implementation, affecting development in bytecode environments.

Question 5

Can Re handle very large regular expressions efficiently?

Accepted Answer

Yes, Re uses DFA-based matching that scales linearly with string length, making it efficient for large patterns. However, for some combinations, DFA building can be slow, though the README notes this is rare in practice based on timing results.

Question 6

What are the main performance trade-offs with Re?

Accepted Answer

Re trades faster compilation time for predictable linear matching speed. While pattern compilation is slower than backtracking libraries, once the DFA is built, matching is extremely fast, especially in native code. The README provides examples where Re outperforms PCRE for repeated matches.

Question 7

Is Re suitable for building file synchronization tools?

Accepted Answer

Yes, Re is well-suited for tools like Unison that require fast, repeated pattern matching against filenames, as shown in the README's timing results where it outperformed PCRE for ignore rules. Its DFA-based matching prevents performance bottlenecks in such applications.

Re

What is Re?

Overview

Use Cases

Best For

Found a gem we're missing?

Not Ideal For

Pros & Cons

Pros

Cons

Frequently Asked Questions