List view
RegExp in Nova currently uses the excellent [regex](https://github.com/rust-lang/regex/) crate; it is fast, it is efficient, it can match over arbitrary byte sequences including WTF-8 strings, and it has guaranteed linear execution time. That is **excellent**! But... it does not match the ECMAScript specification, it cannot match using WTF-8 expressions, it does not support lookaheads or lookbehinds (and maybe never will despite these now being possible in linear time), and it most definitely will never support backreferences (as these cannot be supported in linear time, at least in the general case). So... we probably need to take some other engine and fork it, or create our own. Options are: * regex: fork, support WTF-8 over UTF-8, change engine to work on ECMAScript specification, take up lookahead and lookbehind code, leave backreferences on the table for now and/or possible implement them as the only non-linear matching mode. This'll be hard in many ways. * regress: fork, support WTF-8 over UTF-16. This'll be much easier but gives us a generally non-linear engine AFAIU. * build own: worst of all worlds.
No due date•0/2 issues closedProcess bug reports from beta release, as well as the results of the security audit.
Overdue by 4 month(s)•Due by October 3, 2025•1/3 issues closedA JavaScript engine that cannot stop JavaScript execution to perform garbage collection is mostly a toy. Nova's GC is currently only interleaved behind a flag and this needs to change.
Overdue by 4 month(s)•Due by October 3, 2025•5/6 issues closedFully implement the BuiltinFunction stubs
No due date•31/37 issues closed