Thoughts around using the regular expression parser independently. about duktape HOT 3 CLOSED

Originally I wrote the engine as a separate thing (constrained to the Ecmascript feature set of course), so it should be possible to extract the central parts with a reasonable effort:

• The regexp compilation primitive (duk_parse_disjunction() and parts of duk_parse_regexp())
• The regexp execution primitive (duk__match_regexp()) which attempts a match from a certain offset

These depend on Duktape for a few things: the buffer datatype is used to construct the bytecode, utf-8 helpers are used for codepoint conversions (the regexp bytecode used by Duktape is actually a series of codepoints). Also they call into Duktape error handling, etc.

Above that it gets quite Ecmascript/Duktape specific. To keep footprint minimal there is no abstract data structure for providing a match result (this is something you'd expect to have in a generic library); rather an Ecmascript result is constructed directly. These parts you'd need to reimplement, perhaps checking whether you want to replicate the detailed Ecmascript semantics or not (e.g. how global matching is implemented).

If you end up taking this approach, send me an e-mail and I'll be happy to help if I can.

from duktape.

Thanks for the info, @svaarala! I'll keep you updated.

from duktape.

Closing. Will follow up via e-mail. Looking at this stuff now. Thanks again!

from duktape.