This library implements a small and efficient parser/generator for SDLang documents, providing a range based API. While the parser still uses the GC to allocate identifiers, strings etc., it uses a very efficient pool based allocation scheme that has very low computation and memory overhead.
The motivation for writing another SDLang implementation for D came from the high overhead that the original sdlang-d implementation has. Parsing a particular 200 MB file took well over 30 seconds and used up almost 10GB of memory if parsed into a DOM. The following changes to the parsing approach brought the parsing time down to around 2.5 seconds:
- Using a more efficient allocation scheme
- Using only "native" range implementations instead of the more comfortable fiber-based approach taken by sdlang-d
- Using
TaggedUnion(taggedalgebraic) instead ofVariant
Further substantial improvements at this point are more difficult and likely require the use of bit-level tricks and SIMD for speeding up the lexer, as well as exploiting the properties of pure array inputs.