Oyl

Build and Run

Build with a C11 compiler (GCC/Clang tested on Linux and macOS).

make              # build liboyl.a
make test-all     # run all tests
make test-suite   # run YAML Test Suite (needs submodules)
make bench        # scanner benchmark on your machine

Core Capabilities

Architecture

Oyl keeps scanner and parser responsibilities clean: raw tokens first, structure and properties in parser, then emit text.

Quick API Snapshot

const oyl_event *evt;
while (oyl_parse_next(parser, &evt) == OYL_OK) {
  if (evt->type == OYL_EVT_STREAM_END) break;
  /* consume event */
}

if (oyl_parse_next(parser, &evt) != OYL_OK) {
  const char *msg = oyl_parser_error(parser);
  oyl_mark mark = oyl_parser_error_mark(parser);
  printf("%zu:%zu: %s\n", mark.line, mark.col, msg);
}

Benchmarks

Parse throughput in MB/s on one core, higher is better. Every parser reads the same bytes; each panel has its own scale.

All results were measured here with make bench-compare: Intel Core Ultra 7 155H, one core, GCC 16.2 -O2, median of 11 runs in each of 3 passes, October 2026. The real-file and long-text inputs are rapidyaml’s own benchmark files. All four are measured as event parsers: Oyl, libyaml 0.2.5, libfyaml 0.9.6, and rapidyaml 0.16 through its event parser with buffers reused between parses, its fastest mode, rather than its usual API, which builds a tree. Run it on your own machine, or see the README for details.