Skip to content

Plan

The milestones: M0 and M1 are done, M2 (evaluator) has slices 1 to 3 done (v0.5.0 is current), M3 to M8 are planned.

Status: in development. M0 and M1 are done, and slices 1 to 3 of M2 are done, released as v0.5.0 of go-embedded-r/r (slice 2 was v0.3.0; v0.4.0 makes the Eval limits bounded by default, see Eval API; the parser was v0.1.0, deparse v0.2.0). The plan is bib/r.md of go-scikit/roadmap. Every stage is checked against the pinned GNU R 4.6.2, never against expectations written by hand.

MilestoneContentsState
M0 oracle harnessinternal/oracle, cmd/roracle, rds (serialization reader and writer), golden filesdone
M1 parsersyntax: lexer, parser, AST with source map, tokens, R’s parse errors, deparse, TopLevelNamesdone
M2 evaluatoreval: values, environments, promises, closures, control flow, replacement functions, subsetting, attributes, S3 dispatch, conditions, printing, the Eval API with limits; 102 of the 210 primitives and 94 base closures, environments as values, the apply family, character functions and regular expressions (an approximation, not TRE)slices 1 to 3 done (v0.5.0); see Evaluator, Regular expressions and Eval API
M3 to M8base, stats, regex, graphics, notebook bridge, WebAssembly playground, CRAN portsplanned

Nothing in the M3 to M8 row exists yet. The README names the subjects and does not assign them to milestone numbers one by one, so neither does this page.

Known gaps of M1

  • the effect of #line directives on the line numbers of later errors;
  • parse(file=) encoding options;
  • an invalid UTF-8 input is parsed as R’s lexer does, and for such text R cannot produce parse data, so its tokens are not compared;
  • numeric constants that deparse differently on some platforms: see Deparse.

See The parser for the parity measurements.

Created on by tannevaled

Modified on by tannevaled

The parser