[tannevaled]

Bumps github.com/go-ruby-parser/parser to v0.2.0 (go-ruby-parser/parser#28, issue #27).

The point is not the fourteen constructs themselves. Thirteen language/*_spec.rb files of the pinned corpus could not be parsed at all, so they contributed zero to both columns of the conformance measurement — they were not failing, they were invisible.

Measurement

language/ swept with the pre-bump and post-bump binaries, same corpus, same timeout:

	passing	fail+error
parser v0.1.9	1443	283
parser v0.2.0	1776	436

+333 passing. fail+error rises at the same time because the files that now parse bring their own failures with them, having previously contributed neither. That is the same arithmetic variables_spec produced when it was unblocked: it had 113 passing examples nobody had ever counted.

Per file, from no result at all to a result: pattern_matching 95 · predefined 68 · class 32 · hash 29 · string 23 · alias 22 · metaclass 18 · assignments 18 · undef 6.

Two of the fourteen were silent mis-parses

Worse than a refusal, because nothing diagnoses them:

o.s (:a){ 1 }          # the brace block bound to :a, not to the call
case [0,1,2,3]
in [0, 1, ]            # the trailing comma was dropped, so a partial pattern
  :partial             # became an exact one and this fell through to else
else
  :exact
end

Both verified against ruby 4.0.5 through the rbgo binary built on this bump.

Not included

Four files parse now but stop in rbgo's own compiler, and are consumer work tracked separately from this bump: for_spec (the new additive For.Target field — For.Vars is empty exactly when it is set), block_spec (block-local slots, now carried on Block.Locals), defined_spec (defined?(a += 1) declares a local no slot is allocated for — reproduces on the old parser too), variables_spec (cannot compile *ast.SplatArg).

Gates

CGO_ENABLED=0 go build ./..., full CGO_ENABLED=0 GOWORK=off go test ./... green (exit 0, no pinned test needed updating), gofmt clean, wasip1 builds.

🤖 Generated with Claude Code