syntaks
syntaks
Version 0.3.2
A parser combinator framework.
Usage
require "syntaks"
Example
class ListParser < Syntaks::Parser
rule(:root, Array(String), _os >> "(" >> _os & item_list & ")") { |v| v[3] }
rule(:item_list, Array(String), item >> ~item_list_tail) do |v|
if tail = v[1]
[v[0]] + tail
else
[v[0]]
end
end
rule(:item_list_tail, Array(String), {"," & _os & item >> _os}) do |list|
list.map(&.[2])
end
rule(:item, String, id >> _os) do |t|
t[0]
end
rule(:id, String, /[_a-z][_a-z0-9]*/, &.content)
ignored(:_os, /\s*/)
end
res = ListParser.new.call("(v1,v2)").as(Success)
assert res.value == ["v1", "v2"]
Usage
Rules are specified in a format similar to EBNF:
# literals/tokens
ignored(:space, /\s+/) # generates nil as AST-node
rule(:int, String, /[1-9][0-9]*/) { |m| m.content } # generates String as AST-node
# sequence
rule(:root, String, "var " >> id) { |m| m[1] } # use the second match (id) as AST-node
# sequence with backtracking disabled
rule(:root, String, "var " & id) { |m| m[1] }
# alternatives
rule(:root, String, a | b)
# repetition
rule(:root, Array(Int32), {"," >> int}) { |m| m.map(&.[1]) }
When backtracking is disabled an parsing fails, a syntax error is produced:
class ErrorParser < Syntaks::Parser
rule(:root, String, "var" >> space & /[a-z]+/) { |m| m[2].content }
ignored(:space, /\s+/)
end
res = ErrorParser.new.call("var 1337").as(Error)
assert res.message == <<-MSG
Syntax error at 1:9:
1>var 1337
----------^
Rule: "var" >> space & /[a-z]+/
MSG
A parse log can be printed to inspect the parsing process:
source = <<-SRC
def main
if true
x = 1
⬤
end
end
SRC
ctx = LoggingContext.new(ParseLog.new(Source.new(source)))
res = ExampleParsers::MethodParser.new.call(source, ctx).as(Error)
log = ctx.parse_log.to_s
# The syntax error entry marked with ⚡ should be on the line that contains ⬤
assert(/⚡[^\n]+(29-33)[^\n]+⬤/ === log)
TODO
- simple token macro
- rules that do not generate ast nodes
- error-recovery productions / synchronization tokens
- optional whitespace skippers
- Simple lexer?
- Packrat parsing / memoization
- remember longest parsed prefix when parsing fails
DONE
- Use ameba
- Generated
README.md
with examples - LoggingContext with a ParseLog
- ProfilingContext
- EBNF DSL
- Selectively disable backtracking
- Detailed syntax errors: display production that failed and the location where exactly it failed
Contributing
- Fork it ( https://github.com/ragmaanir/syntaks/fork )
- Create your feature branch (git checkout -b my-new-feature)
- Commit your changes (git commit -am 'Add some feature')
- Push to the branch (git push origin my-new-feature)
- Create a new Pull Request
Contributors
- Ragmaanir - creator, maintainer
Repository
syntaks
Owner
Statistic
- 0
- 0
- 0
- 0
- 2
- over 1 year ago
- September 15, 2017
License
MIT License
Links
Synced at
Thu, 07 Nov 2024 19:25:53 GMT
Languages