parsanol 1.3.60 → 1.3.62

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7b0b6d85c512010d47ad5bb8237ac68ed2e9d16f1d4c7082bad5c77870ad1836
-  data.tar.gz: b97e208a29314adfe0d13344383518ea50114c140dc2208d74fd9b5bdd502677
+  metadata.gz: dc7df8984ab8f887379dc43d351e628656048cf4fd2a4bdd359b0b8b45341ce0
+  data.tar.gz: 38f9a43f7d4bf61621ef843d672f66883d5eee7abbe9af596b065499f89dcb12
 SHA512:
-  metadata.gz: f8d1294a9f3d674464b25087a4e7a0f89522be9e822b44a6307cbacd195f0db4cdf6151c2f1b02eb93d4b46acff1942fc4d5bf57efcbcc3b9c42f26a53177ce5
-  data.tar.gz: e6fd368e0aaae0519280932e4697d0e5f491975d25fbf2343a62f59822a55e076256ec7c6f9ce14c7017f61cde2a1daf45860c476108cee39baa5cab9beb9ba3
+  metadata.gz: 60c1f94221c0f97ebd61f27f6e409667630b0fd013aeabebab44cbb4918a8f22e951878ffce3f9d40b8a48d7a7bf4f4fb99c160efe11fdefe682beac63c48dca
+  data.tar.gz: b3fa201e5e2290698c3532eee6ba251a88547c4627348c791c08406b849ad651136ab65281ffa2f72b9629999b4af261679379f8f5e3d014ad82670ba0360349

data/Cargo.lock

@@ -307,9 +307,9 @@ checksum = "9f7c3e4beb33f85d45ae3e3a1792185706c8e16d043238c593331cc7cd313b50"
 
 [[package]]
 name = "parsanol"
-version = "0.10.1"
+version = "0.10.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3201f3b6340113ce57871a1405679ad7520a420cb8958777983371a2d4e6bf55"
+checksum = "dd43c104276def3f79a54ce6e8d26642ea330c13f97d66380bd94fc057f0a849"
 dependencies = [
  "ahash",
  "getrandom 0.3.4",
@@ -326,9 +326,9 @@ dependencies = [
 
 [[package]]
 name = "parsanol-derive"
-version = "0.10.1"
+version = "0.10.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "07bea1e5fb6a78314596116ac62e08ab5c73c201c62cd83715e75309c44e0977"
+checksum = "eca430ba21d7cc2d674080d40dd37a5567e1a0acdc37faef9a9b3738b8651dd6"
 dependencies = [
  "proc-macro2",
  "quote",

data/ext/parsanol_native/Cargo.toml

@@ -28,7 +28,7 @@ rb-sys = { version = "0.9.124", features = ["global-allocator"] }
 magnus = { version = "0.9" }
 
 # parsanol parser library (crates.io; 0.10 carries the class fast-path fix)
-parsanol = { version = "0.10", features = ["ruby"] }
+parsanol = { version = "0.10.2", features = ["ruby"] }
 
 # Logging
 log = "0.4"

data/lib/parsanol/native.rb

@@ -220,15 +220,21 @@ module Parsanol
         # to recover inputs from grammars native cannot express.
         parsanol_grammar =
           grammar.is_a?(Parsanol::Atoms::Base) || grammar.is_a?(Parsanol::Parser)
-        if parsanol_grammar && (m = error.message.match(NATIVE_POS_MARKER))
+        # The position capture is extracted as a plain String: a MatchData
+        # would pin its subject — the full multi-KB failure message — for
+        # as long as anything downstream retains it, and callers that
+        # rescue-and-collect (pubid's flavor dispatcher) retain millions
+        # across long parses (#123).
+        message = error.message
+        if parsanol_grammar && (pos_str = message[NATIVE_POS_MARKER, 1])
           source = Parsanol::Source.new(input)
           success, value = grammar.run_with_context(source, nil, true)
           return grammar.finalize_result(value) if success
 
-          pos = m[1].to_i
-          msg = error.message.sub(NATIVE_POS_MARKER, "")
+          pos = pos_str.to_i
+          msg = message.sub(NATIVE_POS_MARKER, "")
           cause = Parsanol::Cause.new(msg, source, pos)
-          raise Parsanol::ParseFailed.new(cause.to_s, cause)
+          raise Parsanol::ParseFailed.new(cause.to_s, cause), cause: nil
         end
 
         if parsanol_grammar
@@ -243,17 +249,20 @@ module Parsanol
           value.raise
         end
 
-        if (marker = error.message.match(NATIVE_POS_MARKER))
+        # Same capture-string extraction as above; `cause: nil` keeps the
+        # retained ParseFailed from pinning the original RuntimeError and
+        # its message through the exception-cause chain (#123).
+        if (pos_str = message[NATIVE_POS_MARKER, 1])
           source = Parsanol::Source.new(input)
           cause = Parsanol::Cause.new(
-            error.message.sub(NATIVE_POS_MARKER, ""), source, marker[1].to_i
+            message.sub(NATIVE_POS_MARKER, ""), source, pos_str.to_i
           )
-          raise Parsanol::ParseFailed.new(cause.to_s, cause)
+          raise Parsanol::ParseFailed.new(cause.to_s, cause), cause: nil
         end
 
         source = Parsanol::Source.new(input)
-        cause = Parsanol::Cause.new(error.message, source, source.bytepos)
-        raise Parsanol::ParseFailed.new(cause.to_s, cause)
+        cause = Parsanol::Cause.new(message, source, source.bytepos)
+        raise Parsanol::ParseFailed.new(cause.to_s, cause), cause: nil
       end
 
       # Pre-serialized JSON grammar path (library authors with cached JSON).

data/lib/parsanol/parg/artifact.rb

@@ -14,7 +14,7 @@ module Parsanol
       attr_reader :envelope, :path, :tables_dir
 
       def self.load(path, tables_dir: nil)
-        envelope = JSON.parse(File.read(path))
+        envelope = JSON.parse(PARG.read_utf8(path))
         new(envelope, path, tables_dir || File.dirname(path))
       end
 
@@ -169,7 +169,7 @@ module Parsanol
 
         file = declared
         path = File.join(@tables_dir.to_s, file)
-        raw = path.end_with?(".json") ? JSON.parse(File.read(path)) : YAML.safe_load_file(path, aliases: true)
+        raw = path.end_with?(".json") ? JSON.parse(PARG.read_utf8(path)) : YAML.safe_load_file(path, aliases: true)
         rows = case raw
                when Hash
                  raw.map { |key, value| { "name" => key.to_s }.merge(value.to_h.transform_keys(&:to_s)) }

data/lib/parsanol/parg/authoring.rb

@@ -25,7 +25,7 @@ module Parsanol
       end
 
       def read_file(file)
-        content = File.read(file).gsub(/^#[^\n]*\n/, "")
+        content = PARG.read_utf8(file).gsub(/^#[^\n]*\n/, "")
         header = content.match(/suite\s+\w+(?:\s+for\s+entry\s+(?<entry>\w+))?\s*\{/)
         unless header
           raise ParseError,

data/lib/parsanol/parg/cli.rb

@@ -36,7 +36,8 @@ module Parsanol
       end
 
       def initialize(argv)
-        @argv = argv.dup
+        # Under a non-UTF-8 locale ARGV arrives as BINARY; inputs are UTF-8.
+        @argv = argv.map { |arg| arg.dup.force_encoding(Encoding::UTF_8) }
         @tables_dir = nil
       end
 
@@ -82,7 +83,7 @@ module Parsanol
         source = @argv.shift
         out = flag_value("-o") || default_artifact_name(source)
         envelope = compile_source(source).envelope
-        File.write(out, "#{JSON.generate(envelope)}\n")
+        File.write(out, "#{JSON.generate(envelope)}\n", encoding: Encoding::UTF_8)
         puts "#{out}  #{envelope['checksum']}  warnings=#{envelope['lint']['order_warnings'].size}"
         0
       end
@@ -161,7 +162,7 @@ module Parsanol
         puts "parsanol parg repl — #{file} entry #{entry_name.inspect}; blank line quits"
         loop do
           print "> "
-          input = $stdin.gets
+          input = $stdin.gets&.force_encoding(Encoding::UTF_8)
           break if input.nil? || input.strip.empty?
 
           begin
@@ -209,7 +210,7 @@ module Parsanol
 
       def compile_source(file)
         tables = @tables_dir || default_tables_dir(file)
-        document = Parsanol::PARG::Parser.new(File.read(file)).parse
+        document = Parsanol::PARG::Parser.new(PARG.read_utf8(file)).parse
         Parsanol::PARG::Imports.merge!(document, [File.dirname(file)])
         Parsanol::PARG::Compiler.compile(document, tables_dir: tables)
       end

data/lib/parsanol/parg/compiler.rb

@@ -254,7 +254,7 @@ module Parsanol
         end
 
         raw = if path.end_with?(".json")
-                JSON.parse(File.read(path))
+                JSON.parse(PARG.read_utf8(path))
               else
                 YAML.safe_load_file(path, aliases: true)
               end

data/lib/parsanol/parg/imports.rb

@@ -19,7 +19,7 @@ module Parsanol
           raise CompileError, "import cycle: #{(merging + [name]).join(' -> ')}" if merging.include?(name)
 
           source = find_file(name, import_dirs)
-          used = Parser.new(File.read(source)).parse
+          used = Parser.new(PARG.read_utf8(source)).parse
           merge!(used, import_dirs, merging + [name])
           merge_used(document, used, "#{name}.")
         end

data/lib/parsanol/parg/selfhost.rb

@@ -32,7 +32,7 @@ module Parsanol
       def validate(source)
         raise ArtifactError, "parg artifact not found (set PG_ARTIFACT)" unless available?
 
-        envelope = JSON.parse(File.read(artifact_path))
+        envelope = JSON.parse(PARG.read_utf8(artifact_path))
         artifact = Artifact.new(envelope, nil, nil)
         artifact.parse("file", source)
       end

data/lib/parsanol/parg.rb

@@ -38,5 +38,12 @@ module Parsanol
     autoload :Imports, "parsanol/parg/imports"
     autoload :Suite, "parsanol/parg/authoring"
     autoload :Schema, "parsanol/parg/authoring"
+
+    # PARG sources and JSON artifacts are UTF-8 by definition: read them as
+    # UTF-8 whatever the locale's default external encoding, and skip a
+    # byte-order mark the way YAML.safe_load_file does (GH-121).
+    def self.read_utf8(path)
+      File.read(path, mode: "r:bom|utf-8")
+    end
   end
 end

data/lib/parsanol/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Parsanol
-  VERSION = "1.3.60"
+  VERSION = "1.3.62"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: parsanol
 version: !ruby/object:Gem::Version
-  version: 1.3.60
+  version: 1.3.62
 platform: ruby
 authors:
 - Ribose Inc.