Simplify now that RubySource always has a TruffleString

Author: eregon

src/main/java/org/truffleruby/language/TruffleBootNodes.java

@@ -132,7 +132,7 @@ int main(int argc, long argv, String kind, String toExecute) {
                 if (getContext().getOptions().SYNTAX_CHECK) {
                     checkSyntax.call(coreLibrary().truffleBootModule, "check_syntax", source);
                 } else {
-                    var tstringWithEncoding = source.hasTruffleString() ? source.getTStringWithEncoding() : null;
+                    var tstringWithEncoding = source.getTStringWithEncoding();
                     var sourceTStringPair = Pair.create(source.getSource(), tstringWithEncoding);
                     final RootCallTarget callTarget = getContext()
                             .getCodeLoader()

src/main/java/org/truffleruby/parser/RubySource.java

@@ -35,6 +35,7 @@ public final class RubySource {
     private final String sourcePath;
     private final TruffleString code;
     private byte[] bytes;
+    // FIXME: not always the source encoding, e.g. when loading from a file or eval, need to check the magic encoding comment
     private final RubyEncoding encoding;
     private final boolean isEval;
     private final int lineOffset;
@@ -81,22 +82,15 @@ public String getSourcePath() {
         return sourcePath;
     }
 
-    public boolean hasTruffleString() {
-        return code != null;
-    }
-
     public TruffleString getTruffleString() {
-        assert hasTruffleString();
         return code;
     }
 
     public TStringWithEncoding getTStringWithEncoding() {
-        assert hasTruffleString();
         return new TStringWithEncoding(code, encoding);
     }
 
     public byte[] getBytes() {
-        assert hasTruffleString();
         if (bytes != null) {
             return bytes;
         } else {
@@ -105,7 +99,6 @@ public byte[] getBytes() {
     }
 
     public RubyEncoding getEncoding() {
-        assert hasTruffleString();
         return encoding;
     }
 

src/main/java/org/truffleruby/parser/lexer/LexerSource.java

@@ -52,7 +52,6 @@ public final class LexerSource {
     public ParserRopeOperations parserRopeOperations;
     private final Source source;
     private final String sourcePath;
-    private final boolean fromTruffleString;
 
     private TruffleString sourceBytes;
     private final int sourceByteLength;
@@ -64,18 +63,8 @@ public LexerSource(RubySource rubySource) {
         this.source = rubySource.getSource();
         this.sourcePath = rubySource.getSourcePath();
 
-        fromTruffleString = rubySource.hasTruffleString();
-
-        final RubyEncoding rubyEncoding;
-        if (fromTruffleString) {
-            rubyEncoding = rubySource.getEncoding();
-            this.sourceBytes = rubySource.getTruffleString();
-        } else {
-            rubyEncoding = Encodings.UTF_8;
-            // TODO CS 5-Sep-17 can we get the bytes directly rather than using getCharacters -> toString -> getBytes?
-            var sourceString = source.getCharacters().toString();
-            this.sourceBytes = TStringUtils.utf8TString(sourceString);
-        }
+        final RubyEncoding rubyEncoding = rubySource.getEncoding();
+        this.sourceBytes = rubySource.getTruffleString();
         this.sourceByteLength = sourceBytes.byteLength(rubyEncoding.tencoding);
         this.encoding = rubyEncoding;
         parserRopeOperations = new ParserRopeOperations(this.encoding);
@@ -138,10 +127,6 @@ private int nextNewLine() {
         }
     }
 
-    public boolean isFromTruffleString() {
-        return fromTruffleString;
-    }
-
     public int getLineOffset() {
         return lineOffset;
     }

src/main/java/org/truffleruby/parser/lexer/RubyLexer.java

@@ -460,27 +460,6 @@ protected void setEncoding(TruffleString name) {
             throw argumentError(context, nameString + " is not ASCII compatible");
         }
 
-        if (!src.isFromTruffleString() && !isUTF8Subset(newEncoding)) {
-            /* The source we are lexing came in via a String (or Reader, or File) from the Polyglot API, so we only have
-             * the String - we don't have any access to the original bytes, so we cannot re-interpret them in another
-             * encoding without risking errors. */
-
-            final String description;
-
-            if (src.getSourcePath().equals("-e")) {
-                description = "program from an -e argument";
-            } else {
-                description = "Polyglot API Source";
-            }
-
-            throw argumentError(
-                    context,
-                    String.format(
-                            "%s cannot be used as an encoding for a %s as it is not UTF-8 or a subset of UTF-8",
-                            nameString,
-                            description));
-        }
-
         setEncoding(newEncoding);
     }