[GR-19220] Kernel#sprintf: don't unconditionally grant encoding validity (#1852).

PullRequest: truffleruby/1179

Author: aardvark179

CHANGELOG.md

@@ -36,6 +36,7 @@ Bug fixes:
 * Do not leak TruffleRuby specific method Array#swap (#1816)
 * Fixed `#inspect` on broken UTF-8 sequences (#1842, @chrisseaton).
 * `Truffle::Interop.keys` should report methods of String and Symbol (#1817)
+* `Kernel#sprintf` encoding validity has been fixed (#1852, @XrXr).
 
 Compatibility:
 

spec/ruby/core/kernel/shared/sprintf.rb

@@ -865,6 +865,20 @@ def obj.to_str; end
     end
   end
 
+  describe "encoding" do
+    it "can produce a string with valid encoding" do
+      string = format("good day %{valid}", valid: "e")
+      string.encoding.should == Encoding::UTF_8
+      string.valid_encoding?.should be_true
+    end
+
+    it "can produce a string with invalid encoding" do
+      string = format("good day %{invalid}", invalid: "\x80")
+      string.encoding.should == Encoding::UTF_8
+      string.valid_encoding?.should be_false
+    end
+  end
+
   describe "faulty key" do
     before :all do
       @base_method = @method

src/main/java/org/truffleruby/core/format/FormatNode.java

@@ -12,11 +12,9 @@
 import java.nio.ByteBuffer;
 import java.util.Arrays;
 
-import org.jcodings.specific.USASCIIEncoding;
 import org.truffleruby.core.array.ArrayUtils;
 import org.truffleruby.core.format.exceptions.TooFewArgumentsException;
 import org.truffleruby.core.rope.CodeRange;
-import org.truffleruby.core.rope.RopeOperations;
 import org.truffleruby.language.RubyBaseNode;
 
 import com.oracle.truffle.api.CompilerDirectives;
@@ -162,7 +160,6 @@ protected void writeByte(VirtualFrame frame, byte value) {
         final int outputPosition = getOutputPosition(frame);
         output[outputPosition] = value;
         setOutputPosition(frame, outputPosition + 1);
-        setStringCodeRange(frame, value >= 0 ? CodeRange.CR_7BIT : CodeRange.CR_VALID);
         increaseStringLength(frame, 1);
     }
 
@@ -175,9 +172,6 @@ protected void writeBytes(VirtualFrame frame, byte[] values, int valuesLength) {
         final int outputPosition = getOutputPosition(frame);
         System.arraycopy(values, 0, output, outputPosition, valuesLength);
         setOutputPosition(frame, outputPosition + valuesLength);
-        setStringCodeRange(
-                frame,
-                RopeOperations.isAsciiOnly(values, USASCIIEncoding.INSTANCE) ? CodeRange.CR_7BIT : CodeRange.CR_VALID);
         increaseStringLength(frame, valuesLength);
     }
 

src/main/java/org/truffleruby/core/format/FormatRootNode.java

@@ -52,7 +52,7 @@ public Object execute(VirtualFrame frame) {
         frame.setObject(FormatFrameDescriptor.OUTPUT_SLOT, new byte[expectedLength]);
         frame.setInt(FormatFrameDescriptor.OUTPUT_POSITION_SLOT, 0);
         frame.setInt(FormatFrameDescriptor.STRING_LENGTH_SLOT, 0);
-        frame.setInt(FormatFrameDescriptor.STRING_CODE_RANGE_SLOT, CodeRange.CR_7BIT.toInt());
+        frame.setInt(FormatFrameDescriptor.STRING_CODE_RANGE_SLOT, CodeRange.CR_UNKNOWN.toInt());
         frame.setBoolean(FormatFrameDescriptor.TAINT_SLOT, false);
         frame.setObject(FormatFrameDescriptor.ASSOCIATED_SLOT, null);