externalize CharsetVerifier, add looksLikeText to OpenPgpMetadata object

1 files changed, 18 insertions, 76 deletions

diff --git a/OpenKeychain/src/main/java/org/sufficientlysecure/keychain/operations/InputDataOperation.java b/OpenKeychain/src/main/java/org/sufficientlysecure/keychain/operations/InputDataOperation.java
index 80f9d6368..ff9377581 100644
--- a/OpenKeychain/src/main/java/org/sufficientlysecure/keychain/operations/InputDataOperation.java
+++ b/OpenKeychain/src/main/java/org/sufficientlysecure/keychain/operations/InputDataOperation.java
@@ -23,13 +23,6 @@ import java.io.FileNotFoundException;
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.OutputStream;
-import java.nio.ByteBuffer;
-import java.nio.CharBuffer;
-import java.nio.charset.Charset;
-import java.nio.charset.CharsetDecoder;
-import java.nio.charset.CoderResult;
-import java.nio.charset.CodingErrorAction;
-import java.nio.charset.UnsupportedCharsetException;
 import java.util.ArrayList;
 
 import android.content.ClipDescription;
@@ -75,14 +68,9 @@ import org.sufficientlysecure.keychain.service.input.CryptoInputParcel;
 public class InputDataOperation extends BaseOperation<InputDataParcel> {
 
     private final byte[] buf = new byte[256];
-    private final ByteBuffer bufWrap;
-    private final CharBuffer dummyOutput;
 
     public InputDataOperation(Context context, ProviderHelper providerHelper, Progressable progressable) {
         super(context, providerHelper, progressable);
-
-        bufWrap = ByteBuffer.wrap(buf);
-        dummyOutput = CharBuffer.allocate(256);
     }
 
     Uri mSignedDataUri;
@@ -338,83 +326,37 @@ public class InputDataOperation extends BaseOperation<InputDataParcel> {
                     throw new IOException("Error getting file for writing!");
                 }
 
-                boolean isPossibleTextMimeType = ClipDescription.compareMimeTypes(mimeType, "application/octet-stream")
-                        || ClipDescription.compareMimeTypes(mimeType, "application/x-download")
-                        || ClipDescription.compareMimeTypes(mimeType, "text/*");
-
                 // If this data looks like text, we pipe the incoming data into a charset
                 // decoder, to see if the data is legal for the assumed charset.
-                String charset;
-                boolean charsetIsFaulty;
-                boolean charsetIsGuessed;
-                CharsetDecoder charsetDecoder = null;
-                if (isPossibleTextMimeType) {
-                    charset = bd.getCharset();
-                    // the charset defaults to us-ascii, but we want to default to utf-8
-                    if (charset == null || "us-ascii".equals(charset)) {
-                        charset = "utf-8";
-                        charsetIsGuessed = true;
-                    } else {
-                        charsetIsGuessed = false;
-                    }
-
-                    try {
-                        charsetDecoder = Charset.forName(charset).newDecoder();
-                        charsetDecoder.onMalformedInput(CodingErrorAction.REPORT);
-                        charsetDecoder.onUnmappableCharacter(CodingErrorAction.REPORT);
-                        charsetDecoder.reset();
-                        charsetIsFaulty = false;
-                    } catch (UnsupportedCharsetException e) {
-                        charsetIsFaulty = true;
-                    }
-                } else {
-                    charsetIsFaulty = true;
-                    charsetIsGuessed = false;
-                    charset = null;
-                }
+                String charset = bd.getCharset();
+                CharsetVerifier charsetVerifier = new CharsetVerifier(buf, mimeType, charset);
 
                 int totalLength = 0;
                 do {
                     totalLength += len;
                     out.write(buf, 0, len);
-
-                    if (isPossibleTextMimeType && !charsetIsFaulty) {
-                        bufWrap.rewind();
-                        bufWrap.limit(len);
-                        dummyOutput.rewind();
-                        CoderResult result = charsetDecoder.decode(bufWrap, dummyOutput, false);
-                        if (result.isError()) {
-                            charsetIsFaulty = true;
-                        }
-                    }
+                    charsetVerifier.write(0, len);
                 } while ((len = is.read(buf)) > 0);
 
-                if (!charsetIsFaulty) {
-                    bufWrap.rewind();
-                    bufWrap.limit(0);
-                    dummyOutput.rewind();
-                    CoderResult result = charsetDecoder.decode(bufWrap, dummyOutput, true);
-                    if (result.isError()) {
-                        charsetIsFaulty = true;
-                    }
-                }
+                log.add(LogType.MSG_DATA_MIME_LENGTH, 3, Long.toString(totalLength));
 
-                if (isPossibleTextMimeType) {
-                    if (charsetIsFaulty && charsetIsGuessed) {
-                        log.add(LogType.MSG_DATA_MIME_CHARSET_UNKNOWN, 3, charset);
-                        charset = null;
-                    } else if (charsetIsFaulty) {
-                        log.add(LogType.MSG_DATA_MIME_CHARSET_FAULTY, 3, charset);
-                    } else if (charsetIsGuessed) {
-                        log.add(LogType.MSG_DATA_MIME_CHARSET_GUESS, 3, charset);
+                OpenPgpMetadata metadata;
+                if (charsetVerifier.isDefinitelyBinary()) {
+                    metadata = new OpenPgpMetadata(mFilename, mimeType, 0L, totalLength);
+                } else {
+                    if (charsetVerifier.isCharsetFaulty() && charsetVerifier.isCharsetGuessed()) {
+                        log.add(LogType.MSG_DATA_MIME_CHARSET_UNKNOWN, 3, charsetVerifier.getMaybeFaultyCharset());
+                    } else if (charsetVerifier.isCharsetFaulty()) {
+                        log.add(LogType.MSG_DATA_MIME_CHARSET_FAULTY, 3, charsetVerifier.getCharset());
+                    } else if (charsetVerifier.isCharsetGuessed()) {
+                        log.add(LogType.MSG_DATA_MIME_CHARSET_GUESS, 3, charsetVerifier.getCharset());
                     } else {
-                        log.add(LogType.MSG_DATA_MIME_CHARSET, 3, charset);
+                        log.add(LogType.MSG_DATA_MIME_CHARSET, 3, charsetVerifier.getCharset());
                     }
-                }
-
-                log.add(LogType.MSG_DATA_MIME_LENGTH, 3, Long.toString(totalLength));
 
-                OpenPgpMetadata metadata = new OpenPgpMetadata(mFilename, mimeType, 0L, totalLength, charset);
+                    metadata = new OpenPgpMetadata(mFilename, mimeType, 0L, totalLength,
+                            charsetVerifier.getCharset(), charsetVerifier.isProbablyText());
+                }
 
                 out.close();
                 outputUris.add(uri);