Attachment 45701 Details for Bug 149028 – Improved fix (with same regression test)

[patch] Improved fix (with same regression test)

149028.txt (text/plain), 7.79 KB, created by Jerome Lanneluc

on 2006-07-04 09:09:25 EDT

(hide)

Description:

Filename:

MIME Type:

Creator: Jerome Lanneluc

Created: 2006-07-04 09:09:25 EDT

Size: 7.79 KB

patch

obsolete

>### Eclipse Workspace Patch 1.0
>#P org.eclipse.jdt.core
>Index: compiler/org/eclipse/jdt/internal/compiler/util/Util.java
>===================================================================
>RCS file: /cvsroot/eclipse/org.eclipse.jdt.core/compiler/org/eclipse/jdt/internal/compiler/util/Util.java,v
>retrieving revision 1.55
>diff -u -r1.55 Util.java
>--- compiler/org/eclipse/jdt/internal/compiler/util/Util.java	29 Mar 2006 02:50:01 -0000	1.55
>+++ compiler/org/eclipse/jdt/internal/compiler/util/Util.java	4 Jul 2006 12:23:43 -0000
>@@ -155,33 +155,10 @@
> 
> 		return contents;
> 	}
>-	/*
>-	 * NIO support to get input stream as char array.
>-	 * Not used as with JDK 1.4.2 this support is slower than standard IO one...
>-	 * Keep it as comment for future in case of next JDK versions improve performance
>-	 * in this area...
>-	public static char[] getInputStreamAsCharArray(FileInputStream stream, int length, String encoding)
>-		throws IOException {
>-		
>-		FileChannel channel = stream.getChannel();
>-		int size = (int)channel.size();
>-		if (length >= 0 && length < size) size = length;
>-		Charset charset = encoding==null?systemCharset:Charset.forName(encoding);
>-		if (charset != null) {
>-			MappedByteBuffer bbuffer = channel.map(FileChannel.MapMode.READ_ONLY, 0, size);
>-		    CharsetDecoder decoder = charset.newDecoder();
>-		    CharBuffer buffer = decoder.decode(bbuffer);
>-		    char[] contents = new char[buffer.limit()];
>-		    buffer.get(contents);
>-		    return contents;
>-		}
>-		throw new UnsupportedCharsetException(SYSTEM_FILE_ENCODING);
>-	}
>-	*/
>+	
> 	/**
> 	 * Returns the given input stream's contents as a character array.
>-	 * If a length is specified (ie. if length != -1), only length chars
>-	 * are returned. Otherwise all chars in the stream are returned.
>+	 * If a length is specified (ie. if length != -1), this represents the number of bytes in the stream.
> 	 * Note this doesn't close the stream.
> 	 * @throws IOException if a problem occured reading the stream.
> 	 */
>@@ -197,73 +174,51 @@
> 			reader =  new InputStreamReader(stream);
> 		}
> 		char[] contents;
>+		int totalRead = 0;
> 		if (length == -1) {
> 			contents = CharOperation.NO_CHAR;
>-			int contentsLength = 0;
>-			int amountRead = -1;
>-			do {
>-				int amountRequested = Math.max(stream.available(), DEFAULT_READING_SIZE);  // read at least 8K
>+		} else {
>+			// length is a good guess when the encoding produces less or the same amount of characters than the file length
>+			contents = new char[length]; // best guess
>+		}
> 
>+		while (true) {
>+			int amountRequested;
>+			if (totalRead < length) {
>+				// until known length is met, reuse same array sized eagerly
>+				amountRequested = length - totalRead;
>+			} else {
>+				// reading beyond known length
>+				int current = reader.read(); 
>+				if (current < 0) break;
>+				
>+				amountRequested = Math.max(stream.available(), DEFAULT_READING_SIZE);  // read at least 8K
>+				
> 				// resize contents if needed
>-				if (contentsLength + amountRequested > contents.length) {
>-					System.arraycopy(
>-						contents,
>-						0,
>-						contents = new char[contentsLength + amountRequested],
>-						0,
>-						contentsLength);
>-				}
>-
>-				// read as many chars as possible
>-				amountRead = reader.read(contents, contentsLength, amountRequested);
>-
>-				if (amountRead > 0) {
>-					// remember length of contents
>-					contentsLength += amountRead;
>-				}
>-			} while (amountRead != -1);
>-
>-			// Do not keep first character for UTF-8 BOM encoding
>-			int start = 0;
>-			if (contentsLength > 0 && UTF_8.equals(encoding)) {
>-				if (contents[0] == 0xFEFF) { // if BOM char then skip
>-					contentsLength--;
>-					start = 1;
>-				}
>-			}
>-			// resize contents if necessary
>-			if (contentsLength < contents.length) {
>-				System.arraycopy(
>-					contents,
>-					start,
>-					contents = new char[contentsLength],
>-					0,
>-					contentsLength);
>-			}
>-		} else {
>-			contents = new char[length];
>-			int len = 0;
>-			int readSize = 0;
>-			while ((readSize != -1) && (len != length)) {
>-				// See PR 1FMS89U
>-				// We record first the read size. In this case len is the actual read size.
>-				len += readSize;
>-				readSize = reader.read(contents, len, length - len);
>+				if (totalRead + 1 + amountRequested > contents.length)
>+					System.arraycopy(contents, 	0, 	contents = new char[totalRead + 1 + amountRequested], 0, totalRead);
>+				
>+				// add current character
>+				contents[totalRead++] = (char) current; // coming from totalRead==length
> 			}
>-			// Do not keep first character for UTF-8 BOM encoding
>-			int start = 0;
>-			if (length > 0 && UTF_8.equals(encoding)) {
>-				if (contents[0] == 0xFEFF) { // if BOM char then skip
>-					len--;
>-					start = 1;
>-				}
>+			// read as many chars as possible
>+			int amountRead = reader.read(contents, totalRead, amountRequested);
>+			if (amountRead < 0) break;
>+			totalRead += amountRead;
>+		}
>+
>+		// Do not keep first character for UTF-8 BOM encoding
>+		int start = 0;
>+		if (totalRead > 0 && UTF_8.equals(encoding)) {
>+			if (contents[0] == 0xFEFF) { // if BOM char then skip
>+				totalRead--;
>+				start = 1;
> 			}
>-			// See PR 1FMS89U
>-			// Now we need to resize in case the default encoding used more than one byte for each
>-			// character
>-			if (len != length)
>-				System.arraycopy(contents, start, (contents = new char[len]), 0, len);
> 		}
>+		
>+		// resize contents if necessary
>+		if (totalRead < contents.length)
>+			System.arraycopy(contents, start, contents = new char[totalRead], 	0, 	totalRead);
> 
> 		return contents;
> 	}
>#P org.eclipse.jdt.core.tests.model
>Index: src/org/eclipse/jdt/core/tests/model/EncodingTests.java
>===================================================================
>RCS file: /cvsroot/eclipse/org.eclipse.jdt.core.tests.model/src/org/eclipse/jdt/core/tests/model/EncodingTests.java,v
>retrieving revision 1.42
>diff -u -r1.42 EncodingTests.java
>--- src/org/eclipse/jdt/core/tests/model/EncodingTests.java	29 Mar 2006 04:03:07 -0000	1.42
>+++ src/org/eclipse/jdt/core/tests/model/EncodingTests.java	4 Jul 2006 12:23:53 -0000
>@@ -12,6 +12,7 @@
> 
> import java.io.File;
> import java.io.IOException;
>+import java.io.InputStream;
> import java.io.UnsupportedEncodingException;
> 
> import junit.framework.Test;
>@@ -672,6 +673,50 @@
> 		}
> 	}
> 
>+	/*
>+	 * Ensures that an encoding that a file using an encoding producing more charaters than the file size can
>+	 * be correctly read.
>+	 * (regression test for bug 149028 Limiting number of characters to read with the file size is invalid.)
>+	 */
>+	public void test034() throws CoreException, IOException {
>+		try {
>+			// Create file
>+			IFile file = createFile("/Encoding/Test34.txt", "acegikm");
>+			
>+			// Read file using a transformation where a character is read and the next alphabetical character is
>+			// automaticaly added
>+			final InputStream fileStream = file.getContents();
>+			try {
>+				InputStream in = new InputStream() {
>+					int current = -1;
>+					public int read() throws IOException {
>+						int result;
>+						if (current != -1) {
>+							result = current;
>+							current = -1;
>+						} else {
>+							result = fileStream.read();
>+							if (result == -1)
>+								return -1;
>+							current = result + 1;
>+						}
>+						return result;
>+					}
>+				};
>+				char[] result = org.eclipse.jdt.internal.compiler.util.Util.getInputStreamAsCharArray(in, (int) file.getLocation().toFile().length(), "UTF-8");
>+				assertSourceEquals(
>+					"Unexpected source",
>+					"abcdefghijklmn",
>+					new String(result)					
>+				);
>+			} finally {
>+				fileStream.close();
>+			}
>+		} finally {
>+			deleteFile("Encoding/Test34.txt");
>+		}
>+	}	
>+	
> 	/**
> 	 * Bug 66898: refactor-rename: encoding is not preserved
> 	 * @see "http://bugs.eclipse.org/bugs/show_bug.cgi?id=66898"

Actions: View | Diff

Attachments on bug 149028: 45560 | 45701