[#2544] Correctly parse Multipart-mixed POST HTTP request in case of entity ends with odd number of 0x0D. Port of @fredericBregier 's work.

Motivation:
When an attribute is ending with an odd number of CR (0x0D), the decoder
add an extra CR in the decoded attribute and should not.

Modifications:
Each time a CR is detected, the next byte was tested to be LF or not. If
not, in a number of places, the CR byte was lost while it should not be.
When a CR is detected, if the next byte is not LF, the CR byte should be
saved as the position point to the next byte (not LF). When a CR is
detected, if there is not yet other available bytes, the position is
reset to the position of CR (since a LF could follow).

A new Junit test case is added, using DECODER and variable number of CR
in the final attribute (testMultipartCodecWithCRasEndOfAttribute).

Result:
The attribute is now correctly decoded with the right number of CR
ending bytes.

codec-http/src/main/java/io/netty/handler/codec/http/multipart/HttpPostRequestDecoder.java

@@ -1287,9 +1287,15 @@ public class HttpPostRequestDecoder {
             while (undecodedChunk.isReadable()) {
                 byte nextByte = undecodedChunk.readByte();
                 if (nextByte == HttpConstants.CR) {
-                    nextByte = undecodedChunk.readByte();
+                    // check but do not changed readerIndex
+                    nextByte = undecodedChunk.getByte(undecodedChunk.readerIndex());
                     if (nextByte == HttpConstants.LF) {
+                        // Skip
+                        undecodedChunk.skipBytes(1);
                         return line.toString(charset);
+                    } else {
+                        // Write CR (not followed by LF)
+                        line.writeByte(HttpConstants.CR);
                     }
                 } else if (nextByte == HttpConstants.LF) {
                     return line.toString(charset);
@@ -1332,6 +1338,10 @@ public class HttpPostRequestDecoder {
                         if (nextByte == HttpConstants.LF) {
                             sao.setReadPosition(0);
                             return line.toString(charset);
+                        } else {
+                            // Write CR (not followed by LF)
+                            sao.pos--;
+                            line.writeByte(HttpConstants.CR);
                         }
                     } else {
                         line.writeByte(nextByte);
@@ -1494,6 +1504,11 @@ public class HttpPostRequestDecoder {
                         if (nextByte == HttpConstants.LF) {
                             sao.setReadPosition(0);
                             return sb.toString();
+                        } else {
+                            // error CR without LF
+                            // delimiter not found so break here !
+                            undecodedChunk.readerIndex(readerIndex);
+                            throw new NotEnoughDataDecoderException();
                         }
                     } else {
                         // error since CR must be followed by LF
@@ -1522,6 +1537,11 @@ public class HttpPostRequestDecoder {
                                         if (nextByte == HttpConstants.LF) {
                                             sao.setReadPosition(0);
                                             return sb.toString();
+                                        } else {
+                                            // error CR without LF
+                                            // delimiter not found so break here !
+                                            undecodedChunk.readerIndex(readerIndex);
+                                            throw new NotEnoughDataDecoderException();
                                         }
                                     } else {
                                         // error CR without LF
@@ -1826,7 +1846,13 @@ public class HttpPostRequestDecoder {
                                     newLine = true;
                                     index = 0;
                                     lastPosition = undecodedChunk.readerIndex() - 2;
+                                }  else {
+                                    // Unread second nextByte
+                                    lastPosition = undecodedChunk.readerIndex() - 1;
+                                    undecodedChunk.readerIndex(lastPosition);
                                 }
+                            } else {
+                                lastPosition = undecodedChunk.readerIndex() - 1;
                             }
                         } else if (nextByte == HttpConstants.LF) {
                             newLine = true;
@@ -1845,7 +1871,13 @@ public class HttpPostRequestDecoder {
                                 newLine = true;
                                 index = 0;
                                 lastPosition = undecodedChunk.readerIndex() - 2;
+                            } else {
+                                // Unread second nextByte
+                                lastPosition = undecodedChunk.readerIndex() - 1;
+                                undecodedChunk.readerIndex(lastPosition);
                             }
+                        } else {
+                            lastPosition = undecodedChunk.readerIndex() - 1;
                         }
                     } else if (nextByte == HttpConstants.LF) {
                         newLine = true;
@@ -1930,6 +1962,10 @@ public class HttpPostRequestDecoder {
                                     newLine = true;
                                     index = 0;
                                     lastrealpos = sao.pos - 2;
+                                } else {
+                                    // Unread last nextByte
+                                    sao.pos--;
+                                    lastrealpos = sao.pos;
                                 }
                             }
                         } else if (nextByte == HttpConstants.LF) {
@@ -1949,6 +1985,10 @@ public class HttpPostRequestDecoder {
                                 newLine = true;
                                 index = 0;
                                 lastrealpos = sao.pos - 2;
+                            } else {
+                                // Unread last nextByte
+                                sao.pos--;
+                                lastrealpos = sao.pos;
                             }
                         }
                     } else if (nextByte == HttpConstants.LF) {

codec-http/src/test/java/io/netty/handler/codec/http/multipart/HttpPostRequestDecoderTest.java

@@ -129,6 +129,53 @@ public class HttpPostRequestDecoderTest {
         decoder.destroy();
     }
 
+    // See https://github.com/netty/netty/issues/2544
+    @Test
+    public void testMultipartCodecWithCRasEndOfAttribute() throws Exception {
+        final String boundary = "dLV9Wyq26L_-JQxk6ferf-RT153LhOO";
+
+        // Force to use memory-based data.
+        final DefaultHttpDataFactory inMemoryFactory = new DefaultHttpDataFactory(false);
+        // Build test case
+        String extradata = "aaaa";
+        String [] datas = new String[5];
+        for (int i = 0; i < 4; i++) {
+            datas[i] = extradata;
+            for (int j = 0; j < i ; j++) {
+                datas[i] += '\r';
+            }
+        }
+
+        for (int i = 0; i < 4; i++) {
+            final DefaultFullHttpRequest req = new DefaultFullHttpRequest(HttpVersion.HTTP_1_1, HttpMethod.POST,
+                    "http://localhost");
+            req.setDecoderResult(DecoderResult.SUCCESS);
+            req.headers().add(HttpHeaders.Names.CONTENT_TYPE, "multipart/form-data; boundary=" + boundary);
+            req.headers().add(HttpHeaders.Names.TRANSFER_ENCODING, HttpHeaders.Values.CHUNKED);
+            final String body =
+                    "--" + boundary + "\r\n" +
+                            "Content-Disposition: form-data; name=\"file" + i + "\"\r\n" +
+                            "Content-Type: image/gif\r\n" +
+                            "\r\n" +
+                            datas[i] + "\r\n" +
+                            "--" + boundary + "--\r\n";
+
+            req.content().writeBytes(body.getBytes(CharsetUtil.UTF_8));
+            // Create decoder instance to test.
+            final HttpPostRequestDecoder decoder = new HttpPostRequestDecoder(inMemoryFactory, req);
+            assertFalse(decoder.getBodyHttpDatas().isEmpty());
+            // Check correctness: data size
+            InterfaceHttpData httpdata = decoder.getBodyHttpData("file" + i);
+            assertNotNull(httpdata);
+            Attribute attribute = (Attribute) httpdata;
+            byte []datar = attribute.get();
+            assertNotNull(datar);
+            assertEquals(datas[i].getBytes(CharsetUtil.UTF_8).length, datar.length);
+
+            decoder.destroy();
+        }
+    }
+
     // See https://github.com/netty/netty/issues/1848
     @Test
     public void testNoZeroOut() throws Exception {