重构 MultiPart 解析器 (#99)

scx567888 · Oct 14, 2024 · ed6ff3f · ed6ff3f
1 parent 4629cc4
commit ed6ff3f
Show file tree

Hide file tree

Showing 9 changed files with 94 additions and 48 deletions.
diff --git a/scx-http/src/main/java/cool/scx/http/ScxHttpHeadersHelper.java b/scx-http/src/main/java/cool/scx/http/ScxHttpHeadersHelper.java
@@ -7,7 +7,7 @@ public static ScxHttpHeadersWritable parseHeaders(String headersStr) {
         var lines = headersStr.split("\r\n");
 
         for (var line : lines) {
-            int i = line.indexOf(":");
+            int i = line.indexOf(':');
             if (i != -1) {
                 var key = line.substring(0, i).trim();
                 var value = line.substring(i + 1).trim();

diff --git a/scx-http/src/main/java/cool/scx/http/media/multi_part/MultiPartStream.java b/scx-http/src/main/java/cool/scx/http/media/multi_part/MultiPartStream.java
@@ -1,6 +1,5 @@
 package cool.scx.http.media.multi_part;
 
-import cool.scx.common.util.ArrayUtils;
 import cool.scx.http.ScxHttpHeaders;
 import cool.scx.http.ScxHttpHeadersWritable;
 import cool.scx.io.InputStreamDataSupplier;
@@ -16,15 +15,12 @@ public class MultiPartStream implements MultiPart, Iterator<MultiPartPart> {
 
     protected static final byte[] CRLF_CRLF_BYTES = "\r\n\r\n".getBytes();
     protected final LinkedDataReader linkedDataReader;
-    protected final byte[] boundaryHeadCRLFBytes;
-    protected final byte[] boundaryENDBytes;
+    protected final byte[] boundaryBytes;
     protected boolean hasNextPart;
     protected String boundary;
 
     public MultiPartStream(InputStream inputStream, String boundary) {
-        var boundaryHeadBytes = ArrayUtils.concat("--".getBytes(), boundary.getBytes());
-        this.boundaryHeadCRLFBytes = ArrayUtils.concat(boundaryHeadBytes, "\r\n".getBytes());
-        this.boundaryENDBytes = ArrayUtils.concat(boundaryHeadBytes, "--".getBytes());
+        this.boundaryBytes = ("--" + boundary).getBytes();
         this.linkedDataReader = new LinkedDataReader(new InputStreamDataSupplier(inputStream));
         this.hasNextPart = readNext();
     }
@@ -34,36 +30,42 @@ public ScxHttpHeadersWritable readToHeaders() {
         // head /r/n
         // /r/n
         // content
-        var headersBytes = linkedDataReader.readMatch(CRLF_CRLF_BYTES);
+        var headersBytes = linkedDataReader.readUntil(CRLF_CRLF_BYTES);
         var headersStr = new String(headersBytes);
         return ScxHttpHeaders.of(headersStr);
     }
 
     public byte[] readContentToByte() throws IOException {
-        //我们需要查找终结点 先假设不是最后一个 那我们就需要查找下一个开始位置 
+        //因为正常的表单一定是 --xxxxxx 结尾的 所以我们只需要找 下一个分块的起始位置作为结束位置即可 
         try {
-            var i = linkedDataReader.indexOf(boundaryHeadCRLFBytes);
+            var i = linkedDataReader.indexOf(boundaryBytes);
             // i - 2 因为我们不需要读取内容结尾的 \r\n  
             var bytes = linkedDataReader.read(i - 2);
             //跳过 \r\n 方便后续读取
             linkedDataReader.skip(2);
             return bytes;
         } catch (NoMatchFoundException e) {
-            //可能是最后一个查找 最终终结点
-            var i = linkedDataReader.indexOf(boundaryENDBytes);
-            var bytes = linkedDataReader.read(i - 2);
-            //跳过 \r\n 方便后续读取
-            linkedDataReader.skip(2);
-            return bytes;
+            // 理论上一个正常的 MultiPart 不会有这种情况
+            throw new RuntimeException("异常状态 !!!");
         }
     }
 
     public boolean readNext() {
-        //查找 --xxxxxxxxx\r\n 没有代表 读取到结尾
+        //查找 --xxxxxxxxx
         try {
-            var i = linkedDataReader.indexOf(boundaryHeadCRLFBytes);
-            linkedDataReader.skip(i + boundaryHeadCRLFBytes.length);
-            return true;
+            var i = linkedDataReader.indexOf(boundaryBytes);
+            linkedDataReader.skip(i + boundaryBytes.length);
+            //向后读取两个字节 
+            var a = linkedDataReader.read();
+            var b = linkedDataReader.read();
+            // 判断 是 \r\n or -- 
+            if (a == '\r' && b == '\n') { //还有数据
+                return true;
+            } else if (a == '-' && b == '-') { // 读取到了终结符
+                return false;
+            } else { // 理论上一个正常的 MultiPart 不会有这种情况
+                throw new RuntimeException("未知字符 !!! ");
+            }
         } catch (NoMatchFoundException e) {
             return false;
         }
@@ -91,10 +93,11 @@ public MultiPartPart next() {
         }
         try {
 
+            var part = new MultiPartPartImpl();
+
             // 读取当前部分的头部信息
             var headers = readToHeaders();
-
-            var part = new MultiPartPartImpl().headers(headers);
+            part.headers(headers);
 
             //读取内容
             var content = readContentToByte();

diff --git a/scx-http/src/main/java/cool/scx/http/media/multi_part/MultiPartStreamCached.java b/scx-http/src/main/java/cool/scx/http/media/multi_part/MultiPartStreamCached.java
@@ -37,17 +37,14 @@ public Path readContentToPath(Path path) throws IOException {
         try (output) {
             //我们需要查找终结点 先假设不是最后一个 那我们就需要查找下一个开始位置 
             try {
-                var i = linkedDataReader.indexOf(boundaryHeadCRLFBytes);
+                var i = linkedDataReader.indexOf(boundaryBytes);
                 // i - 2 因为我们不需要读取内容结尾的 \r\n  
                 linkedDataReader.read(output, i - 2);
                 //跳过 \r\n 方便后续读取
                 linkedDataReader.skip(2);
             } catch (NoMatchFoundException e) {
-                //可能是最后一个查找 最终终结点
-                var i = linkedDataReader.indexOf(boundaryENDBytes);
-                linkedDataReader.read(output, i - 2);
-                //跳过 \r\n 方便后续读取
-                linkedDataReader.skip(2);
+                // 理论上一个正常的 MultiPart 不会有这种情况
+                throw new RuntimeException("异常状态 !!!");
             }
         }
 
@@ -61,10 +58,11 @@ public MultiPartPart next() {
         }
         try {
 
+            var part = new MultiPartPartImpl();
+
             // 读取当前部分的头部信息
             var headers = readToHeaders();
-
-            var part = new MultiPartPartImpl().headers(headers);
+            part.headers(headers);
 
             var b = needCached(headers);
             if (b) {
@@ -74,6 +72,7 @@ public MultiPartPart next() {
                 var content = readContentToByte();
                 part.body(content);
             }
+
             // 检查是否有下一个部分
             hasNextPart = readNext();
 

diff --git a/scx-http/src/test/java/cool/scx/http/test/HeadersTest.java b/scx-http/src/test/java/cool/scx/http/test/HeadersTest.java
@@ -0,0 +1,28 @@
+package cool.scx.http.test;
+
+import cool.scx.http.MediaType;
+import cool.scx.http.ScxHttpHeaders;
+import cool.scx.http.content_type.ContentType;
+
+import java.nio.charset.StandardCharsets;
+
+public class HeadersTest {
+
+    public static void main(String[] args) {
+        test1();
+    }
+
+    public static void test1() {
+        long l = System.nanoTime();
+        for (int i = 0; i < 9999; i++) {
+            var h = ScxHttpHeaders.of();
+            h.add("Content-Disposition", "form-data; name=myname");
+            h.contentLength(100);
+            h.contentType(ContentType.of(MediaType.APPLICATION_JSON).charset(StandardCharsets.UTF_8));
+            var s = h.encode();
+            var nw = ScxHttpHeaders.of(s);
+        }
+        System.out.println((System.nanoTime() - l) / 1000_000);
+    }
+
+}
diff --git a/scx-http/src/test/java/cool/scx/http/test/MultiPartTest.java b/scx-http/src/test/java/cool/scx/http/test/MultiPartTest.java
@@ -1,5 +1,6 @@
 package cool.scx.http.test;
 
+import cool.scx.common.util.ArrayUtils;
 import cool.scx.http.MediaType;
 import cool.scx.http.ScxHttpHeaders;
 import cool.scx.http.content_type.ContentType;
@@ -36,6 +37,9 @@ public static void test1() {
         ss.write(b);
         byte[] byteArray = b.toByteArray();
 
+        //复制两遍查看是否会产生错误的读取
+        byteArray = ArrayUtils.concat(byteArray, byteArray);
+
 
         long l = System.nanoTime();
         for (int j = 0; j < 9999; j++) {

diff --git a/scx-io/src/main/java/cool/scx/io/ByteArrayDataReader.java b/scx-io/src/main/java/cool/scx/io/ByteArrayDataReader.java
@@ -53,7 +53,7 @@ public void read(OutputStream outputStream, int maxLength) throws NoMoreDataExce
     }
 
     @Override
-    public byte get() throws NoMoreDataException {
+    public byte peek() throws NoMoreDataException {
         try {
             return bytes[position];
         } catch (ArrayIndexOutOfBoundsException e) {
@@ -62,7 +62,7 @@ public byte get() throws NoMoreDataException {
     }
 
     @Override
-    public byte[] get(int maxLength) throws NoMoreDataException {
+    public byte[] peek(int maxLength) throws NoMoreDataException {
         int availableLength = bytes.length - position;
         if (availableLength <= 0) {
             throw new NoMoreDataException();
@@ -74,7 +74,7 @@ public byte[] get(int maxLength) throws NoMoreDataException {
     }
 
     @Override
-    public void get(OutputStream outputStream, int maxLength) throws NoMoreDataException {
+    public void peek(OutputStream outputStream, int maxLength) throws NoMoreDataException {
         int availableLength = bytes.length - position;
         if (availableLength <= 0) {
             throw new NoMoreDataException();

diff --git a/scx-io/src/main/java/cool/scx/io/DataReader.java b/scx-io/src/main/java/cool/scx/io/DataReader.java
@@ -11,6 +11,7 @@ public interface DataReader {
      * 当没有更多的数据时会抛出异常
      *
      * @return byte
+     * @throws NoMoreDataException 没有更多数据时抛出
      */
     byte read() throws NoMoreDataException;
 
@@ -20,6 +21,7 @@ public interface DataReader {
      *
      * @param maxLength 最大长度
      * @return bytes
+     * @throws NoMoreDataException 没有更多数据时抛出
      */
     byte[] read(int maxLength) throws NoMoreDataException;
 
@@ -28,6 +30,7 @@ public interface DataReader {
      * 当没有更多的数据时会抛出异常
      *
      * @param maxLength 最大长度
+     * @throws NoMoreDataException 没有更多数据时抛出
      */
     void read(OutputStream outputStream, int maxLength) throws NoMoreDataException;
 
@@ -36,31 +39,35 @@ public interface DataReader {
      * 当没有更多的数据时会抛出异常
      *
      * @return byte
+     * @throws NoMoreDataException 没有更多数据时抛出
      */
-    byte get() throws NoMoreDataException;
+    byte peek() throws NoMoreDataException;
 
     /**
      * 读取指定长度字节 (指针不会移动)
      * 当没有更多的数据时会抛出异常
      *
      * @param maxLength 最大长度
      * @return byte
+     * @throws NoMoreDataException 没有更多数据时抛出
      */
-    byte[] get(int maxLength) throws NoMoreDataException;
+    byte[] peek(int maxLength) throws NoMoreDataException;
 
     /**
      * 向 outputStream 写入指定长度字节 (指针不会移动)
      * 当没有更多的数据时会抛出异常
      *
      * @param maxLength 最大长度
+     * @throws NoMoreDataException 没有更多数据时抛出
      */
-    void get(OutputStream outputStream, int maxLength) throws NoMoreDataException;
+    void peek(OutputStream outputStream, int maxLength) throws NoMoreDataException;
 
     /**
      * 查找 指定字节 第一次出现的 index (指针不会移动)
      *
      * @param b 指定字节
      * @return index 或者 -1 (未找到)
+     * @throws NoMatchFoundException 没有匹配时抛出
      */
     int indexOf(byte b) throws NoMatchFoundException;
 
@@ -69,6 +76,7 @@ public interface DataReader {
      *
      * @param b 指定字节数组
      * @return index 或者 -1 (未找到)
+     * @throws NoMatchFoundException 没有匹配时抛出
      */
     int indexOf(byte[] b) throws NoMatchFoundException;
 
@@ -84,8 +92,9 @@ public interface DataReader {
      *
      * @param b 指定字节
      * @return bytes
+     * @throws NoMatchFoundException 没有匹配时抛出
      */
-    default byte[] readMatch(byte b) throws NoMatchFoundException {
+    default byte[] readUntil(byte b) throws NoMatchFoundException {
         var index = indexOf(b);
         var data = read(index);
         skip(1);
@@ -97,8 +106,9 @@ default byte[] readMatch(byte b) throws NoMatchFoundException {
      *
      * @param b 指定字节
      * @return bytes
+     * @throws NoMatchFoundException 没有匹配时抛出
      */
-    default byte[] readMatch(byte[] b) throws NoMatchFoundException {
+    default byte[] readUntil(byte[] b) throws NoMatchFoundException {
         var index = indexOf(b);
         var data = read(index);
         skip(b.length);
@@ -110,21 +120,23 @@ default byte[] readMatch(byte[] b) throws NoMatchFoundException {
      *
      * @param b 指定字节
      * @return bytes
+     * @throws NoMatchFoundException 没有匹配时抛出
      */
-    default byte[] getMatch(byte b) throws NoMatchFoundException {
+    default byte[] peekUntil(byte b) throws NoMatchFoundException {
         var index = indexOf(b);
-        return get(index);
+        return peek(index);
     }
 
     /**
      * 读取 直到 找到匹配的字节 (指针不会移动)
      *
      * @param b 指定字节
      * @return bytes
+     * @throws NoMatchFoundException 没有匹配时抛出
      */
-    default byte[] getMatch(byte[] b) throws NoMatchFoundException {
+    default byte[] peekUntil(byte[] b) throws NoMatchFoundException {
         var index = indexOf(b);
-        return get(index);
+        return peek(index);
     }
 
 }
diff --git a/scx-io/src/main/java/cool/scx/io/LinkedDataReader.java b/scx-io/src/main/java/cool/scx/io/LinkedDataReader.java
@@ -127,13 +127,13 @@ public void read(OutputStream outputStream, int maxLength) throws NoMoreDataExce
     }
 
     @Override
-    public byte get() throws NoMoreDataException {
+    public byte peek() throws NoMoreDataException {
         ensureAvailable();
         return head.bytes[head.position];
     }
 
     @Override
-    public byte[] get(int maxLength) throws NoMoreDataException {
+    public byte[] peek(int maxLength) throws NoMoreDataException {
         ensureAvailable(); // 确保至少有一个字节可读
         var result = new byte[maxLength];
         var remaining = maxLength; //剩余字节数
@@ -169,7 +169,7 @@ public byte[] get(int maxLength) throws NoMoreDataException {
     }
 
     @Override
-    public void get(OutputStream outputStream, int maxLength) throws NoMoreDataException {
+    public void peek(OutputStream outputStream, int maxLength) throws NoMoreDataException {
         ensureAvailable(); // 确保至少有一个字节可读
         var result = new byte[maxLength];
         var remaining = maxLength; //剩余字节数

diff --git a/scx-io/src/test/java/cool/scx/io/test/DataReaderTest.java b/scx-io/src/test/java/cool/scx/io/test/DataReaderTest.java
@@ -22,15 +22,15 @@ public static void test1() {
         var dataReader = new ByteArrayDataReader("11112345678".getBytes(StandardCharsets.UTF_8));
 
         //不会影响读取
-        dataReader.get(99);
+        dataReader.peek(99);
 
         dataReader.indexOf("1".getBytes(StandardCharsets.UTF_8));
 
-        var index = dataReader.readMatch("123".getBytes());
+        var index = dataReader.readUntil("123".getBytes());
 
         try {
             //第二次应该匹配失败
-            var index1 = dataReader.readMatch("123".getBytes());
+            var index1 = dataReader.readUntil("123".getBytes());
         } catch (NoMatchFoundException _) {
 
         }