Add unescapeCsvFields to parse a CSV line and implement CombinedHttpH…

…eaders.getAll

Motivation:

See netty#4855

Modifications:

Unfortunately, unescapeCsv cannot be used here because the input could be a CSV line like `"a,b",c`. Hence this patch adds unescapeCsvFields to parse a CSV line and split it into multiple fields and unescaped them. The unit tests should define the behavior of unescapeCsvFields.

Then this patch just uses unescapeCsvFields to implement `CombinedHttpHeaders.getAll`.

Result:

`CombinedHttpHeaders.getAll` will return the unescaped values of a header.

codec-http/src/main/java/io/netty/handler/codec/http/CombinedHttpHeaders.java

@@ -23,6 +23,7 @@
 
 import java.util.Collection;
 import java.util.Iterator;
+import java.util.List;
 import java.util.Map;
 
 import static io.netty.util.AsciiString.CASE_INSENSITIVE_HASHER;
@@ -77,6 +78,18 @@ public CombinedHttpHeadersImpl(HashingStrategy<CharSequence> nameHashingStrategy
             super(nameHashingStrategy, valueConverter, nameValidator);
         }
 
+        @Override
+        public List<CharSequence> getAll(CharSequence name) {
+            List<CharSequence> values = super.getAll(name);
+            if (values.isEmpty()) {
+                return values;
+            }
+            if (values.size() != 1) {
+                throw new IllegalStateException("CombinedHttpHeaders should only have one value");
+            }
+            return StringUtil.unescapeCsvFields(values.get(0));
+        }
+
         @Override
         public CombinedHttpHeadersImpl add(Headers<? extends CharSequence, ? extends CharSequence, ?> headers) {
             // Override the fast-copy mechanism used by DefaultHeaders
@@ -158,6 +171,12 @@ public CombinedHttpHeadersImpl set(CharSequence name, Iterable<? extends CharSeq
             return this;
         }
 
+        @Override
+        public CombinedHttpHeadersImpl setObject(CharSequence name, Object value) {
+            super.set(name, commaSeparate(objectEscaper(), value));
+            return this;
+        }
+
         @Override
         public CombinedHttpHeadersImpl setObject(CharSequence name, Object... values) {
             super.set(name, commaSeparate(objectEscaper(), values));

codec-http/src/test/java/io/netty/handler/codec/http/CombinedHttpHeadersTest.java

@@ -18,6 +18,7 @@
 import io.netty.handler.codec.http.HttpHeadersTestUtils.HeaderValue;
 import org.junit.Test;
 
+import java.util.Arrays;
 import java.util.Collections;
 
 import static io.netty.util.AsciiString.contentEquals;
@@ -101,15 +102,15 @@ public void addCharSequencesCsvWithValueContainingComma() {
         final CombinedHttpHeaders headers = newCombinedHttpHeaders();
         headers.add(HEADER_NAME, HeaderValue.SIX_QUOTED.subset(4));
         assertTrue(contentEquals(HeaderValue.SIX_QUOTED.subsetAsCsvString(4), headers.get(HEADER_NAME)));
-        assertTrue(contentEquals(HeaderValue.SIX_QUOTED.subsetAsCsvString(4), headers.getAll(HEADER_NAME).get(0)));
+        assertEquals(HeaderValue.SIX_QUOTED.subset(4), headers.getAll(HEADER_NAME));
     }
 
     @Test
     public void addCharSequencesCsvWithValueContainingCommas() {
         final CombinedHttpHeaders headers = newCombinedHttpHeaders();
         headers.add(HEADER_NAME, HeaderValue.EIGHT.subset(6));
         assertTrue(contentEquals(HeaderValue.EIGHT.subsetAsCsvString(6), headers.get(HEADER_NAME)));
-        assertTrue(contentEquals(HeaderValue.EIGHT.subsetAsCsvString(6), headers.getAll(HEADER_NAME).get(0)));
+        assertEquals(HeaderValue.EIGHT.subset(6), headers.getAll(HEADER_NAME));
     }
 
     @Test (expected = NullPointerException.class)
@@ -168,7 +169,7 @@ public void addIterableCsvSingleValue() {
     public void addIterableCsvEmtpy() {
         final CombinedHttpHeaders headers = newCombinedHttpHeaders();
         headers.add(HEADER_NAME, Collections.<CharSequence>emptyList());
-        assertTrue(contentEquals("", headers.getAll(HEADER_NAME).get(0)));
+        assertEquals(Arrays.asList(""), headers.getAll(HEADER_NAME));
     }
 
     @Test
@@ -234,7 +235,7 @@ private static CombinedHttpHeaders newCombinedHttpHeaders() {
 
     private static void assertCsvValues(final CombinedHttpHeaders headers, final HeaderValue headerValue) {
         assertTrue(contentEquals(headerValue.asCsv(), headers.get(HEADER_NAME)));
-        assertTrue(contentEquals(headerValue.asCsv(), headers.getAll(HEADER_NAME).get(0)));
+        assertEquals(headerValue.asList(), headers.getAll(HEADER_NAME));
     }
 
     private static void assertCsvValue(final CombinedHttpHeaders headers, final HeaderValue headerValue) {
@@ -253,4 +254,21 @@ private static void addObjectValues(final CombinedHttpHeaders headers, HeaderVal
             headers.add(HEADER_NAME, v.toString());
         }
     }
+
+    @Test
+    public void testGetAll() {
+        final CombinedHttpHeaders headers = newCombinedHttpHeaders();
+        headers.set(HEADER_NAME, Arrays.asList("a", "b", "c"));
+        assertEquals(Arrays.asList("a", "b", "c"), headers.getAll(HEADER_NAME));
+        headers.set(HEADER_NAME, Arrays.asList("a,", "b,", "c,"));
+        assertEquals(Arrays.asList("a,", "b,", "c,"), headers.getAll(HEADER_NAME));
+        headers.set(HEADER_NAME, Arrays.asList("a\"", "b\"", "c\""));
+        assertEquals(Arrays.asList("a\"", "b\"", "c\""), headers.getAll(HEADER_NAME));
+        headers.set(HEADER_NAME, Arrays.asList("\"a\"", "\"b\"", "\"c\""));
+        assertEquals(Arrays.asList("a", "b", "c"), headers.getAll(HEADER_NAME));
+        headers.set(HEADER_NAME, "a,b,c");
+        assertEquals(Arrays.asList("a,b,c"), headers.getAll(HEADER_NAME));
+        headers.set(HEADER_NAME, "\"a,b,c\"");
+        assertEquals(Arrays.asList("a,b,c"), headers.getAll(HEADER_NAME));
+    }
 }

common/src/main/java/io/netty/util/internal/StringUtil.java

@@ -16,6 +16,7 @@
 package io.netty.util.internal;
 
 import java.io.IOException;
+import java.util.ArrayList;
 import java.util.Formatter;
 import java.util.List;
 
@@ -425,6 +426,76 @@ public static CharSequence unescapeCsv(CharSequence value) {
     }
 
     /**
+     * Unescapes the specified escaped CSV fields according to
+     * <a href="https://tools.ietf.org/html/rfc4180#section-2">RFC-4180</a>.
+     *
+     * @param value A string with multiple CSV escaped fields which will be unescaped according to
+     *              <a href="https://tools.ietf.org/html/rfc4180#section-2">RFC-4180</a>
+     * @return {@link List} the list of unescaped fields
+     */
+    public static List<CharSequence> unescapeCsvFields(CharSequence value) {
+        List<CharSequence> unescaped = new ArrayList<CharSequence>(2);
+        StringBuilder current = InternalThreadLocalMap.get().stringBuilder();
+        boolean quoted = false;
+        int last = value.length() - 1;
+        for (int i = 0; i <= last; i++) {
+            char c = value.charAt(i);
+            if (quoted) {
+                switch (c) {
+                    case DOUBLE_QUOTE:
+                        if (i == last) {
+                            // Add the last field and return
+                            unescaped.add(current.toString());
+                            return unescaped;
+                        }
+                        char next = value.charAt(++i);
+                        if (next == DOUBLE_QUOTE) {
+                            // 2 double-quotes should be unescaped to one
+                            current.append(DOUBLE_QUOTE);
+                            break;
+                        }
+                        if (next == COMMA) {
+                            // This is the end of a field. Let's start to parse the next field.
+                            quoted = false;
+                            unescaped.add(current.toString());
+                            current.setLength(0);
+                            break;
+                        }
+                        // double-quote followed by other character is invalid
+                        throw newInvalidEscapedCsvFieldException(value, i - 1);
+                    default:
+                        current.append(c);
+                }
+            } else {
+                switch (c) {
+                    case COMMA:
+                        // Start to parse the next field
+                        unescaped.add(current.toString());
+                        current.setLength(0);
+                        break;
+                    case DOUBLE_QUOTE:
+                        if (current.length() == 0) {
+                            quoted = true;
+                            break;
+                        }
+                        // double-quote appears without being enclosed with double-quotes
+                    case LINE_FEED:
+                    case CARRIAGE_RETURN:
+                        // special characters appears without being enclosed with double-quotes
+                        throw newInvalidEscapedCsvFieldException(value, i);
+                    default:
+                        current.append(c);
+                }
+            }
+        }
+        if (quoted) {
+            throw newInvalidEscapedCsvFieldException(value, last);
+        }
+        unescaped.add(current.toString());
+        return unescaped;
+    }
+
+    /**s
      * Validate if {@code value} is a valid csv field without double-quotes.
      *
      * @throws IllegalArgumentException if {@code value} needs to be encoded with double-quotes.

common/src/test/java/io/netty/util/internal/StringUtilTest.java

@@ -15,6 +15,7 @@
  */
 package io.netty.util.internal;
 
+import java.util.Arrays;
 import org.junit.Test;
 
 import static io.netty.util.internal.StringUtil.*;
@@ -376,6 +377,47 @@ private void assertEscapeCsvAndUnEscapeCsv(String value) {
         assertEquals(value, unescapeCsv(StringUtil.escapeCsv(value)));
     }
 
+    @Test
+    public void testUnescapeCsvFields() {
+        assertEquals(Arrays.asList(""), unescapeCsvFields(""));
+        assertEquals(Arrays.asList("", ""), unescapeCsvFields(","));
+        assertEquals(Arrays.asList("a", ""), unescapeCsvFields("a,"));
+        assertEquals(Arrays.asList("", "a"), unescapeCsvFields(",a"));
+        assertEquals(Arrays.asList("\""), unescapeCsvFields("\"\"\"\""));
+        assertEquals(Arrays.asList("\"", "\""), unescapeCsvFields("\"\"\"\",\"\"\"\""));
+        assertEquals(Arrays.asList("netty"), unescapeCsvFields("netty"));
+        assertEquals(Arrays.asList("hello", "netty"), unescapeCsvFields("hello,netty"));
+        assertEquals(Arrays.asList("hello,netty"), unescapeCsvFields("\"hello,netty\""));
+        assertEquals(Arrays.asList("hello", "netty"), unescapeCsvFields("\"hello\",\"netty\""));
+        assertEquals(Arrays.asList("a\"b", "c\"d"), unescapeCsvFields("\"a\"\"b\",\"c\"\"d\""));
+        assertEquals(Arrays.asList("a\rb", "c\nd"), unescapeCsvFields("\"a\rb\",\"c\nd\""));
+    }
+
+    @Test(expected = IllegalArgumentException.class)
+    public void unescapeCsvFieldsWithCRWithoutQuote() {
+        unescapeCsvFields("a,\r");
+    }
+
+    @Test(expected = IllegalArgumentException.class)
+    public void unescapeCsvFieldsWithLFWithoutQuote() {
+        unescapeCsvFields("a,\r");
+    }
+
+    @Test(expected = IllegalArgumentException.class)
+    public void unescapeCsvFieldsWithQuote() {
+        unescapeCsvFields("a,\"");
+    }
+
+    @Test(expected = IllegalArgumentException.class)
+    public void unescapeCsvFieldsWithQuote2() {
+        unescapeCsvFields("\",a");
+    }
+
+    @Test(expected = IllegalArgumentException.class)
+    public void unescapeCsvFieldsWithQuote3() {
+        unescapeCsvFields("a\"b,a");
+    }
+
     @Test
     public void testSimpleClassName() throws Exception {
         testSimpleClassName(String.class);