Merge branch '2.11' into 2.12

cowtowncoder · cowtowncoder · commit 21efc0773bb9 · 2020-09-23T19:27:26.000-07:00
diff --git a/csv/src/main/java/com/fasterxml/jackson/dataformat/csv/impl/CsvEncoder.java b/csv/src/main/java/com/fasterxml/jackson/dataformat/csv/impl/CsvEncoder.java
@@ -17,20 +17,19 @@
  */
 public class CsvEncoder
 {
-
-    /*
-     * default set of escaped characters.
-     */
+    // Default set of escaped characters (none)
     private static final int [] sOutputEscapes = new int[0];
 
     final protected static char[] HEX_CHARS = CharTypes.copyHexChars();
 
-    /* As an optimization we try coalescing short writes into
+    /**
+     * As an optimization we try coalescing short writes into
      * buffer; but pass longer directly.
      */
     final protected static int SHORT_WRITE = 32;
 
-    /* Also: only do check for optional quotes for short
+    /**
+     * Also: only do check for optional quotes for short
      * values; longer ones will always be quoted.
      */
     final protected static int MAX_QUOTE_CHECK = 24;
@@ -221,9 +220,9 @@ public CsvEncoder(IOContext ctxt, int csvFeatures, Writer out, CsvSchema schema)
         _cfgMaxQuoteCheckChars = MAX_QUOTE_CHECK;
 
         _cfgQuoteCharEscapeChar = _getQuoteCharEscapeChar(
-          _cfgEscapeQuoteCharWithEscapeChar,
-          _cfgQuoteCharacter,
-          _cfgEscapeCharacter
+                _cfgEscapeQuoteCharWithEscapeChar,
+                _cfgQuoteCharacter,
+                _cfgEscapeCharacter
         );
 
         _cfgControlCharEscapeChar = _cfgEscapeCharacter > 0 ? (char) _cfgEscapeCharacter : '\\';
@@ -256,31 +255,31 @@ public CsvEncoder(CsvEncoder base, CsvSchema newSchema)
         _cfgMinSafeChar = _calcSafeChar();
         _columnCount = newSchema.size();
         _cfgQuoteCharEscapeChar = _getQuoteCharEscapeChar(
-          base._cfgEscapeQuoteCharWithEscapeChar,
-          newSchema.getQuoteChar(),
-          newSchema.getEscapeChar()
+                base._cfgEscapeQuoteCharWithEscapeChar,
+                newSchema.getQuoteChar(),
+                newSchema.getEscapeChar()
         );
         _cfgControlCharEscapeChar = _cfgEscapeCharacter > 0 ? (char) _cfgEscapeCharacter : '\\';
     }
 
     private final char _getQuoteCharEscapeChar(
-        final boolean escapeQuoteCharWithEscapeChar,
-        final int quoteCharacter,
-        final int escapeCharacter) {
-
-      final char quoteEscapeChar;
-
-      if (_cfgEscapeQuoteCharWithEscapeChar && _cfgEscapeCharacter > 0) {
-        quoteEscapeChar = (char) _cfgEscapeCharacter;
-      }
-      else if (_cfgQuoteCharacter > 0) {
-        quoteEscapeChar = (char) _cfgQuoteCharacter;
-      }
-      else {
-        quoteEscapeChar = '\\';
-      }
-
-      return quoteEscapeChar;
+            final boolean escapeQuoteCharWithEscapeChar,
+            final int quoteCharacter,
+            final int escapeCharacter)
+    {
+        final char quoteEscapeChar;
+
+        if (_cfgEscapeQuoteCharWithEscapeChar && _cfgEscapeCharacter > 0) {
+            quoteEscapeChar = (char) _cfgEscapeCharacter;
+        }
+        else if (_cfgQuoteCharacter > 0) {
+            quoteEscapeChar = (char) _cfgQuoteCharacter;
+        }
+        else {
+            quoteEscapeChar = '\\';
+        }
+
+        return quoteEscapeChar;
     }
 
     private final int _calcSafeChar()
@@ -289,6 +288,8 @@ private final int _calcSafeChar()
         int min = Math.max(_cfgColumnSeparator, _cfgQuoteCharacter);
         // 06-Nov-2015, tatu: We will NOT apply escape character, because it usually
         //    has higher ascii value (with backslash); better handle separately.
+        // 23-Sep-2020, tatu: Should not actually need to consider anything but the
+        //    first character when checking... but leaving rest for now
         for (int i = 0; i < _cfgLineSeparatorLength; ++i) {
             min = Math.max(min, _cfgLineSeparator[i]);
         }
@@ -312,7 +313,7 @@ public CsvEncoder overrideFormatFeatures(int feat) {
         return this;
     }
 
-    public CsvEncoder setOutputEscapes(int [] esc) {
+    public CsvEncoder setOutputEscapes(int[] esc) {
         _outputEscapes = (esc != null) ? esc : sOutputEscapes;
         return this;
     }
@@ -934,6 +935,7 @@ private final void _writeLongQuotedAndEscaped(String text, char esc) throws IOEx
         final int len = text.length();
         // NOTE: caller should guarantee quote char is valid (not -1) at this point:
         final char q = (char) _cfgQuoteCharacter;
+        // 23-Sep-2020, tatu: Why was this defined but not used? Commented out in 2.11.3
 //        final char quoteEscape = _cfgEscapeQuoteCharWithEscapeChar ? esc : q;
         for (int i = 0; i < len; ++i) {
             if (_outputTail >= _outputEnd) {
@@ -1063,7 +1065,7 @@ protected final boolean _needsQuotingLoose(String value, int esc)
         }
         return false;
     }
-    
+
     /**
      * @since 2.4
      */
@@ -1073,12 +1075,16 @@ protected boolean _needsQuotingStrict(String value)
 
         final int[] escCodes = _outputEscapes;
         final int escLen = escCodes.length;
+        // 23-Sep-2020, tatu: [dataformats-text#217] Must also ensure line separator
+        //   leads to quoting
+        final int lfFirst = (_cfgLineSeparatorLength == 0) ? 0 : _cfgLineSeparator[0];
 
         for (int i = 0, len = value.length(); i < len; ++i) {
             int c = value.charAt(i);
             if (c < minSafe) {
                 if (c == _cfgColumnSeparator || c == _cfgQuoteCharacter
                         || (c < escLen && escCodes[c] != 0)
+                        || (c == lfFirst)
                         // 31-Dec-2014, tatu: Comment lines start with # so quote if starts with #
                         || (c == '#' && i == 0)) {
                     return true;
@@ -1094,15 +1100,18 @@ protected boolean _needsQuotingStrict(String value)
     protected boolean _needsQuotingStrict(String value, int esc)
     {
         final int minSafe = _cfgMinSafeChar;
-
         final int[] escCodes = _outputEscapes;
         final int escLen = escCodes.length;
+        // 23-Sep-2020, tatu: [dataformats-text#217] Must also ensure line separator
+        //   leads to quoting
+        final int lfFirst = (_cfgLineSeparatorLength == 0) ? 0 : _cfgLineSeparator[0];
 
         for (int i = 0, len = value.length(); i < len; ++i) {
             int c = value.charAt(i);
             if (c < minSafe) {
                 if (c == _cfgColumnSeparator || c == _cfgQuoteCharacter
                         || (c < escLen && escCodes[c] != 0)
+                        || (c == lfFirst)
                         // 31-Dec-2014, tatu: Comment lines start with # so quote if starts with #
                         || (c == '#' && i == 0)) {
                     return true;
diff --git a/csv/src/test/java/com/fasterxml/jackson/dataformat/csv/deser/TrailingCommaCSVTest.java b/csv/src/test/java/com/fasterxml/jackson/dataformat/csv/deser/TrailingCommaCSVTest.java
@@ -1,7 +1,5 @@
 package com.fasterxml.jackson.dataformat.csv.deser;
 
-import java.util.List;
-
 import com.fasterxml.jackson.annotation.JsonPropertyOrder;
 import com.fasterxml.jackson.databind.JsonMappingException;
 import com.fasterxml.jackson.databind.MappingIterator;
diff --git a/csv/src/test/java/com/fasterxml/jackson/dataformat/csv/ser/GeneratorQuotingTest.java b/csv/src/test/java/com/fasterxml/jackson/dataformat/csv/ser/GeneratorQuotingTest.java
@@ -0,0 +1,73 @@
+package com.fasterxml.jackson.dataformat.csv.ser;
+
+import com.fasterxml.jackson.dataformat.csv.*;
+
+public class GeneratorQuotingTest extends ModuleTestBase
+{
+    /*
+    /**********************************************************************
+    /* Test methods
+    /**********************************************************************
+     */
+
+    private final CsvMapper MAPPER = mapperForCsv();
+
+    // [dataformats-text#220]
+    public void testQuotingOfLinefeedsStd() throws Exception
+    {
+        final CsvSchema schema = MAPPER.schemaFor(IdDesc.class)
+                .withLineSeparator("\n");
+        String csv;
+
+        csv = MAPPER.writer(schema)
+                .without(CsvGenerator.Feature.STRICT_CHECK_FOR_QUOTING)
+                .writeValueAsString(new IdDesc("12\n3", "Foo"));
+        assertEquals("\"12\n3\",Foo\n", csv);
+
+        csv = MAPPER.writer(schema.withEscapeChar('\\'))
+                .without(CsvGenerator.Feature.STRICT_CHECK_FOR_QUOTING)
+                .writeValueAsString(new IdDesc("12\n3", "Foo"));
+        assertEquals("\"12\n3\",Foo\n", csv);
+        
+        csv = MAPPER.writer(schema)
+                .with(CsvGenerator.Feature.STRICT_CHECK_FOR_QUOTING)
+                .writeValueAsString(new IdDesc("12\n3", "Foo"));
+        assertEquals("\"12\n3\",Foo\n", csv);
+        csv = MAPPER.writer(schema.withEscapeChar('\\'))
+                .with(CsvGenerator.Feature.STRICT_CHECK_FOR_QUOTING)
+                .writeValueAsString(new IdDesc("12\n3", "Foo"));
+        assertEquals("\"12\n3\",Foo\n", csv);
+    }
+
+    public void testQuotingOfLinefeedsCustom() throws Exception
+    {
+        // '-' is bigger than max('"', ','):
+        final CsvSchema schema = MAPPER.schemaFor(IdDesc.class)
+                .withLineSeparator("-");
+        final IdDesc value = new IdDesc("12-3", "Foo");
+
+        // with loose(default) quoting
+        String csv = MAPPER.writer(schema)
+                .without(CsvGenerator.Feature.STRICT_CHECK_FOR_QUOTING)
+                .writeValueAsString(value);
+        assertEquals("\"12-3\",Foo-", csv);
+
+        // with loose(default) quoting and escape char
+        csv = MAPPER.writer(schema.withEscapeChar('\\'))
+            .without(CsvGenerator.Feature.STRICT_CHECK_FOR_QUOTING)
+            .writeValueAsString(value);
+        assertEquals("\"12-3\",Foo-", csv);
+
+        // with strict/optimal
+        csv = MAPPER.writer(schema)
+                .with(CsvGenerator.Feature.STRICT_CHECK_FOR_QUOTING)
+                .writeValueAsString(value);
+        assertEquals("\"12-3\",Foo-", csv);
+
+        // with strict/optimal and escape char
+        csv = MAPPER.writer(schema.withEscapeChar('\\'))
+                .with(CsvGenerator.Feature.STRICT_CHECK_FOR_QUOTING)
+                .writeValueAsString(value);
+        assertEquals("\"12-3\",Foo-", csv);
+    }
+}
diff --git a/csv/src/test/java/com/fasterxml/jackson/dataformat/csv/ser/TestGenerator.java b/csv/src/test/java/com/fasterxml/jackson/dataformat/csv/ser/TestGenerator.java
@@ -8,7 +8,6 @@
 import com.fasterxml.jackson.core.JsonGenerator;
 
 import com.fasterxml.jackson.databind.JsonMappingException;
-import com.fasterxml.jackson.databind.ObjectMapper;
 import com.fasterxml.jackson.databind.ObjectWriter;
 import com.fasterxml.jackson.databind.node.ObjectNode;
 import com.fasterxml.jackson.dataformat.csv.*;
diff --git a/release-notes/VERSION-2.x b/release-notes/VERSION-2.x
@@ -19,9 +19,15 @@ Modules:
  (requested by pnepywoda@github)
 - Add Gradle Module Metadata (https://blog.gradle.org/alignment-with-gradle-module-metadata)
 
+2.11.3 (not yet released)
+
+#217: (csv) Should quote strings with line separator under STRICT_CHECK_FOR_QUOTING
+  mode
+ (reported, fix contributed by wkwkhautbois@github)
+
 2.11.2 (02-Aug-2020)
 
-#204: `CsvParser.Feature.ALLOW_TRAILING_COMMA` doesn't work with header columns
+#204: (csv) `CsvParser.Feature.ALLOW_TRAILING_COMMA` doesn't work with header columns
  (reported by Björn M)
 
 2.11.1 (25-Jun-2020)
