Changeset 5570

Timestamp:

Feb 16, 2011, 2:49:40 PM (11 years ago)

Author:

Nicklas Nordborg

Message:

Fixes #1581: Show common character sets first in the list of character sets

This is now a configuration option that can be set in base.config[commonCharsets]. It is also possible to specifiy a regular expression to filter character sets that are not used. The default setting will filter all character sets starting with 'x-' or 'IBM'.

Location:

trunk

Files:

• config/dist/base.config (modified) (1 diff)
• src/core/net/sf/basedb/core/Config.java (modified) (5 diffs)
• src/core/net/sf/basedb/util/ComparableComparator.java (added)
• src/core/net/sf/basedb/util/PriorityComparator.java (added)
• src/plugins/core/net/sf/basedb/plugins/util/Parameters.java (modified) (2 diffs)
• www/common/import/select_file.jsp (modified) (1 diff)
• www/common/plugin/test_with_file.jsp (modified) (1 diff)
• www/filemanager/files/edit_file.jsp (modified) (2 diffs)
• www/filemanager/files/edit_multiple_file.jsp (modified) (2 diffs)
• www/filemanager/upload/select.jsp (modified) (2 diffs)

trunk/config/dist/base.config

@@ -221 +221 @@
 # locale.country = 
 # locacle.variant =
+
+# The default character set that will be used to parse text files that
+# hasn't specified a character set. The default value for this setting
+# is ISO-8859-1.
+#
+# defaultCharset = ISO-8859-1
+
+# A comma-separated list of the most commonly used character sets for text
+# files. This list is used to order selection lists so that the common values
+# are listed at the top of the list.
+#
+# commonCharsets = UTF-8,US-ASCII,ISO-8859-1
+
+# A regular expression that used to filter character sets that are not used 
+# in order to keep the selection list shorter. If no value is specified all
+# character sets are included. The standard setting will filter out all 
+# character sets starting with 'x-' or 'IBM'. Character sets that are specified
+# in the commonCharsets setting are never filtered.
+#
+ignoreCharsets = ((x|X)\-|IBM).*
+
 
 # ===============

trunk/src/core/net/sf/basedb/core/Config.java

@@ -23 +23 @@
 package net.sf.basedb.core;
 
+import java.util.Arrays;
+import java.util.Collections;
 import java.util.Locale;
 import java.util.Properties;
+import java.util.SortedSet;
+import java.util.TreeSet;
+import java.util.regex.Pattern;
+import java.util.regex.PatternSyntaxException;
 import java.io.FileNotFoundException;
 import java.io.InputStream;
 import java.net.URL;
-
+import java.nio.charset.Charset;
+
+import org.apache.commons.collections.CollectionUtils;
+import org.apache.commons.collections.Predicate;
+
+import net.sf.basedb.util.ComparableComparator;
 import net.sf.basedb.util.FileUtil;
+import net.sf.basedb.util.PriorityComparator;
 
 /**
@@ -60 +72 @@
   private static Locale locale = Locale.getDefault();
   
+  private static SortedSet<String> commonCharsets;
+  
+  private static SortedSet<String> allCharsets;
+  
   /**
     Loads the settings from the configuration file.
@@ -93 +109 @@
       if (is != null) FileUtil.close(is);
     }
+
     isInitialised = true;
-    
+  
     // Load locale
     String language = getString("locale.language", "");
@@ -107 +124 @@
       locale = Locale.getDefault();
     }
+    
+    // Load character sets
+    String charsets = getString("commonCharsets", "UTF-8,US-ASCII,ISO-8859-1").trim();
+    commonCharsets = Collections.unmodifiableSortedSet(new TreeSet<String>(Arrays.asList(charsets.split("\\s*,\\s*"))));
+
+    SortedSet<String> tmp = new TreeSet<String>(
+        new PriorityComparator<String>(commonCharsets, 
+        new ComparableComparator<String>()));
+    tmp.addAll(Charset.availableCharsets().keySet());
+    String filter = getString("ignoreCharsets", null);
+    if (filter != null)
+    {
+      try
+      {
+        final Pattern p = Pattern.compile(filter);
+        CollectionUtils.filter(tmp, new Predicate()
+        {
+          @Override
+          public boolean evaluate(Object o)
+          {
+            return !p.matcher((String)o).matches() || commonCharsets.contains(o);
+          }
+        });
+      }
+      catch (PatternSyntaxException ex)
+      {
+        isInitialised = false;
+        throw new ConfigurationException("Invalid pattern for 'ignoreCharsets' property in '/base.config'. ", ex);
+      }
+    }
+    allCharsets = Collections.unmodifiableSortedSet(tmp);
   }
 
@@ -274 +322 @@
 
   /**
+    Get a set containing the most common character sets used by text files.
+    The set is sorted by name.
+    @return A sorted set
+    @since 2.17
+  */
+  public static SortedSet<String> getCommonCharsets()
+  {
+    return commonCharsets;
+  }
+  
+  /**
+    Get a set containing all character sets known to the current java virtual machine.
+    The set is sorted by name with the most common character sets first.
+    @return A sorted set
+    @since 2.17
+  */
+  public static SortedSet<String> getAllCharsets()
+  {
+    return allCharsets;
+  }
+  
+  /**
     Get the default locale configured for the server.
     @since 2.16

trunk/src/plugins/core/net/sf/basedb/plugins/util/Parameters.java

@@ -22 +22 @@
 package net.sf.basedb.plugins.util;
 
-import java.nio.charset.Charset;
 import java.text.SimpleDateFormat;
 import java.util.ArrayList;
@@ -307 +306 @@
   private static final StringParameterType charsetType = 
     new StringParameterType(255, Config.getCharset(), false, 1, 0, 0, 
-    new ArrayList<String>(Charset.availableCharsets().keySet()));
+    new ArrayList<String>(Config.getAllCharsets()));
     
   

trunk/www/common/import/select_file.jsp

@@ -217 +217 @@
           <%
           String defaultCharset = Config.getCharset();
-          for (String charset : Charset.availableCharsets().keySet())
+          int numCommon = Config.getCommonCharsets().size();
+          int i = 0;
+          for (String charset : Config.getAllCharsets())
           {
             %>
             <option value="<%=charset%>" 
-              <%=defaultCharset.equalsIgnoreCase(charset) ? "selected" : ""%>><%=charset%>
+              <%=defaultCharset.equalsIgnoreCase(charset) ? "selected" : ""%> 
+              <%=i==numCommon ? "style=\"border-top: 1px solid #666666;\"" : "" %>><%=charset%>
             <%
+            i++;
           }
           %>

trunk/www/common/plugin/test_with_file.jsp

@@ -339 +339 @@
           <%
           String defaultCharset = Config.getCharset();
-          for (String charset : Charset.availableCharsets().keySet())
+          int numCommon = Config.getCommonCharsets().size();
+          int i = 0;
+          for (String charset : Config.getAllCharsets())
           {
             %>
             <option value="<%=charset%>" 
-              <%=defaultCharset.equalsIgnoreCase(charset) ? "selected" : ""%>><%=charset%>
+              <%=defaultCharset.equalsIgnoreCase(charset) ? "selected" : ""%> 
+              <%=i==numCommon ? "style=\"border-top: 1px solid #666666;\"" : "" %>><%=charset%>
             <%
+            i++;
           }
           %>

trunk/www/filemanager/files/edit_file.jsp

@@ -29 +29 @@
   import="net.sf.basedb.core.DbControl"
   import="net.sf.basedb.core.Item"
+  import="net.sf.basedb.core.Config"
   import="net.sf.basedb.core.SystemItems"
   import="net.sf.basedb.core.ItemContext"
@@ -340 +341 @@
           String currentCharset = file == null ? 
             Values.getString(cc.getPropertyValue("characterSet")) : file.getCharacterSet();
+          int numCommon = Config.getCommonCharsets().size();
+          int i = 0;
           boolean hasMatched = false;
-          for (String charset : Charset.availableCharsets().keySet())
+          for (String charset : Config.getAllCharsets())
           {
-            boolean selected = charset.equalsIgnoreCase(currentCharset);
-            hasMatched |= selected;
+            boolean selected = false;
+            if (!hasMatched)
+            {
+              selected = charset.equalsIgnoreCase(currentCharset);
+              hasMatched |= selected;
+            }
             %>
-            <option value="<%=charset%>" <%=selected ? "selected" : ""%>><%=charset%>
+            <option value="<%=charset%>" <%=selected ? "selected" : ""%> 
+              <%=i==numCommon ? "style=\"border-top: 1px solid #666666;\"" : "" %>><%=charset%>
             <%
+            i++;
           }
           if (!hasMatched && currentCharset != null)

trunk/www/filemanager/files/edit_multiple_file.jsp

@@ -29 +29 @@
   import="net.sf.basedb.core.DbControl"
   import="net.sf.basedb.core.Item"
+  import="net.sf.basedb.core.Config"
   import="net.sf.basedb.core.SystemItems"
   import="net.sf.basedb.core.ItemContext"
@@ -182 +183 @@
           <option value="">- n/a -
           <%
-          for (String charset : Charset.availableCharsets().keySet())
+          int numCommon = Config.getCommonCharsets().size();
+          int i = 0;
+          for (String charset : Config.getAllCharsets())
           {
             %>
-            <option value="<%=charset%>"><%=charset%>
+            <option value="<%=charset%>" <%=i==numCommon ? "style=\"border-top: 1px solid #666666;\"" : "" %>><%=charset%>
             <%
+            i++;
           }
           %>

trunk/www/filemanager/upload/select.jsp

@@ -30 +30 @@
   import="net.sf.basedb.core.SessionControl"
   import="net.sf.basedb.core.DbControl"
+  import="net.sf.basedb.core.Config"
   import="net.sf.basedb.core.Item"
   import="net.sf.basedb.core.SystemItems"
@@ -486 +487 @@
           <%
           String currentCharset = file == null ? null : file.getCharacterSet();
+          int numCommon = Config.getCommonCharsets().size();
+          int i = 0;
           boolean hasSelected = false;
-          for (String charset : Charset.availableCharsets().keySet())
+          for (String charset : Config.getAllCharsets())
           {
-            boolean selected = charset.equalsIgnoreCase(currentCharset);
-            hasSelected |= selected;
+            boolean selected = false;
+            if (!hasSelected)
+            {
+              selected = charset.equalsIgnoreCase(currentCharset);
+              hasSelected |= selected;
+            }
             %>
-            <option value="<%=charset%>" <%=selected ? "selected" : ""%>><%=charset%>
+            <option value="<%=charset%>" <%=selected ? "selected" : ""%> 
+              <%=i==numCommon ? "style=\"border-top: 1px solid #666666;\"" : "" %>><%=charset%>
             <%
+            i++;
           }
           if (!hasSelected && currentCharset != null)