Now String is mapped to PyUnicode on ClassicPyObjectAdapter (instead of mapping it to PyString). Thus, for APIs which really need to return bytestrings (such as PyFile) we now have to explicitely declare the return type as PyString (and on code written in Python, use StringUtil.asPyString(String)). Fixes #1128

leosoto · leosoto · commit 71f8ff257118 · 2008-10-14T03:20:10.000Z
diff --git a/Lib/javapath.py b/Lib/javapath.py
@@ -20,6 +20,8 @@
 from java.lang import System
 import os
 
+from org.python.core.util.StringUtil import asPyString
+
 
 def _tostr(s, method):
     if isinstance(s, basestring):
@@ -40,7 +42,7 @@ def _type_name(obj):
 def dirname(path):
     """Return the directory component of a pathname"""
     path = _tostr(path, "dirname")
-    result = File(path).getParent()
+    result = asPyString(File(path).getParent())
     if not result:
         if isabs(path):
             result = path # Must be root
@@ -51,7 +53,7 @@ def dirname(path):
 def basename(path):
     """Return the final component of a pathname"""
     path = _tostr(path, "basename")
-    return File(path).getName()
+    return asPyString(File(path).getName())
 
 def split(path):
     """Split a pathname.
@@ -128,7 +130,7 @@ def join(path, *args):
             if a == "":
                 a = os.sep
             f = File(f, a)
-    return f.getPath()
+    return asPyString(f.getPath())
 
 def normcase(path):
     """Normalize case of pathname.
@@ -137,7 +139,7 @@ def normcase(path):
 
     """
     path = _tostr(path, "normcase")
-    return File(path).getPath()
+    return asPyString(File(path).getPath())
 
 def commonprefix(m):
     "Given a list of pathnames, return the longest common leading component"
@@ -197,7 +199,7 @@ def expanduser(path):
         if not c:
             return gethome()
         if c == os.sep:
-            return File(gethome(), path[2:]).getPath()
+            return asPyString(File(gethome(), path[2:]).getPath())
     return path
 
 def getuser():
@@ -252,7 +254,7 @@ def abspath(path):
 def _abspath(path):
     # Must use normpath separately because getAbsolutePath doesn't normalize
     # and getCanonicalPath would eliminate symlinks.
-    return normpath(File(sys.getPath(path)).getAbsolutePath())
+    return normpath(asPyString(File(sys.getPath(path)).getAbsolutePath()))
 
 def realpath(path):
     """Return an absolute path normalized and symbolic links eliminated"""
@@ -261,7 +263,7 @@ def realpath(path):
     
 def _realpath(path):
     try:
-        return File(sys.getPath(path)).getCanonicalPath()
+        return asPyString(File(sys.getPath(path)).getCanonicalPath())
     except java.io.IOException:
         return _abspath(path)
 
diff --git a/Lib/os.py b/Lib/os.py
@@ -48,6 +48,7 @@
 from java.io import File
 from org.python.core import PyFile
 from org.python.core.io import FileDescriptors, FileIO, IOBase
+from org.python.core.util.StringUtil import asPyString
 
 # Mapping of: os._name: [name list, shell command list]
 _os_map = dict(nt=[
@@ -264,7 +265,7 @@ def listdir(path):
     l = File(sys.getPath(path)).list()
     if l is None:
         raise OSError(0, 'No such directory', path)
-    return list(l)
+    return [asPyString(entry) for entry in l]
 
 def chmod(path, mode):
     """chmod(path, mode)
@@ -629,7 +630,7 @@ def read(fd, buffersize):
     from org.python.core.util import StringUtil
     rawio = FileDescriptors.get(fd)
     buf = _handle_oserror(rawio.read, buffersize)
-    return str(StringUtil.fromBytes(buf))
+    return asPyString(StringUtil.fromBytes(buf))
 
 def write(fd, string):
     """write(fd, string) -> byteswritten
diff --git a/Lib/zlib.py b/Lib/zlib.py
@@ -1,7 +1,9 @@
 import jarray, binascii
 
 from java.util.zip import Adler32, Deflater, Inflater
-from java.lang import Long, String, StringBuffer
+from java.lang import Long, String
+
+from cStringIO import StringIO
 
 class error(Exception):
     pass
@@ -132,18 +134,19 @@ def flush(self, length=None):
 
 def _get_deflate_data(deflater):
     buf = jarray.zeros(1024, 'b')
-    sb = StringBuffer()
+    s = StringIO()
     while not deflater.finished():
         l = deflater.deflate(buf)
         if l == 0:
             break
-        sb.append(String(buf, 0, 0, l))
-    return sb.toString()
+        s.write(String(buf, 0, 0, l))
+    s.seek(0)
+    return s.read()
 
         
 def _get_inflate_data(inflater, max_length=0):
     buf = jarray.zeros(1024, 'b')
-    sb = StringBuffer()
+    s = StringIO()
     total = 0
     while not inflater.finished():
         if max_length:
@@ -154,7 +157,8 @@ def _get_inflate_data(inflater, max_length=0):
             break
 
         total += l
-        sb.append(String(buf, 0, 0, l))
+        s.write(String(buf, 0, 0, l))
         if max_length and total == max_length:
             break
-    return sb.toString()
+    s.seek(0)
+    return s.read()
diff --git a/src/com/ziclix/python/sql/DataHandler.java b/src/com/ziclix/python/sql/DataHandler.java
@@ -177,7 +177,7 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in
 
                 case Types.LONGVARCHAR:
                     if (object instanceof PyFile) {
-                        object = new PyString(((PyFile) object).read());
+                        object = ((PyFile) object).read();
                     }
 
                     String varchar = (String) object.__tojava__(String.class);
@@ -192,7 +192,7 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in
 
                 default :
                     if (object instanceof PyFile) {
-                        object = new PyString(((PyFile) object).read());
+                        object = ((PyFile) object).read();
                     }
 
                     stmt.setObject(index, object.__tojava__(Object.class), type);
diff --git a/src/com/ziclix/python/sql/JDBC20DataHandler.java b/src/com/ziclix/python/sql/JDBC20DataHandler.java
@@ -62,7 +62,7 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in
 
             case Types.CLOB:
                 if (object instanceof PyFile) {
-                    object = new PyString(((PyFile) object).read());
+                    object = ((PyFile) object).read();
                 }
 
                 String clob = (String) object.__tojava__(String.class);
diff --git a/src/com/ziclix/python/sql/Jython22DataHandler.java b/src/com/ziclix/python/sql/Jython22DataHandler.java
@@ -163,7 +163,7 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in
 
                 case Types.LONGVARCHAR:
                     if (object instanceof PyFile) {
-                        object = new PyString(((PyFile) object).read());
+                        object = ((PyFile) object).read();
                     }
 
                     String varchar = (String) object.__tojava__(String.class);
@@ -178,7 +178,7 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in
 
                 default :
                     if (object instanceof PyFile) {
-                        object = new PyString(((PyFile) object).read());
+                        object = ((PyFile) object).read();
                     }
 
                     stmt.setObject(index, object.__tojava__(Object.class), type);
diff --git a/src/com/ziclix/python/sql/handler/MySQLDataHandler.java b/src/com/ziclix/python/sql/handler/MySQLDataHandler.java
@@ -54,17 +54,19 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in
     switch (type) {
 
       case Types.LONGVARCHAR:
-        String varchar;
+          // XXX: Only works with ASCII data!
+          byte[] bytes;
         if (object instanceof PyFile) {
-          varchar = ((PyFile) object).read();
+              bytes = ((PyFile) object).read().toBytes();
         } else {
-          varchar = (String) object.__tojava__(String.class);
+              String varchar = (String) object.__tojava__(String.class);
+              bytes = StringUtil.toBytes(varchar);
         }
-        InputStream stream = new ByteArrayInputStream(StringUtil.toBytes(varchar));
+          InputStream stream = new ByteArrayInputStream(bytes);
 
         stream = new BufferedInputStream(stream);
 
-        stmt.setAsciiStream(index, stream, varchar.length());
+          stmt.setAsciiStream(index, stream, bytes.length);
         break;
 
       default :
diff --git a/src/com/ziclix/python/sql/handler/PostgresqlDataHandler.java b/src/com/ziclix/python/sql/handler/PostgresqlDataHandler.java
@@ -111,7 +111,7 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in
         String varchar;
         // Postgresql driver can't handle the setCharacterStream() method so use setObject() instead
         if (object instanceof PyFile) {
-          varchar = ((PyFile) object).read();
+          varchar = ((PyFile) object).read().asString();
         } else {
           varchar = (String) object.__tojava__(String.class);
         }
diff --git a/src/org/python/core/PyFile.java b/src/org/python/core/PyFile.java
@@ -286,16 +286,16 @@ private String parseMode(String mode) {
     }
 
     @ExposedMethod(defaults = {"-1"})
-    final synchronized String file_read(int n) {
+    final synchronized PyString file_read(int n) {
         checkClosed();
-        return file.read(n);
+        return new PyString(file.read(n));
     }
 
-    public String read(int n) {
+    public PyString read(int n) {
         return file_read(n);
     }
 
-    public String read() {
+    public PyString read() {
         return file_read(-1);
     }
 
@@ -310,16 +310,16 @@ public int readinto(PyObject buf) {
     }
 
     @ExposedMethod(defaults = {"-1"})
-    final synchronized String file_readline(int max) {
+    final synchronized PyString file_readline(int max) {
         checkClosed();
-        return file.readline(max);
+        return new PyString(file.readline(max));
     }
 
-    public String readline(int max) {
+    public PyString readline(int max) {
         return file_readline(max);
     }
 
-    public String readline() {
+    public PyString readline() {
         return file_readline(-1);
     }
 
diff --git a/src/org/python/core/PyFunction.java b/src/org/python/core/PyFunction.java
@@ -143,13 +143,13 @@ static final PyObject function___new__(PyNewWrapper new_, boolean init, PyType s
     }
 
     @ExposedGet(name = "func_name")
-    public String getFuncName() {
-        return __name__;
+    public PyString getFuncName() {
+        return new PyString(__name__);
     }
 
     @ExposedSet(name = "func_name")
-    public void setFuncName(String func_name) {
-        __name__ = func_name;
+    public void setFuncName(PyString func_name) {
+        __name__ = func_name.asString();
     }
 
     @ExposedGet(name = "func_doc")
diff --git a/src/org/python/core/PyStringMap.java b/src/org/python/core/PyStringMap.java
@@ -141,7 +141,9 @@ public String toString() {
         for (Entry<Object, PyObject> entry : table.entrySet()) {
             Object key = entry.getKey();
             if (key instanceof String) {
-                buf.append(Py.java2py(key).__repr__().toString());
+                // This is a bit complicated, but prevents us to duplicate
+                // PyString#__repr__ logic here.
+                buf.append(new PyString((String)key).__repr__().toString());
             } else {
                 buf.append(((PyObject)key).__repr__().toString());
             }
diff --git a/src/org/python/core/PySystemState.java b/src/org/python/core/PySystemState.java
@@ -41,7 +41,7 @@ public class PySystemState extends PyObject
     private static final String JAR_URL_PREFIX = "jar:file:";
     private static final String JAR_SEPARATOR = "!";
 
-    public static String version = Version.getVersion();
+    public static PyString version = new PyString(Version.getVersion());
     public static int hexversion = ((Version.PY_MAJOR_VERSION << 24) |
                                     (Version.PY_MINOR_VERSION << 16) |
                                     (Version.PY_MICRO_VERSION <<  8) |
@@ -101,8 +101,8 @@ public class PySystemState extends PyObject
     public PyList path_hooks;
     public PyObject path_importer_cache;
 
-    public static String platform = "java";
-    public static String byteorder = "big";
+    public static PyString platform = new PyString("java");
+    public static PyString byteorder = new PyString("big");
 
     public PyObject ps1 = new PyString(">>> ");
     public PyObject ps2 = new PyString("... ");
@@ -288,8 +288,8 @@ public void setprofile(PyObject profilefunc) {
         }
     }
 
-    public String getdefaultencoding() {
-        return codecs.getDefaultEncoding();
+    public PyString getdefaultencoding() {
+        return new PyString(codecs.getDefaultEncoding());
     }
 
     public void setdefaultencoding(String encoding) {
@@ -431,7 +431,7 @@ public static void determinePlatform(Properties props) {
         if (version.equals("12")) {
             version = "1.2";
         }
-        platform = "java" + version;
+        platform = new PyString("java" + version);
     }
 
     private static void initRegistry(Properties preProperties, Properties postProperties, 
diff --git a/src/org/python/core/PyTraceback.java b/src/org/python/core/PyTraceback.java
@@ -62,11 +62,12 @@ private String getLine(String filename, int lineno) {
             return null;
         }
 
+
         String line = null;
         int i = 0;
         try {
             for (i = 0; i < tb_lineno; i++) {
-                line = pyFile.readline();
+                line = pyFile.readline().asString();
                 if (line.equals("")) {
                     break;
                 }
diff --git a/src/org/python/core/__builtin__.java b/src/org/python/core/__builtin__.java
@@ -1028,7 +1028,7 @@ public static PyObject range(PyObject start, PyObject stop) {
 
     private static PyString readline(PyObject file) {
         if (file instanceof PyFile) {
-            return new PyString(((PyFile) file).readline());
+            return ((PyFile) file).readline();
         } else {
             PyObject ret = file.invoke("readline");
             if (!(ret instanceof PyString)) {
diff --git a/src/org/python/core/adapter/ClassicPyObjectAdapter.java b/src/org/python/core/adapter/ClassicPyObjectAdapter.java
@@ -9,8 +9,8 @@
 import org.python.core.PyLong;
 import org.python.core.PyObject;
 import org.python.core.PyProxy;
-import org.python.core.PyString;
 import org.python.core.PyType;
+import org.python.core.PyUnicode;
 
 /**
  * Implements the algorithm originally used in {@link Py#java2py} to adapt objects.
@@ -60,7 +60,7 @@ public PyObject adapt(Object o) {
 		add(new ClassAdapter(String.class) {
 
 			public PyObject adapt(Object o) {
-				return new PyString((String) o);
+                return new PyUnicode((String) o);
 			}
 
 		});
diff --git a/src/org/python/core/util/StringUtil.java b/src/org/python/core/util/StringUtil.java
@@ -5,6 +5,7 @@
 import java.nio.ByteBuffer;
 
 import org.python.core.Py;
+import org.python.core.PyString;
 
 /**
  * String Utility methods.
@@ -83,4 +84,8 @@ public static String decapitalize(String string) {
         chars[0] = Character.toLowerCase(c0);
         return new String(chars);
     }
+
+    public static PyString asPyString(String string) {
+        return new PyString(string);
+}
 }
diff --git a/src/org/python/modules/binascii.java b/src/org/python/modules/binascii.java
diff --git a/src/org/python/modules/cPickle.java b/src/org/python/modules/cPickle.java
diff --git a/src/org/python/modules/cStringIO.java b/src/org/python/modules/cStringIO.java
diff --git a/src/org/python/modules/struct.java b/src/org/python/modules/struct.java
diff --git a/src/org/python/modules/time/Time.java b/src/org/python/modules/time/Time.java

Original file line number	Diff line number	Diff line change
`@@ -177,7 +177,7 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in`
`177`	`177`
`178`	`178`	`case Types.LONGVARCHAR:`
`179`	`179`	`if (object instanceof PyFile) {`
`180`		`- object = new PyString(((PyFile) object).read());`
	`180`	`+ object = ((PyFile) object).read();`
`181`	`181`	`}`
`182`	`182`
`183`	`183`	`String varchar = (String) object.__tojava__(String.class);`
`@@ -192,7 +192,7 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in`
`192`	`192`
`193`	`193`	`default :`
`194`	`194`	`if (object instanceof PyFile) {`
`195`		`- object = new PyString(((PyFile) object).read());`
	`195`	`+ object = ((PyFile) object).read();`
`196`	`196`	`}`
`197`	`197`
`198`	`198`	`stmt.setObject(index, object.__tojava__(Object.class), type);`
Original file line number	Diff line number	Diff line change
`@@ -62,7 +62,7 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in`
`62`	`62`
`63`	`63`	`case Types.CLOB:`
`64`	`64`	`if (object instanceof PyFile) {`
`65`		`- object = new PyString(((PyFile) object).read());`
	`65`	`+ object = ((PyFile) object).read();`
`66`	`66`	`}`
`67`	`67`
`68`	`68`	`String clob = (String) object.__tojava__(String.class);`
Original file line number	Diff line number	Diff line change
`@@ -163,7 +163,7 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in`
`163`	`163`
`164`	`164`	`case Types.LONGVARCHAR:`
`165`	`165`	`if (object instanceof PyFile) {`
`166`		`- object = new PyString(((PyFile) object).read());`
	`166`	`+ object = ((PyFile) object).read();`
`167`	`167`	`}`
`168`	`168`
`169`	`169`	`String varchar = (String) object.__tojava__(String.class);`
`@@ -178,7 +178,7 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in`
`178`	`178`
`179`	`179`	`default :`
`180`	`180`	`if (object instanceof PyFile) {`
`181`		`- object = new PyString(((PyFile) object).read());`
	`181`	`+ object = ((PyFile) object).read();`
`182`	`182`	`}`
`183`	`183`
`184`	`184`	`stmt.setObject(index, object.__tojava__(Object.class), type);`
Original file line number	Diff line number	Diff line change
`@@ -111,7 +111,7 @@ public void setJDBCObject(PreparedStatement stmt, int index, PyObject object, in`
`111`	`111`	`String varchar;`
`112`	`112`	`// Postgresql driver can't handle the setCharacterStream() method so use setObject() instead`
`113`	`113`	`if (object instanceof PyFile) {`
`114`		`- varchar = ((PyFile) object).read();`
	`114`	`+ varchar = ((PyFile) object).read().asString();`
`115`	`115`	`} else {`
`116`	`116`	`varchar = (String) object.__tojava__(String.class);`
`117`	`117`	`}`
Original file line number	Diff line number	Diff line change
`@@ -286,16 +286,16 @@ private String parseMode(String mode) {`
`286`	`286`	`}`
`287`	`287`
`288`	`288`	`@ExposedMethod(defaults = {"-1"})`
`289`		`- final synchronized String file_read(int n) {`
	`289`	`+ final synchronized PyString file_read(int n) {`
`290`	`290`	`checkClosed();`
`291`		`- return file.read(n);`
	`291`	`+ return new PyString(file.read(n));`
`292`	`292`	`}`
`293`	`293`
`294`		`- public String read(int n) {`
	`294`	`+ public PyString read(int n) {`
`295`	`295`	`return file_read(n);`
`296`	`296`	`}`
`297`	`297`
`298`		`- public String read() {`
	`298`	`+ public PyString read() {`
`299`	`299`	`return file_read(-1);`
`300`	`300`	`}`
`301`	`301`
`@@ -310,16 +310,16 @@ public int readinto(PyObject buf) {`
`310`	`310`	`}`
`311`	`311`
`312`	`312`	`@ExposedMethod(defaults = {"-1"})`
`313`		`- final synchronized String file_readline(int max) {`
	`313`	`+ final synchronized PyString file_readline(int max) {`
`314`	`314`	`checkClosed();`
`315`		`- return file.readline(max);`
	`315`	`+ return new PyString(file.readline(max));`
`316`	`316`	`}`
`317`	`317`
`318`		`- public String readline(int max) {`
	`318`	`+ public PyString readline(int max) {`
`319`	`319`	`return file_readline(max);`
`320`	`320`	`}`
`321`	`321`
`322`		`- public String readline() {`
	`322`	`+ public PyString readline() {`
`323`	`323`	`return file_readline(-1);`
`324`	`324`	`}`
`325`	`325`
Original file line number	Diff line number	Diff line change
`@@ -143,13 +143,13 @@ static final PyObject function___new__(PyNewWrapper new_, boolean init, PyType s`
`143`	`143`	`}`
`144`	`144`
`145`	`145`	`@ExposedGet(name = "func_name")`
`146`		`- public String getFuncName() {`
`147`		`- return __name__;`
	`146`	`+ public PyString getFuncName() {`
	`147`	`+ return new PyString(__name__);`
`148`	`148`	`}`
`149`	`149`
`150`	`150`	`@ExposedSet(name = "func_name")`
`151`		`- public void setFuncName(String func_name) {`
`152`		`- __name__ = func_name;`
	`151`	`+ public void setFuncName(PyString func_name) {`
	`152`	`+ __name__ = func_name.asString();`
`153`	`153`	`}`
`154`	`154`
`155`	`155`	`@ExposedGet(name = "func_doc")`