Libguestfs - May 2017 - [PATCH 2/2] python: unicode decode handler error scheme setter

The Python 3 bindings currently are unable to deal with non UTF8 characters.
This series continues what proposed in RHBZ#1406906.

A new function 'set_decode_error_handler' allows the User to specify
how to deal with decoding errors.

The default behaviour will be raising a UnicodeDecodeError.
If the handler is changed to 'surrogateescape', non UTF8 characters will
be escaped
in a similar manner as for Python 2.

See PEP383 for reference.

This series fixes also a bug introduced in commit
9d25b4e56471f9c33ea6229a8b620fc800c240f8.

Matteo Cafasso (2):
  python: return bytes when return value is RBufferOut
  python: unicode decode handler error scheme setter

 generator/python.ml            | 19 ++++++++++++++++++-
 python/handle.c                | 28 ++++++++++++++++++++++++++--
 python/t/test830RHBZ1406906.py |  6 ++++++
 3 files changed, 50 insertions(+), 3 deletions(-)

--
2.11.0

Signed-off-by: Matteo Cafasso <noxdafox@gmail.com>
---
 generator/python.ml |  3 ++-
 python/handle.c     | 10 ++++++++++
 2 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/generator/python.ml b/generator/python.ml
index 4cae24757..f7c1f80bb 100644
--- a/generator/python.ml
+++ b/generator/python.ml
@@ -93,6 +93,7 @@ extern PyObject *guestfs_int_py_put_string_list (char * const
* const argv);
 extern PyObject *guestfs_int_py_put_table (char * const * const argv);
 extern PyObject *guestfs_int_py_fromstring (const char *str);
 extern PyObject *guestfs_int_py_fromstringsize (const char *str, size_t size);
+extern PyObject *guestfs_int_py_bytesfromstringsize (const char *str, size_t
size);
 extern char *guestfs_int_py_asstring (PyObject *obj);

 ";
@@ -513,7 +514,7 @@ and generate_python_actions actions ()             pr "
guestfs_int_free_string_list (r);\n";
            pr "  if (py_r == NULL) goto out;\n";
        | RBufferOut _ ->
-           pr "  py_r = guestfs_int_py_fromstringsize (r, size);\n";
+           pr "  py_r = guestfs_int_py_bytesfromstringsize (r,
size);\n";
            pr "  free (r);\n";
            pr "  if (py_r == NULL) goto out;\n";
       );
diff --git a/python/handle.c b/python/handle.c
index d93f2f021..52c36f1d2 100644
--- a/python/handle.c
+++ b/python/handle.c
@@ -400,6 +400,16 @@ guestfs_int_py_fromstringsize (const char *str, size_t
size)
 #endif
 }

+PyObject *
+guestfs_int_py_bytesfromstringsize (const char *str, size_t size)
+{
+#ifdef HAVE_PYSTRING_ASSTRING
+  return PyString_FromStringAndSize (str, size);
+#else
+  return PyBytes_FromStringAndSize (str, size);
+#endif
+}
+
 char *
 guestfs_int_py_asstring (PyObject *obj)
 {
--
2.11.0

The set_decode_error_handler function allows the User to set the
decoding error scheme to be used when non UTF8 characters are
encountered in Python 3.

The function has no effect in Python 2.

Signed-off-by: Matteo Cafasso <noxdafox@gmail.com>
---
 generator/python.ml            | 16 ++++++++++++++++
 python/handle.c                | 18 ++++++++++++++++--
 python/t/test830RHBZ1406906.py |  6 ++++++
 3 files changed, 38 insertions(+), 2 deletions(-)

diff --git a/generator/python.ml b/generator/python.ml
index f7c1f80bb..66bb7f27d 100644
--- a/generator/python.ml
+++ b/generator/python.ml
@@ -82,6 +82,7 @@ put_handle (guestfs_h *g)
 }

 extern void guestfs_int_py_extend_module (PyObject *module);
+extern PyObject *guestfs_int_py_set_decode_error_handler (PyObject *self,
PyObject *args);

 extern PyObject *guestfs_int_py_create (PyObject *self, PyObject *args);
 extern PyObject *guestfs_int_py_close (PyObject *self, PyObject *args);
@@ -577,6 +578,8 @@ and generate_python_module () 
   (* Table of functions. *)
   pr "static PyMethodDef methods[] = {\n";
+  pr "  { (char *) \"set_decode_error_handler\", \n";
+  pr "    guestfs_int_py_set_decode_error_handler, METH_VARARGS, NULL
},\n";
   pr "  { (char *) \"create\", guestfs_int_py_create,
METH_VARARGS, NULL },\n";
   pr "  { (char *) \"close\", guestfs_int_py_close,
METH_VARARGS, NULL },\n";
   pr "  { (char *) \"set_event_callback\",\n";
@@ -728,6 +731,19 @@ class ClosedHandle(ValueError):
     pass


+def set_decode_error_handler(handler):
+    \"\"\"Set the error handling scheme to use for the handling
+    of decoding errors.
+    The default is 'strict' meaning that decoding errors raise a
+    UnicodeDecodeError.
+
+    The other possible value is 'surrogateescape', see PEP383 for
reference.
+
+    Return the previous error handler.
+    \"\"\"
+    return libguestfsmod.set_decode_error_handler(handler)
+
+
 class GuestFS(object):
     \"\"\"Instances of this class are libguestfs API
handles.\"\"\"

diff --git a/python/handle.c b/python/handle.c
index 52c36f1d2..b665bb899 100644
--- a/python/handle.c
+++ b/python/handle.c
@@ -35,6 +35,8 @@

 #include "actions.h"

+static const char *decode_error_handler = "strict";
+
 static PyObject **get_all_event_callbacks (guestfs_h *g, size_t *len_rtn);

 void
@@ -45,6 +47,17 @@ guestfs_int_py_extend_module (PyObject *module)
 }

 PyObject *
+guestfs_int_py_set_decode_error_handler (PyObject *self, PyObject *args)
+{
+  const char *previous_handler = decode_error_handler;
+
+  if (!PyArg_ParseTuple (args, (char *) "s:set_decode_error_handler",
&decode_error_handler))
+    return NULL;
+
+  return guestfs_int_py_fromstring (previous_handler);
+}
+
+PyObject *
 guestfs_int_py_create (PyObject *self, PyObject *args)
 {
   guestfs_h *g;
@@ -386,7 +399,8 @@ guestfs_int_py_fromstring (const char *str)
 #ifdef HAVE_PYSTRING_ASSTRING
   return PyString_FromString (str);
 #else
-  return PyUnicode_FromString (str);
+  Py_ssize_t size = strlen(str);
+  return PyUnicode_DecodeUTF8 (str, size, decode_error_handler);
 #endif
 }

@@ -396,7 +410,7 @@ guestfs_int_py_fromstringsize (const char *str, size_t size)
 #ifdef HAVE_PYSTRING_ASSTRING
   return PyString_FromStringAndSize (str, size);
 #else
-  return PyUnicode_FromStringAndSize (str, size);
+  return PyUnicode_DecodeUTF8 (str, size, decode_error_handler);
 #endif
 }

diff --git a/python/t/test830RHBZ1406906.py b/python/t/test830RHBZ1406906.py
index 17b875226..0bb1ac1d0 100644
--- a/python/t/test830RHBZ1406906.py
+++ b/python/t/test830RHBZ1406906.py
@@ -55,3 +55,9 @@ class Test830RHBZ1406906(unittest.TestCase):
         elif sys.version_info >= (2, 0):
             self.assertTrue(
                 any(path for path in g.find("/") if non_utf8_fname in
path))
+
+        # change decoding error handler
+        self.assertEqual(
+            guestfs.set_decode_error_handler("surrogateescape"),
'strict')
+        self.assertTrue(
+            any(path for path in g.find("/") if non_utf8_fname in
path))
--
2.11.0

On Sunday, 21 May 2017 18:29:03 CEST Matteo Cafasso
wrote:
> The set_decode_error_handler function allows the User to set the
> decoding error scheme to be used when non UTF8 characters are
> encountered in Python 3.
s/User/user/, and s/UTF8/UTF-8/
> The function has no effect in Python 2.
> 
> Signed-off-by: Matteo Cafasso <noxdafox@gmail.com>
> ---
>  generator/python.ml            | 16 ++++++++++++++++
>  python/handle.c                | 18 ++++++++++++++++--
>  python/t/test830RHBZ1406906.py |  6 ++++++
>  3 files changed, 38 insertions(+), 2 deletions(-)
> 
> diff --git a/generator/python.ml b/generator/python.ml
> index f7c1f80bb..66bb7f27d 100644
> --- a/generator/python.ml
> +++ b/generator/python.ml
> @@ -82,6 +82,7 @@ put_handle (guestfs_h *g)
>  }
> 
>  extern void guestfs_int_py_extend_module (PyObject *module);
> +extern PyObject *guestfs_int_py_set_decode_error_handler (PyObject *self,
PyObject *args);
> 
>  extern PyObject *guestfs_int_py_create (PyObject *self, PyObject *args);
>  extern PyObject *guestfs_int_py_close (PyObject *self, PyObject *args);
> @@ -577,6 +578,8 @@ and generate_python_module () > 
>    (* Table of functions. *)
>    pr "static PyMethodDef methods[] = {\n";
> +  pr "  { (char *) \"set_decode_error_handler\", \n";
> +  pr "    guestfs_int_py_set_decode_error_handler, METH_VARARGS, NULL
},\n";
This is implemented as global for the whole module, which means
changing the behaviour for an handle changes it for all the existing
handles (and in a racy behaviour, even).  This IMHO should be a
per-handle setting.
>    pr "  { (char *) \"create\", guestfs_int_py_create,
METH_VARARGS, NULL },\n";
>    pr "  { (char *) \"close\", guestfs_int_py_close,
METH_VARARGS, NULL },\n";
>    pr "  { (char *) \"set_event_callback\",\n";
> @@ -728,6 +731,19 @@ class ClosedHandle(ValueError):
>      pass
> 
> 
> +def set_decode_error_handler(handler):
'handler' usually is a function/callback, while in this case is a
behaviour/mode, so I'd use a different naming.
> +    \"\"\"Set the error handling scheme to use for the
handling
> +    of decoding errors.
> +    The default is 'strict' meaning that decoding errors raise a
> +    UnicodeDecodeError.
> +
> +    The other possible value is 'surrogateescape', see PEP383 for
reference.
> +
> +    Return the previous error handler.
> +    \"\"\"
> +    return libguestfsmod.set_decode_error_handler(handler)
> +
> +
>  class GuestFS(object):
>      \"\"\"Instances of this class are libguestfs API
handles.\"\"\"
> 
> diff --git a/python/handle.c b/python/handle.c
> index 52c36f1d2..b665bb899 100644
> --- a/python/handle.c
> +++ b/python/handle.c
> @@ -35,6 +35,8 @@
> 
>  #include "actions.h"
> 
> +static const char *decode_error_handler = "strict";
> +
>  static PyObject **get_all_event_callbacks (guestfs_h *g, size_t *len_rtn);
> 
>  void
> @@ -45,6 +47,17 @@ guestfs_int_py_extend_module (PyObject *module)
>  }
> 
>  PyObject *
> +guestfs_int_py_set_decode_error_handler (PyObject *self, PyObject *args)
> +{
> +  const char *previous_handler = decode_error_handler;
> +
> +  if (!PyArg_ParseTuple (args, (char *)
"s:set_decode_error_handler", &decode_error_handler))
> +    return NULL;
I really doubt "decode_error_handler" will hold a valid pointer after
guestfs_int_py_set_decode_error_handler is done (and the args PyObject
is disposed.

-- 
Pino Toscano