Re: Python-list Digest, Vol 24, Issue 451

Elke Hohls <elke.hohls@xxxxxxxxx> · Fri, 30 Sep 2005 12:57:10 +0200

Sorry, the last line is wrong:
   PySDLXMLNodeType = PyMyType
...above the correction
// == PyMyExtention.c =================================================
..
:
typedef struct {
    PyObject_HEAD
    long lAttribute;
} PyMyObject;
static PyObject* PyMyObject_NewC (long lAttribute)
{
    PyMyObject *self;
    PyMyObject *type;
    self = new PyMyObject
    self->lAttribute = lAttribute;
    return (PyObject*)self;
}
static PyMethodDef PyMyObject_methods[] = {
    {"PyMyObject_NewC", (PyCFunction)PyMyObject_NewC, METH_NOARGS,
    "Create PyMyObject_NewC from C-Code"},
    {NULL}  /* Sentinel */
};
:
static PyTypeObject PyMyType = {
    PyObject_HEAD_INIT(NULL)
    :
};
//===================================================================/
python-list-request@xxxxxxxxxx wrote:

Send Python-list mailing list submissions to
	python-list@xxxxxxxxxx
To subscribe or unsubscribe via the World Wide Web, visit
	http://mail.python.org/mailman/listinfo/python-list
or, via email, send a message with subject or body 'help' to
	python-list-request@xxxxxxxxxx
You can reach the person managing the list at
	python-list-owner@xxxxxxxxxx
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Python-list digest..."

------------------------------------------------------------------------
Today's Topics:
   1. Re: Soap Question (WSDL) (Adriaan Renting)
   2. Re: Will python never intend to support private, protected
      and public? (Gregor Horvath)
   3. return (PyObject*)myPyType;  ...segmentation fault! (elho)
   4. Re: Self reordering list in Python (zooko)

------------------------------------------------------------------------
Subject:
Re: Soap Question (WSDL)
From:
"Adriaan Renting" <renting@xxxxxxxxx>
Date:
Fri, 30 Sep 2005 11:22:36 +0200
To:
<python-list@xxxxxxxxxx>
To:
<python-list@xxxxxxxxxx>

You need the WSDL file if you want external probrams to be able to discover what WebService you are running, so it depends on your need if you need to use one. You can perfectly run a SOAP service without a WSDL file, using SOAPpy, only then external programs do not have a way to find out how to talk to you.
A WSDL file just defines what messages, operations, urls etc. you accept/send/offer.
If your external applications know how to talk to you, you can do without a WSDL file.
It contains stuff like:
   <wsdl:message name="sayHelloResponse1">
      <wsdl:part name="sayHelloReturn" type="soapenc:string"/>
   </wsdl:message>
...
      <wsdl:operation name="sayHello">
         <wsdlsoap:operation soapAction=""/>
         <wsdl:input name="sayHelloRequest1">
            <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/"; namespace="urn:something.test" use="encoded"/>
         </wsdl:input>
         <wsdl:output name="sayHelloResponse1">
            <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/"; namespace="urn:something.test" use="encoded"/>
         </wsdl:output>
      </wsdl:operation>

"Armin" <akiany@xxxxxxxxx> 09/30/05 12:56 am >>> 

Hey everyone, 

I am trying to write a web app. that connects to flickr using SOAP. The 
book 'Dive into python' says I need to have a WSDL file to connect, 
while the only useful soap related url flickr api 
(flickr.com/services/api) provides is the following: 

The SOAP Server Endpoint URL is http://www.flickr.com/services/soap/ 

What am I supposed to do here? Help is very much appreciated at this 
point. 

Thanks, 
Armin 

------------------------------------------------------------------------
Subject:
Re: Will python never intend to support private, protected and public?
From:
Gregor Horvath <g.horvath@xxxxxx>
Date:
Fri, 30 Sep 2005 11:31:59 +0200
To:
python-list@xxxxxxxxxx
To:
python-list@xxxxxxxxxx

Paul Rubin wrote:
Gregor Horvath <g.horvath@xxxxxx> writes:

Someone has a problem and tweaks a private variable as a workaround.

They should have patched the source instead.

I think they are going to do that. In the meantime our friend has a 
working solution otherwise he would have nothing but broken code today.

Believe it or not, not all development environments are that
disorganized.

Martians?
Examples?
This has nothing to do with organisation but a lot with natural 
influances and constraints of software development (except really simple 
programs)
--
Greg

------------------------------------------------------------------------
Subject:
return (PyObject*)myPyType; ...segmentation fault!
From:
elho <eh1@xxxxxxxxx>
Date:
Fri, 30 Sep 2005 11:50:42 +0200
To:
python-list@xxxxxxxxxx
To:
python-list@xxxxxxxxxx

I called a own python type 'PyType' with a c function and returned it 
into my python programm - there it fault.

It is said that the object has a NULL-Pointer when I try to debug it?
Here are the importent snips from my code:

// == test.py =========================================================
.
:
myNewPyType = PyMyExtention.GetValue ("xxx")
# printings for testing
print "...back to python... test.py"
print "pp\t ...PyMyType.PyMyObject:", type(tySdlXml)
//===================================================================/

// == PyMyExtention.c =================================================
.
:
static PyObject* wrap_GetValue (PyObject* self, PyObject* args)
{
    char*      pchXXX;
    if (!PyArg_ParseTuple(args, "s", &pchXXX))
    {
        return 0;
    }
    long llong = CFunktion::CallMe(pchXXX);
    // returning Python-Objekt
    PyObject *pyType = PyMyObject_NewC (llong);
    cout << "cc ..." << ((PyMyType*)pyType)->lAttribute << endl;
    cout << "\t ...proof object-valid pointer?" << (void*)pyType << endl;
    return (PyObject*)pyType;
}
.
:
//===================================================================/

// == PyMyExtention.c =================================================
.
:
typedef struct {
    PyObject_HEAD
    long lAttribute;
} PyMyObject;
static PyObject* PyMyObject_NewC (long lAttribute)
{
    PySDLXMLNode *self;
    PySDLXMLNode *type;
    self = new PySDLXMLNode;
    self->lAttribute = lAttribute;
    return (PyObject*)self;
}
static PyMethodDef PyMyObject_methods[] = {
    {"PyMyObject_NewC", (PyCFunction)PyMyObject_NewC, METH_NOARGS,
    "Create PyMyObject_NewC from C-Code"},
    {NULL}  /* Sentinel */
};
:
static PyTypeObject PySDLXMLNodeType = {
    PyObject_HEAD_INIT(NULL)
    :
};
//===================================================================/

// :::   output  ::::::::::::::::::::::::::::::::::::::::::::::::::::
cc ...135603272
    t ...proof object-valid pointer?: 0x8165940
...back to python... test.py
Segmentation fault
//===================================================================/

...you see: It returns to python but over there the object is something 
bad. So what is wrong?

------------------------------------------------------------------------
Subject:
Re: Self reordering list in Python
From:
"zooko" <zooko@xxxxxxxxx>
Date:
30 Sep 2005 02:54:08 -0700
To:
python-list@xxxxxxxxxx
To:
python-list@xxxxxxxxxx

I've implemented such an LRU Cache in Python.  My technique was to
weave a doubly-linked list into the dict, so that it is O(dict) for all
LRU operations.  I benchmarked it against someone's Python-list-based
implementation from the ActiveState cookbook and noted that on my
machine the better constant factors of the Python list win out when the
list is cache contains fewer than about 16000 elements.  Of course,
once you exceed that cross-over point, the asymptotically worse
behavior of the list-based implementation becomes a big factor.  If you
have more than 16000 or so elements then you really oughtn't use a
list-based LRU cache.
http://zooko.com/repos/pyutil/pyutil/pyutil/cache.py
I haven't benchmarked it against Evan Podromou's heap implementation
yet, but obviously inserting and removing things from a heapq heap is
O(N).
You can find unit tests and benchmarking tools in the pyutil/test
directory.
Regards,
Zooko
P.S.  I read this list sporadically, so if you want me to read your
response, please Cc: zooko@xxxxxxxxxx  Thanks.

--
Kind regards / Mit freundlichen Grüßen
Elke Hohls
--------------------------------------------------------------------------
Elke Hohls
delair Air Traffic Systems GmbH
Lilienthalplatz 3
38108 Braunschweig
Germany
Tel: +49 (0)531 215 36-210
Fax: +49 (0)531 215 36-19
E-mail: elke.hohls@xxxxxxxxx
http://www.delair.de
.

Re: Python-list Digest, Vol 24, Issue 451

Relevant Pages