setup.py and file extensions like ".c++"

[garyjefferson123]

Is there any way to get setup.py to recognize file extensions like .c++
in lieu of .cpp?  I'd love to not have to rename the source files for
the library I'm trying to wrap in a python C extension.

I get:

error: unknown file type '.c++' (from 'parser.c++')

when I type 'python setup.py build'

thanks,
Gary

-- 
http://mail.python.org/mailman/listinfo/python-list

accepting cStringIO in an extension

[garyjefferson123]

I want to accept a cStringIO object in a function in a python extension
module.  How do I do this?

e.g.,

static PyObject *myfunc(PyObject *self, PyObject *args)
{
PyObject *cstringio;
if (!PyArg_ParseTuple(args, "O:cStringIO", &cstringio)) {
PyErr_SetString(PyExc_ValueError,
"value must be a cstringio");
return NULL;
}
/* how to get at the data buffer or call read() on cstringio? */

}

I understand that I probably want PyEval_CallObject(), but I don't know
how to get at the read() method of cstringio.

Thanks,
Gary

-- 
http://mail.python.org/mailman/listinfo/python-list

customizing a logging logger

[garyjefferson123]

Suppose I have some sort of context variable that I want to appear in
log messages.  E.g.:

logger = logging.getLogger("somelogger")
class SomeOp:
def __init__(self, ctx):
self.ctx = ctx
def method1(self):
logger.info("%s: here's a message", self.ctx)

What's the idiomatic way to abstract away the inclusion of self.ctx
from the calls in logger.info() et al?  Is there some way I can
declare a

@property
def info(self): return logger.info

but have it insert the '"%s: " % self.ctx' bit for me in one place
instead of the dozens of places I currently do it in the class?

Thanks,
Gary

-- 
http://mail.python.org/mailman/listinfo/python-list

Re: customizing a logging logger

[garyjefferson123]

I think the following question is clearer.

I want to make it so that method1 below can be transformed:

logger = logging.getLogger("somelogger")
class SomeOp:
def __init__(self, ctx):
self.ctx = ctx
def method1(self):
logger.info("%s: here's a message", self.ctx)
logger.debug("%s: msg #2", self.ctx)
v = 'yessir'
logger.debug("%s: msg #3 %s", self.ctx, v)

into something like:

def method1(self):
logger.info("here's a message")
logger.debug("msg #2")
v = 'yessir'
logger.debug("msg #3 %s", v)


What is the best way to do this, so that I don't have to manually put
the self.ctx in the log string in hundreds of different places?

Thanks,
Gary

-- 
http://mail.python.org/mailman/listinfo/python-list

Re: customizing a logging logger

[garyjefferson123]

On Sep 12, 10:46 am, Vinay Sajip <[EMAIL PROTECTED]> wrote:
> On 11 Sep, 17:14, [EMAIL PROTECTED] wrote:
>
> > What is the best way to do this, so that I don't have to manually put
> > the self.ctx in the log string in hundreds of different places?
>
> Another way is to generate your log messages via a factory method
> which prepends the context: you can do this via a mixin, too.
>
> def log_message(self, msg):
> return "%s: %s" % (self.ctx, msg)
>
> Note also that in recent versions, an optional "extra" parameter can
> be used, to pass info into the LogRecord.
>
> Seehttp://mail.python.org/pipermail/patches/2007-January/021535.html
> which may be of some help - it's a similar use case.
>
> Best regards,
>
> Vinay Sajip


The email thread you pointed to was very informative, especially since
I was about to go down the road of creating one logger per context
(eek!).

Whichever method I choose, I am particularly concerned with the
following aspect:

I like to use the "logger.debug("msg %s %s", s1, s2)" format, as I
understand that when the logging for this log level is turned off, we
don't have to do the potentially expensive string building that a '%'
operator would imply.

It seems like using something other than a literal string in the
message is the way to go, since I assume that its __repr__() won't get
called unless the logger is actually going to log a message for it.
Is that right?

Are any of the other methods likely to provide as-good or better
performance?

thanks,
Gary

-- 
http://mail.python.org/mailman/listinfo/python-list

Accessing overridden __builtin__s?

[garyjefferson123]

I'm having a scoping problem.  I have a module called SpecialFile,
which defines:

def open(fname, mode):
  return SpecialFile(fname, mode)

class SpecialFile:

  def __init__(self, fname, mode):
self.f = open(fname, mode)
...


The problem, if it isn't obvioius, is that the open() call in __init__
no longer refers to the builtin open(), but to the module open().  So,
if I do:

f = SpecialFile.open(name, mode)

I get infinite recursion.

How do I tell my class that I want to refer to the __builtin__ open(),
and not the one defined in the module?

Thanks,
Gary

-- 
http://mail.python.org/mailman/listinfo/python-list