Automatic debugging of copy by reference errors?

[Niels L Ellegaard]

Is there a module that allows me to find errors that occur due to copy
by reference? I am looking for something like the following:

>>import mydebug
mydebug.check copybyreference = True
a=2
b=[a]
a=4

Traceback (most recent call last):
File "<stdin>", line 1, in ?
CopyByReference Error: Variable b refers to variable a, so please do not
change variable a.

Does such a module exist?
Would it be possible to code such a module?
Would it be possible to add the functionality to array-copying in
numpy?
What would be the extra cost in terms of memory and CPU power?

I realize that this module would disable some useful features of the
language. On the other hand it could be helpful for new python users.

Niels

[Gabriel Genellina]

On 9 dic, 02:22, "Niels L Ellegaard" <niels.ellega.. [e-mail] e:

Is there a module that allows me to find errors that occur due to copy
by reference?

What do you mean by "copy by reference"?

I am looking for something like the following:

>import mydebug
mydebug.checkc opybyreference = True
a=2
b=[a]
a=4

Traceback (most recent call last):
File "<stdin>", line 1, in ?
CopyByReference Error: Variable b refers to variable a, so please do not
change variable a.

(I won't be pedantic to say that Python has no "variables" ). What's
wrong with that code? You are *not* changing "variable a", you are
binding the integer 4 to the name "a". That name used previously to be
bound to another integer, 2 - what's wrong with it? Anyway it has no
effect on the list referred by the name "b", still holds b==[2]
What do you want? Forbid the re-use of names? So once you say a=2, you
can't modify it? That could be done, yes, a "write-once-read-many"
namespace. But I don't see the usefullness.

Does such a module exist?

No

Would it be possible to code such a module?

I don't know what do you want to do exactly, but I feel it's not
useful.

Would it be possible to add the functionality to array-copying in
numpy?

No idea.

What would be the extra cost in terms of memory and CPU power?

No idea yet.

I realize that this module would disable some useful features of the
language.

Not only "some useful features", most programs would not even work!

On the other hand it could be helpful for new python users.

I think you got in trouble with something and you're trying to avoid it
again - but perhaps this is not the right way. Could you provide some
example?

--
Gabriel Genellina

[Niels L Ellegaard]

Gabriel Genellina wrote:

I think you got in trouble with something and you're trying to avoid it
again - but perhaps this is not the right way. Could you provide some
example?

I have been using scipy for some time now, but in the beginning I made
a few mistakes with copying by reference. The following example is
exagerated
for clarity, but the principle is the same:

import os
output=[]
firstlines =[0,0]
for filename in os.listdir('.') :
try:
firstlines[0] = open(filename," r").readline s()[0]
firstlines[1] = open(filename," r").readline s()[1]
output.append(( filename,firstl ines))
except:continue
print output

Now some of my fortran-using friends would like to use python to
analyze their data files. I wanted them to avoid making the same
mistakes as I did so I thought it would be good if they could get some
nanny-like warnings saying: "Watch out young man. If do this, then
python will behave differently from fortran and matlab". That could
teach them to do things the right way.

I am not an expert on all this, but I guessed that it would be possible
to make a set of constraints that could catch a fair deal of simple
errors such as the one above, but still allow for quite a bit of
programming.

Niels

[Marc 'BlackJack' Rintsch]

In <11************ **********@n67g 2000cwd.googleg roups.com>, Niels L
Ellegaard wrote:

I have been using scipy for some time now, but in the beginning I made
a few mistakes with copying by reference.

But "copying by reference" is the way Python works. Python never copies
objects unless you explicitly ask for it. So what you want is a warning
for *every* assignment.

Ciao,
Marc 'BlackJack' Rintsch

[Niels L Ellegaard]

Marc 'BlackJack' Rintsch wrote:

In <11************ **********@n67g 2000cwd.googleg roups.com>, Niels L
Ellegaard wrote:

I have been using scipy for some time now, but in the beginning I made
a few mistakes with copying by reference.

But "copying by reference" is the way Python works. Python never copies
objects unless you explicitly ask for it. So what you want is a warning
for *every* assignment.

Maybe I am on the wrong track here, but just to clarify myself:

I wanted a each object to know whether or not it was being referred to
by a living object, and I wanted to warn the user whenever he tried to
change an object that was being refered to by a living object. As far
as I can see the garbage collector module would allow to do some of
this, but one would still have to edit the assignment operators of each
of the standard data structures:

http://docs.python.org/lib/module-gc.html

Anyway you are probably right that the end result would be a somewhat
crippled version of python
Niels

[Fredrik Lundh]

Niels L Ellegaard wrote:

I wanted a each object to know whether or not it was being referred to
by a living object, and I wanted to warn the user whenever he tried to
change an object that was being refered to by a living object. As far
as I can see the garbage collector module would allow to do som

*all* objects in Python are referred to by "living" objects; those that
don't are garbage, and are automatically destroyed sooner or later.

(maybe you've missed that namespaces are objects too?)

</F>

[Stuart D. Gathman]

On Sat, 09 Dec 2006 05:58:22 -0800, Niels L Ellegaard wrote:

I wanted a each object to know whether or not it was being referred to
by a living object, and I wanted to warn the user whenever he tried to
change an object that was being refered to by a living object. As far
as I can see the garbage collector module would allow to do some of
this, but one would still have to edit the assignment operators of each
of the standard data structures:

I think what you want is a namespace that requires each object to have
exactly one reference - the namespace. Of course, additional references
will be created during evaluation of expressions. So the best you can do
is provide a function that checks reference counts for a namespace when
called, and warns about objects with multiple references. If that could
be called for every statement (i.e. not during expression evaluation -
something like C language "sequence points"), it would probably catch the
type of error you are looking for. Checking such a thing efficiently would
require deep changes to the interpreter.

The better approach is to revel in the ease with which data can be
referenced rather than copied. I'm not sure it's worth turning python
into fortran - even for selected namespaces.

--
Stuart D. Gathman <st****@bmsi.co m>
Business Management Systems Inc. Phone: [phone] Fax: [phone]
"Confutatis maledictis, flamis acribus addictis" - background song for
a Microsoft sponsored "Where do you want to go from here?" commercial.

[Gabriel Genellina]

I think what you want is a namespace that requires each object to have

exactly one reference - the namespace. Of course, additional references
will be created during evaluation of expressions. So the best you can do

It's not enough. It won't catch the case where a list holds many
references to the same object - the example provided by the OP.

--
Gabriel Genellina

[Gabriel Genellina]

On 9 dic, 09:08, "Niels L Ellegaard" <niels.ellega.. [e-mail] e:

Now some of my fortran-using friends would like to use python to
analyze their data files. I wanted them to avoid making the same
mistakes as I did so I thought it would be good if they could get some
nanny-like warnings saying: "Watch out young man. If do this, then
python will behave differently from fortran and matlab". That could
teach them to do things the right way.

Best bet is to learn to use python the right way. It's not so hard...

I am not an expert on all this, but I guessed that it would be possible
to make a set of constraints that could catch a fair deal of simple
errors such as the one above, but still allow for quite a bit of
programming.

The problem is, it's an error only becasuse it's not the *intent* of
the programmer - but it's legal Python code, and useful. It's hard for
the computer to guess the intent of the programmer -yet!...-

--
Gabriel Genellina

[greg]

Niels L Ellegaard wrote:

I wanted to warn the user whenever he tried to
change an object that was being refered to by a living object.

To see how fundamentally misguided this idea is,
consider that, under your proposal, the following
program would produce a warning:

a = 1

The reason being that the assignment is modifying
the dictionary holding the namespace of the main
module, which is referred to by the main module
itself. So you are "changing an object that is
being referred to by a living object".

--
Greg