'Using is not None, may not always work'

[Doug Fort]

Hi,

Since I installed 2.4a2 I've been getting a warning from pychecker: Using
is not None, may not always work'. I thought 'is not None' was the right
thing to do. I've had problems with 'if not x:', because some objects
return False in this context.

--
Doug Fort, Consulting Programmer

http://www.dougfort.com

[Heiko Wundram]

Re: 'Using is not None, may not always work'

Am Freitag, 6. August 2004 15:15 schrieb Doug Fort:[color=blue]
> Since I installed 2.4a2 I've been getting a warning from pychecker: Using
> is not None, may not always work'. I thought 'is not None' was the right
> thing to do. I've had problems with 'if not x:', because some objects
> return False in this context.[/color]

That's probably what pychecker warns you about: that you might get an object
in the respective context which evaluates to boolean false, but is not
None... Although I'd find this strange...

Heiko.

[Michael Hudson]

Re: 'Using is not None, may not always work'

Heiko Wundram <heikowu@ceosg. de> writes:
[color=blue]
> Am Freitag, 6. August 2004 15:15 schrieb Doug Fort:[color=green]
> > Since I installed 2.4a2 I've been getting a warning from pychecker: Using
> > is not None, may not always work'. I thought 'is not None' was the right
> > thing to do. I've had problems with 'if not x:', because some objects
> > return False in this context.[/color]
>
> That's probably what pychecker warns you about: that you might get an object
> in the respective context which evaluates to boolean false, but is not
> None... Although I'd find this strange...[/color]

Does seem an odd warning, seeing as when I write "x is not None" I'm
generally avoiding the case of, e.g., x being the empty list.

Cheers,
mwh

--
Well, you pretty much need Microsoft stuff to get misbehaviours
bad enough to actually tear the time-space continuum. Luckily
for you, MS Internet Explorer is available for Solaris.
-- Calle Dybedahl, alt.sysadmin.re covery

[Grant Edwards]

Re: 'Using is not None, may not always work'

On 2004-08-06, Doug Fort <dougfort@dougf ort.com> wrote:
[color=blue]
> Since I installed 2.4a2 I've been getting a warning from
> pychecker: Using is not None, may not always work'.[/color]

Huh?!

That implies that there are cases where "None is not None" will
evaluate as True. I find that extremely hard to believe.
[color=blue]
> I thought 'is not None' was the right thing to do.[/color]

It's certainly what I want to do if I want to know if a name is
bound to None or not.
[color=blue]
> I've had problems with 'if not x:', because some objects
> return False in this context.[/color]

Exactly.

--
Grant Edwards grante Yow! I was in a HOT
at TUB! I was NORMAL! I was
visi.com ITALIAN!! I enjoyed th'
EARTHQUAKE!

[Peter Otten]

Re: 'Using is not None, may not always work'

Doug Fort wrote:
[color=blue]
> Since I installed 2.4a2 I've been getting a warning from pychecker: Using
> is not None, may not always work'. I thought 'is not None' was the right
> thing to do. I've had problems with 'if not x:', because some objects
> return False in this context.[/color]

This is harmless. Starting with 2.4a2 None is a constant:

Python 2.3.3 (#1, Jan 3 2004, 13:57:08)
[GCC 3.2] on linux2
Type "help", "copyright" , "credits" or "license" for more information.[color=blue][color=green][color=darkred]
>>> import dis
>>> def f():[/color][/color][/color]
.... if x is None: pass
....[color=blue][color=green][color=darkred]
>>> dis.dis(f)[/color][/color][/color]
2 0 LOAD_GLOBAL 0 (x)
3 LOAD_GLOBAL 1 (None)
6 COMPARE_OP 8 (is)
9 JUMP_IF_FALSE 4 (to 16)
12 POP_TOP
13 JUMP_FORWARD 1 (to 17)[color=blue][color=green]
>> 16 POP_TOP
>> 17 LOAD_CONST 0 (None)[/color][/color]
20 RETURN_VALUE[color=blue][color=green][color=darkred]
>>>[/color][/color][/color]

Python 2.4a2 (#1, Aug 6 2004, 16:38:38)
[GCC 3.2] on linux2
Type "help", "copyright" , "credits" or "license" for more information.
[python 2.4a2][color=blue][color=green][color=darkred]
>>> import dis
>>> def f():[/color][/color][/color]
.... if x is None: pass
....[color=blue][color=green][color=darkred]
>>> dis.dis(f)[/color][/color][/color]
2 0 LOAD_GLOBAL 0 (x)
3 LOAD_CONST 0 (None)
6 COMPARE_OP 8 (is)
9 JUMP_IF_FALSE 4 (to 16)
12 POP_TOP
13 JUMP_FORWARD 1 (to 17)[color=blue][color=green]
>> 16 POP_TOP
>> 17 LOAD_CONST 0 (None)[/color][/color]
20 RETURN_VALUE


When PyChecker sees the constant it supposes you are doing something like

if x is "some string": pass

which always had the LOAD_CONST op-code and is indeed dangerous. PyChecker
needs to be fixed to special-case None for 2.4.

Peter

[Doug Fort]

Re: 'Using is not None, may not always work'

On Fri, 06 Aug 2004 17:02:43 +0200, Peter Otten wrote:
[color=blue]
> Doug Fort wrote:
>[color=green]
>> Since I installed 2.4a2 I've been getting a warning from pychecker: Using
>> is not None, may not always work'. I thought 'is not None' was the right
>> thing to do. I've had problems with 'if not x:', because some objects
>> return False in this context.[/color]
>
> This is harmless. Starting with 2.4a2 None is a constant:[/color]

Thanks, that it.
[color=blue]
>
> Python 2.3.3 (#1, Jan 3 2004, 13:57:08)
> [GCC 3.2] on linux2
> Type "help", "copyright" , "credits" or "license" for more information.[color=green][color=darkred]
>>>> import dis
>>>> def f():[/color][/color]
> ... if x is None: pass
> ...[color=green][color=darkred]
>>>> dis.dis(f)[/color][/color]
> 2 0 LOAD_GLOBAL 0 (x)
> 3 LOAD_GLOBAL 1 (None)
> 6 COMPARE_OP 8 (is)
> 9 JUMP_IF_FALSE 4 (to 16)
> 12 POP_TOP
> 13 JUMP_FORWARD 1 (to 17)[color=green][color=darkred]
> >> 16 POP_TOP
> >> 17 LOAD_CONST 0 (None)[/color][/color]
> 20 RETURN_VALUE[color=green][color=darkred]
>>>>[/color][/color]
>
> Python 2.4a2 (#1, Aug 6 2004, 16:38:38)
> [GCC 3.2] on linux2
> Type "help", "copyright" , "credits" or "license" for more information.
> [python 2.4a2][color=green][color=darkred]
>>>> import dis
>>>> def f():[/color][/color]
> ... if x is None: pass
> ...[color=green][color=darkred]
>>>> dis.dis(f)[/color][/color]
> 2 0 LOAD_GLOBAL 0 (x)
> 3 LOAD_CONST 0 (None)
> 6 COMPARE_OP 8 (is)
> 9 JUMP_IF_FALSE 4 (to 16)
> 12 POP_TOP
> 13 JUMP_FORWARD 1 (to 17)[color=green][color=darkred]
> >> 16 POP_TOP
> >> 17 LOAD_CONST 0 (None)[/color][/color]
> 20 RETURN_VALUE
>
>
> When PyChecker sees the constant it supposes you are doing something like
>
> if x is "some string": pass
>
> which always had the LOAD_CONST op-code and is indeed dangerous. PyChecker
> needs to be fixed to special-case None for 2.4.
>
> Peter[/color]

--
Doug Fort, Consulting Programmer

http://www.dougfort.com