[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Dong-hee Na]

Dong-hee Na  added the comment:

rm-finite.patch was applied by bpo-39415.

> Once that's merged, let's close this issue as rejected.

I close this issue to a rejected state as discussed above.
Thank you to all who participated in this discussion.

--
nosy: +corona10
resolution:  -> rejected
stage: patch review -> resolved
status: open -> closed

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[STINNER Victor]

Changes by STINNER Victor :


--
nosy: +haypo

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Mariatta Wijaya]

Mariatta Wijaya added the comment:

Martin,

I reviewed the patch (rm-finite.patch) and ran the tests. The patch no longer 
applies cleanly to the default branch, but otherwise looks good. 

Assigning this back to you as per Raymond's suggestion.

Thanks :)

--
assignee: Mariatta -> martin.panter

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Raymond Hettinger]

Raymond Hettinger added the comment:

Mariatta, please review the rm-finite patch.  We're already agreed that the 
removal makes sense.  What you're checking for is whether patch is complete 
(that there aren't any dangling references to the removed code and that it 
doesn't remove too much).  Also, try out the patch and run the test suite in in 
a debug mode.  

Whether you find any issues or are ready to bless the patch, make a comment in 
the tracker and assign back to Martin (the core dev who posts the patch is 
usually the one to apply it).  After merging, he will mark the class as 
closed/rejected.

--
assignee:  -> Mariatta
nosy: +Mariatta

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Mark Dickinson]

Mark Dickinson added the comment:

rm-finite.patch LGTM.

Once that's merged, let's close this issue as rejected. If decimal is ever 
added as a first-class type (literal support, fixed-width decimal128 or 
decimal64, etc.), we may want to reconsider, but that day is probably a long 
way off.

--

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Martin Panter]

Martin Panter added the comment:

While I don’t have much opinion either way, here is a patch to remove the 
existing dead code should that be the eventual outcome.

If the default implementations in the base class deferred to math.isfinite() 
etc, that may help with compatibility.

--
versions: +Python 3.7 -Python 3.4
Added file: http://bugs.python.org/file46451/rm-finite.patch

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Serhiy Storchaka]

Serhiy Storchaka added the comment:

I concur with Raymond. This expands the API too much. Not just the float API, 
but the API of all numeric classes, including third-party classes. And since 
the existence of additional method in third-party classes (e.g. NumPy classes) 
can't be guarantied, this couldn't help in the case of msg286441.

The less harmful way is making math.isfinite() and like supporting non-float 
numeric types. Always return False for integer types and fall back to the 
is_finite() method. But even this can be too expensive.

--

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Raymond Hettinger]

Raymond Hettinger added the comment:

I would like to caution against expansion of core APIs in cases where we 
already have a way to do it.  In almost every corner, Python's APIs are 
growing, resulting in the language becoming massive, where very few people on 
the planet can claim to know what is in it and it is becoming much harder to 
teach as it gets bigger.

Also, we should place very little value on anecdotal evidence from one core 
developer who says he once wrote a single line of code that would have 
benefitted from a single part of the proposed changes.  In general, core devs 
tend to write much more generic code than end users, so our needs are often 
atypical and do not reflect user needs (in this case, no non-core-dev numeric 
user has ever requested this behavior or had a real use case that couldn't be 
met by the existing APIs).  The bar for new APIs is much higher than "I wrote a 
line of code once".  

We should be looking for APIs additions that significantly reduce complexity or 
solve problems that can't easily be met with existing APIs.  Minor expansions 
create long term maintenance burdens, increase the size of docs making them 
less usable, cause other implementations to have to follow suit, cause 
subclassers and users of the numeric module to have to implement additional 
methods, and make Python more difficult to learn.  There is a reason that the 
zen expresses a preference for only one way to do it.

Replicating parts the decimal API should always be suspect as well.  In 
general, that module is much more difficult to use and learn than regular 
floats.  Many parts of the API are simply worthless and are there only because 
they were part of the spec.  We have no evidence that any of these proposed 
methods are actually being used and are of real benefit to real users.

Also, remember that Decimal is not registered as a Real for a reason.  Guido 
does not consider them to be interoperable with binary floats.

--

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Mark Dickinson]

Mark Dickinson added the comment:

@Martin: the dead code should definitely be removed from floatobject.c and 
longobject.c. (Though that's kinda independent of this issue.)

--

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Martin Panter]

Martin Panter added the comment:

Of course, somehow I missed that

--

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Steven D'Aprano]

Steven D'Aprano added the comment:

On Sun, Jan 29, 2017 at 08:23:05AM +, Martin Panter wrote:
> Why do you name the methods is_finite() etc with underscores, when the 
> existing methods math.isfinite() etc do not have underscores? Seems it 
> would add unnecessary confusion.

The idea is to enable duck-typing between float and Decimal. Instead of:

if isinstance(x, float):
math.isfinite(x)
else:
x.is_finite()

we can just say

x.is_finite() 

on both floats and decimals.

--

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Martin Panter]

Martin Panter added the comment:

FWIW, here is an attempt to add Argument Clinic to the Objects/floatobject.c 
and Objects/longobject.c implementations:

https://bugs.python.org/file33943/issue20185_conglomerate_v4.diff
https://bugs.python.org/file33989/clinic_longobject_v3.patch

If the methods are rejected, the dead code should be removed to avoid wasting 
people’s time.

Why do you name the methods is_finite() etc with underscores, when the existing 
methods math.isfinite() etc do not have underscores? Seems it would add 
unnecessary confusion.

--
nosy: +martin.panter

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Stefan Krah]

Changes by Stefan Krah stefan-use...@bytereef.org:


--
nosy:  -skrah

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue18842
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Serhiy Storchaka]

Serhiy Storchaka added the comment:

Should we add these methods to other concrete Number subclasses (as Fraction 
and complex)?

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue18842
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Steven D'Aprano]

Steven D'Aprano added the comment:

On Sat, Nov 16, 2013 at 04:33:36PM +, Serhiy Storchaka wrote:
 
 Should we add these methods to other concrete Number subclasses (as Fraction 
 and complex)?

Seems like a good idea to me. Is it worth making them part of the Number 
ABC, or is that too much of a change?

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue18842
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Mark Dickinson]

Mark Dickinson added the comment:

+1 from me.  You want `float.is_infinite` rather than `float.is_infinity`.  
`is_signed` is another one that may be worth considering.

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue18842
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Christian Heimes]

Christian Heimes added the comment:

The code is already there so the patch is really small.

http://hg.python.org/cpython/annotate/5bc7b20dc04a/Objects/floatobject.c#l1046

I love my time machine. *g*

--
keywords: +patch
nosy: +christian.heimes
Added file: http://bugs.python.org/file32089/float_is.patch

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue18842
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Christian Heimes]

Christian Heimes added the comment:

Here is a longer patch that also adds the methods to int and numbers.Real. It 
comes with tests and doc updates, too.

--
stage:  - patch review
Added file: http://bugs.python.org/file32090/issue18842.patch

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue18842
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Serhiy Storchaka]

Serhiy Storchaka added the comment:

Fraction? complex? Complex? Integral? Number?

--
nosy: +serhiy.storchaka

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue18842
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Christian Heimes]

Christian Heimes added the comment:

Pardon?

The methods could be added to complex, too. cmath implements the methods as:

is_finite:
Py_IS_FINITE(z.real)  Py_IS_FINITE(z.imag)
is_infinite:
Py_IS_INFINITY(z.real) || Py_IS_INFINITY(z.imag)
is_nan:
Py_IS_NAN(z.real) || Py_IS_NAN(z.imag)

For numbers.Real: We can't make the methods abstractmethods because it would be 
an incompatible change.

--

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue18842
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Steven D'Aprano]

New submission from Steven D'Aprano:

On issue 15544 Mark Dickinson suggested adding methods to float to match 
methods on Decimal, giving type-agnostic ways of testing real numbers that 
don't rely on converting to float. I don't see any sign that Mark raised a 
feature request, so I'm taking the liberty of doing so myself.

Note that the math.is* functions convert to float first, which means that they 
behave differently. Example: math.isnan(Decimal('sNAN')) raises ValueError, 
rather than returning True.

float.is_nan
float.is_infinity
float.is_finite

would mirror the spelling of Decimal methods, rather than the math module. As 
float doesn't support signalling NANs, there probably isn't any need to include 
is_snan and is_qnan.

For what it's worth, I have code that would use this. I currently write 
something like:

if isinstance(x, Decimal) and x.is_nan() or math.isnan(x): ...

in order to prevent triggering the ValueError on signalling NANs.

--
messages: 196219
nosy: stevenjd
priority: normal
severity: normal
status: open
title: Add float.is_finite is_nan is_infinite to match Decimal methods
type: enhancement
versions: Python 3.4

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue18842
___
___
Python-bugs-list mailing list
Unsubscribe: 
http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue18842] Add float.is_finite is_nan is_infinite to match Decimal methods

[Serhiy Storchaka]

Changes by Serhiy Storchaka storch...@gmail.com:


--
nosy: +facundobatista, mark.dickinson, rhettinger, skrah

___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue18842
___
___
Python-bugs-list mailing list
Unsubscribe: 
http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com