[issue38445] os.path.exists() takes bool as argument and returns True

[Serhiy Storchaka]

Serhiy Storchaka  added the comment:

It is a large change, so I take it.

--
assignee:  -> serhiy.storchaka

___
Python tracker 

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

[issue38445] os.path.exists() takes bool as argument and returns True

[Irit Katriel]

Change by Irit Katriel :


--
components: +IO
versions: +Python 3.10, Python 3.9 -Python 3.7

___
Python tracker 

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

[issue38445] os.path.exists() takes bool as argument and returns True

[Tal Einat]

Change by Tal Einat :


--
nosy:  -taleinat

___
Python tracker 

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

[issue38445] os.path.exists() takes bool as argument and returns True

[STINNER Victor]

Change by STINNER Victor :


--
nosy:  -vstinner

___
Python tracker 

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

[issue38445] os.path.exists() takes bool as argument and returns True

[Eryk Sun]

Eryk Sun  added the comment:

> Maybe os.fspath() can be used

I think that would already be the case if genericpath.exists didn't have to 
support file descriptors. It's documented that the path argument "refers to an 
existing path or an open file descriptor".

Modifying _fd_converter would provide consistent behavior for all calls that 
ultimately use argument clinic's path_t(allow_fd=True) and/or dir_fd. Serhiy's 
suggestion to raise a warning sounds good. 

Here are some examples that currently pass silently in Linux:

>>> os.chown(False, 1000, 1000)
>>> os.chmod(False, 0o666)
>>> os.utime(False, (15, 15))

Probably os.fstat and other "f" functions (e.g. fstatvfs, fchdir, fchown, 
fchmod, ftruncate, fdatasync, fsync, and fpathconf) should also raise a warning 
when passed a bool. For example, the following would raise a warning instead of 
passing silently:

>>> os.fstat(False).st_size
0

These cases could be addressed by consistently using an argument clinic type. 
Some of them already us the fildes type (e.g. fchdir, fsync, fdatasync). 
However, fildes_converter calls PyObject_AsFileDescriptor, which also supports 
objects with a fileno() method. That's documented behavior for 
PyObject_AsFileDescriptor, but nothing in the documentation of fchdir, fsync, 
and fdatasync suggests to me that they support objects with a fileno() method:

>>> os.fchdir(sys.stdin)
Traceback (most recent call last):
  File "", line 1, in 
NotADirectoryError: [Errno 20] Not a directory

If not for this behavior, we could simply change all of the "f" functions to 
use the fildes type.

PyObject_AsFileDescriptor is used in various other places as well, such as the 
select module. In the latter case, supporting objects with a fileno() method is 
clearly documented.

Also consider including open() and os.fdopen by modifying _io_open_impl in 
Modules/_io/_iomodule.c. For example:

   >>> open(False, closefd=False)
<_io.TextIOWrapper name=False mode='r' encoding='UTF-8'>

Currently it calls PyNumber_Check(file). If true, a warning could be raised if 
PyBool_Check(file) is also true.

--

___
Python tracker 

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

[issue38445] os.path.exists() takes bool as argument and returns True

[Tal Einat]

Tal Einat  added the comment:

Mariano Anaya, you're welcome to work on this and create a PR.

There's no need to have this issue assigned to you: feel free to just begin 
working. Anyone reading this will know that you intended to work on this.

(FYI, we only assign issues to core developers, and assigning means something 
slightly different than "who is working on this.")

--
nosy: +taleinat

___
Python tracker 

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

[issue38445] os.path.exists() takes bool as argument and returns True

[STINNER Victor]

STINNER Victor  added the comment:

Maybe os.fspath() can be used:

>>> os.fspath(True)
Traceback (most recent call last):
  File "", line 1, in 
TypeError: expected str, bytes or os.PathLike object, not bool

--

___
Python tracker 

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

[issue38445] os.path.exists() takes bool as argument and returns True

[Serhiy Storchaka]

Serhiy Storchaka  added the comment:

Eryk, I think testing for bool in _fd_converter and issuing a warning is a good 
idea. Do you want to create a PR?

--
nosy: +serhiy.storchaka

___
Python tracker 

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

[issue38445] os.path.exists() takes bool as argument and returns True

[Joannah Nanjekye]

Change by Joannah Nanjekye :


--
nosy: +vstinner

___
Python tracker 

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

[issue38445] os.path.exists() takes bool as argument and returns True

[Eryk Sun]

Eryk Sun  added the comment:

Note that the underlying stat call supports file descriptors, which are 
non-negative integers. This is a supported and tested capability for 
genericpath.exists (see GenericTest.test_exists_fd in 
Lib/test/test_genericpath.py).

False and True are integers with the values 0 and 1: 

>>> issubclass(bool, int)
True
>>> False + 0
0
>>> True + 0
1

That can be useful, but there may be cases where we don't want to conflate 
bools and integers. IMO, a bool should not be supported as a file descriptor. 
It's likely a bug that should be caught early instead of meaninglessly 
propagated. 

A high-level solution would check for bool instances in genericpath.exists. A 
low-level solution, to make this policy consistent in general, would be to 
modify _fd_converter in Modules/posixmodule.c to disallow bool instances.

--
nosy: +eryksun

___
Python tracker 

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

[issue38445] os.path.exists() takes bool as argument and returns True

[Mariano Anaya]

Mariano Anaya  added the comment:

Hi, I would like to take a look at this issue. Could someone please assign it 
to me? Thanks

--
nosy: +Mariano Anaya

___
Python tracker 

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

[issue38445] os.path.exists() takes bool as argument and returns True

[Nika]

New submission from Nika :

os.path.exists() accepts either True or False as argument and returns True.
Reproducible on Windows, e. g.,  in jupyter notebook or in Eclipse, although 
not in IDLE, which returns False, as expected.

import os
print(os.path.exists(False))

--
messages: 354440
nosy: Nika
priority: normal
severity: normal
status: open
title: os.path.exists() takes bool as argument and returns True
type: behavior
versions: Python 3.7

___
Python tracker 

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