[Python-Dev] Updates to PEP 471, the os.scandir() proposal

[MRAB]

On 2014-07-09 23:50, Ethan Furman wrote:
> On 07/09/2014 02:33 PM, Ben Hoyt wrote:
>>>
>>> On a system which did not supply is_dir automatically I would write that as:
>>>
>>>    for entry in os.scandir(path):
>>>        if ignore_entry(entry.name):
>>>            continue
>>>        if os.path.isdir(entry.full_name):
>>>            # do something interesting
>>>
>>> Not hard to read or understand, no time wasted in unnecessary lstat calls.
>>
>> No, but how do you know whether you're on "a system which did not
>> supply is_dir automatically"? The above is not cross-platform, or at
>> least, not efficient cross-platform, which defeats the whole point of
>> scandir -- the above is no better than listdir().
>
> Hit a directory with 100,000 entries and you'll change your mind.  ;)
>
> Okay, so the issue is you /want/ to write an efficient, cross-platform routine...
>
> hrmmm.....
>
> thinking........
>
> Okay, marry the two ideas together:
>
>     scandir(path, info=None, onerror=None)
>         """
>         Return a generator that returns one directory entry at a time in a DirEntry object

Should that be "that yields one directory entry at a time"?

>         info:  None --> DirEntries will have whatever attributes the O/S provides
>                'type'  --> DirEntries will already have at least the file/dir distinction
>                'stat'  --> DirEntries will also already have stat information
>         """
>
>     DirEntry.is_dir()
>        Return True if this is a directory-type entry; may call os.lstat if the cache is empty.
>
>     DirEntry.is_file()
>        Return True if this is a file-type entry; may call os.lstat if the cache is empty.
>
>     DirEntry.is_symlink()
>        Return True if this is a symbolic link; may call os.lstat if the cache is empty.
>
>     DirEntry.stat
>        Return the stat info for this link; may call os.lstat if the cache is empty.
>
Why is "is_dir", et al, functions, but "stat" not a function?

>
> This way both paradigms are supported.
>