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

[Victor Stinner]

2014-07-09 21:59 GMT+02:00 Ben Hoyt <benhoyt at gmail.com>:
> Other python-devers, please chime in with your thoughts or votes.

Sorry, I didn't follow the whole discussion. IMO DirEntry must use
methods and you should not expose nor document which infos are already
provided by the OS or not. DirEntry should be a best-effort black-box
object providing an API similar to pathlib.Path. is_dir() may be fast?
fine, but don't say it in the documentation because Python must remain
portable and you should not write code specific to one specific
platform.

is_dir(), is_file(), is_symlink() and lstat() can fail as any other
Python function, no need to specialize them with custom error handler.
If you care, just use a very standard try/except.

I'm also against ensure_lstat=True or ideas like that fetching all
datas transparently in the generator. The behaviour would be too
different depending on the OS, and usually you don't need all
informations. And it raises errors in the generator, which is
something unusual, and difficult to handle (I don't like the onerror
callback).

Example where you may sometimes need is_dir(), but not always
---
for entry in os.scandir(path):
  if ignore_entry(entry.name):
     # this entry is not interesting, lstat_result is useless here
     continue
  if entry.is_dir():  # fetch required data if needed
     continue
  ...
---

I don't understand why you are all focused on handling os.stat() and
os.lstat() errors. See for example the os.walk() function which is an
old function (python 2.6!): it doesn't catch erros on isdir(), even if
it has an onerror parameter... It only handles errors on listdir().
IMO errors on os.stat() and os.lstat() are very rare under very
specific conditions. The most common case is that you can get the
status if you can list files.

Victor