[Python-ideas] Adding '**' recursive search to glob.glob (original) (raw)

Steven D'Aprano steve at pearwood.info
Mon Jan 14 17:24:20 CET 2013

On 15/01/13 02:46, Vinay Sajip wrote:

Paul Moore<p.f.moore at ...> writes:

I'd like it if the glob module supported the (relatively common) facility to use ** to mean recursively search subdirectories.

+1

One obvious downside is that if used carelessly, it can make globbing pretty slow. So I'd propose that it be added as an optional extension enabled using a flag argument (glob(pat, allowrecursive=True)) which is false by default. That would also mean that backward compatibility should not be an issue. Isn't the requirement to recurse implied by the presence of '**' in the pattern? What's to be gained by specifying it using allowrecursive as well?

Not necessarily. At the moment, a glob like "/**/spam" is equivalent to "/*/spam":

[steve at ando /]$ touch /tmp/spam [steve at ando /]$ mkdir /tmp/ham [steve at ando /]$ touch /tmp/ham/spam [steve at ando /]$ python3.3 -c "import glob; print(glob.glob('/**/spam'))" ['/tmp/spam']

With the suggested new functionality, the meaning of the glob will change.

From a backwards-compatibility point of view, one might not want to enable the new semantics by default. But, from a future-compatibility point of view, I don't know that it is a good idea to require a flag every time a new globbing feature is added.

glob.glob(pattern, allow_recurse=True, allow_spam=True, allow_ham=True, allow_eggs=True, ...)

Rather than a flag, I suggest a version number:

glob.glob(pattern, version=1) # current behaviour, as of 3.3 glob.glob(pattern, version=2) # adds ** recursion in Python 3.4

Then in Python 3.5 or 3.6 support for version 1 globs could be dropped.

Will having allowrecursive=True have any effect if '**' is not in the pattern?

I would expect that it will not have any effect unless ** is present. After all, it simply allows ** to recurse, and no other glob metacharacter can recurse.

If you specify a pattern with '**' and allowrecursive=False, does that mean that '**' effectively acts as '*' would (i.e. one directory level only)?

I expect that without allow_recursive=True, ** would behave identically to a single *

-- Steven

More information about the Python-ideas mailing list