CODEWITHSUNDEEP
perlstat
allmail
allmail

Reputation: 133

perl File::Find and _

I am using File::Find to go through a ton of files and then doing cleanup based on their size, age etc. I am accessing these files through windows network shares so the stat calls are expensive

The code looked something like this

sub findItems
{
  my $item = $File::Find::name;
  my $dir  = $File::Find::dir;
  my $itemName = $_;

  if    (-f $item) { .. }
  elsif (-d $item) { .. }

  my $age = -M $item;
  my $size = -s $item;

Figured out that it was taking a lot of time. The culprit was the 3 calls -f/-d, -M and -s, all needing a fresh stat. So I changed it to

sub findItems
{
  my $item = $File::Find::name;
  my $dir  = $File::Find::dir;
  my $itemName = $_;

  if    (-f $item) { .. }
  elsif (-d $item) { .. }

  my $age = -M _;
  my $size = -s _;

This almost halved the time as I went from 4 stat calls to 2. I said 4 because I know there is one stat call in File::find as well, and 3 are in my code.

My question is can I rely on the stat call in File::Find and change my code to this

sub findItems
{
  my $item = $File::Find::name;
  my $dir  = $File::Find::dir;
  my $itemName = $_;

  if    (-f _) { .. }
  elsif (-d _) { .. }

  my $age = -M _;
  my $size = -s _;

This will again halve the time it takes. I guess I will have to actually check the File::Find code to see if there might be some other stat call on anything other than $item

thanks

Upvotes: 4

Views: 117

Answers (1)

Tanktalus
Tanktalus

Reputation: 22254

Based on its documentation, "It is guaranteed that an lstat has been called before the user's wanted() function is called. This enables fast file checks involving _. Note that this guarantee no longer holds if follow or follow_fast are not set."

So I'd suggest that, under any other circumstance, you cannot rely on it.

Upvotes: 3

Related Questions