Hello fionbarr,

Yes, but apparently you can’t rely on the crypt function being available under Windows. From perlport#crypt:

May not be available if library or source was not provided when building perl. (Win32)

This may be important to bear in mind if your script is intended to run on platforms over which you have no direct control.

Hope that helps,

That warning is well over 20 years old. I doubt if there is any perl version for Windows where crypt isn't properly implemented.

.. and, related to that, I wonder if the different line-endings (in Linux and Windows) might result in a different crypt result. I only have Windows here at work, so I cannot test this hypothesis right now.

Alex / talexb / Toronto

Thanks PJ. We owe you so much. Groklaw -- RIP -- 2003 to 2013.

Line endings should only be an issue in I/O. By default perl strings use a single newline character as a record separator. PerlIO translates between this and whatever your OS requires. When processing files from another OS, specify the required translation on the open statement.
Bill

Excellent point....thanks

Appears to be:

Windows:

c:\> perl -E "say crypt 'string', 'sa'"
saQ9i9dEI8LLA
[download]

Ubuntu:

$ perl -E 'say crypt "string", "sa"'
saQ9i9dEI8LLA
[download]

Looks good, thank you

Is the 'crypt' function identical on Windows and *nix platforms?

No. See the thread starting at Re^2: replacing password in a file. Short: Many implementations use DES, but the libc is free to choose any other algorithm: http://pubs.opengroup.org/onlinepubs/9699919799/functions/crypt.html (<-- link is missing in that thread). Relevant quote:

The crypt() function is a string encoding function. The algorithm is implementation-defined.

Alexander