Minimal Perl for Unix and Linux People | Page 5

[previous]

6.3.2 Finding files by pathname matching

You've seen that Perl can mimic the behavior of find's –name option by matching within a pathname's final segment. But Perl can do something the POSIX find can't do—it can match anywhere within the pathname.

I use this feature periodically when I need to find one of my scripts by name. Because I know it resides in one of my *bin directories, I can use this knowledge to avoid matches with like-named files that reside in other directories, such as *man and *lib.

For instance, here's how Homer would locate his scripts for converting images in other formats to JPEGs, using the fact that he employs a *2jpg convention in naming them:

Notice that the mX syntax of the matching operator is used to specify the ":" character as the delimiter, overriding the default of the slash. This allows the slash at the end of "bin/" to unambiguously represent the directory separator in the pathname, which creates a context that ensures the \w+2jpg$ pattern is only matched as a whole filename under a *bin directory. By using this technique of matching filenames only within directories matching specified patterns, Homer can avoid undesirable matches such as these:

Next, you'll see how to use a special kind of find | perl pipeline for filtering out undesirable arguments for Unix utilities, an d how to use Unix utilities for validating arguments for Perl programs.

Excerpted from Minimal Perl for Unix and Linux People by Tim Maher. Copyright © 2006. Used with permission of Manning Publications Co..

[previous]

URL: