Course of Raku / Objects, I/O, and exceptions / Input and output / Running external programs 🆕

Capturing output

To read what an external program prints rather than letting it go to the screen, pass :out when launching it. The returned Proc then has an .out handle you can read from:

my $proc = run 'echo', 'hello', :out;
my $output = $proc.out.slurp(:close);
say $output.chomp; # hello

:out redirects the program’s standard output into the Proc. The .out handle works like the file handles from the previous section, and .slurp reads everything from it at once. The :close adverb is a handy shortcut: after reading, it closes the handle for you — the same tidy-up you did explicitly with .close on a file handle — so nothing is left open. Finally, .chomp removes the trailing newline that echo adds.

The same works with shell, which is handy for capturing the result of a pipeline:

my $proc = shell 'echo hello | tr a-z A-Z', :out;
say $proc.out.slurp(:close).chomp; # HELLO

You can capture the error stream too, with :err and the matching .err handle. Passing both :out and :err keeps the two streams apart, each in its own handle, so you can read them separately:

my $proc = run 'sh', '-c', 'echo out-line; echo err-line >&2', :out, :err;

say 'stdout: ', $proc.out.slurp(:close).chomp; # stdout: out-line
say 'stderr: ', $proc.err.slurp(:close).chomp; # stderr: err-line

Here, the shell command prints one line to standard output and one to standard error (>&2 redirects to it). Because we asked for both streams, .out and .err hold them independently — the program’s normal output never gets mixed up with its diagnostics.

Capturing output is how you use an external program as a building block — run it, read its result, and carry on with that data in your own program.

Course navigation

← Quiz — Running programs   |   The Proc object →

💪 Or jump directly to the exercises in this section.