NAME

Shell::Perl - A read-eval-print loop in Perl

SYNOPSYS

use Shell::Perl;
Shell::Perl->run_with_args;

DESCRIPTION

This is the implementation of a command-line interpreter for Perl. I wrote this because I was tired of using irb when needing a calculator with a real language within. Ah, that and because it was damn easy to write it.

This module is the heart of the pirl script provided with Shell-Perl distribution, along with this module.

EXAMPLE SESSION

$ pirl
Welcome to the Perl shell. Type ':help' for more information


pirl @> 1+1
2

pirl @> use YAML qw(Load Dump);
()

pirl @> $data = Load("--- { a: 1, b: [ 1, 2, 3] }\n");
{ a => 1, b => [1, 2, 3] }

pirl @> $var = 'a 1 2 3'; $var =~ /(\w+) (\d+) (\d+)/
("a", 1, 2)

pirl @> :q

COMMANDS

Most of the time, the shell reads Perl statements, evaluates them and outputs the result.

There are a few commands (started by ':') that are handled by the shell itself.

:h(elp)

Handy for remembering what the shell commands are.

:q(uit)

Leave the shell. The Perl statement exit will work too.

SYNONYMS: :exit, :x

:set out (D|DD|DDS|Y|P)

Changes the dumper for the expression results used before output. The current supported are:

D

Data::Dump

DD

Data::Dumper, the good and old core module

DDS

Data::Dump::Streamer

Y

YAML

P

a plain dumper ("$ans" or "@ans")

When creating the shell, the dump format is searched among the available ones in the order "D", "DD", "DDS", "Y" and "P". That means Data::Dump is preferred and will be used if available/installed. Otherwise, Data::Dumper is tried, and so on.

Read more about dumpers at Shell::Perl::Dumper.

:set ctx (scalar|list|void|s|l|v|$|@|_)

Changes the default context used to evaluate the entered expression. The default is 'list'.

Intuitively, 'scalar', 's' and '$' are synonyms, just like 'list', 'l', and '@' or 'void', 'v', '_'.

There is a nice way to override the default context in a given expression. Just a '#' followed by one of 'scalar|list|void|s|l|v|$|@|_' at the end of the expression.

pirl @> $var = 'a 1 2 3'; $var =~ /(\w+) (\d+) (\d+)/
("a", 1, 2)

pirl @> $var = 'a 1 2 3'; $var =~ /(\w+) (\d+) (\d+)/ #scalar
1
:set perl_version

Changes the perl version (and current feature bundle) used to evaluate each statement. Usage examples are:

:set perl_version 5.008
:set perl_version v5.10
:set perl_version        # current perl version, $]

Default is to use the current perl version, which works like eval "use $];".

Set to an empty string, as in

:set perl_version ''

for the behavior of pirl 0.0023 or earlier.

:reset

Resets the environment, erasing the symbols created at the current evaluation package. See the section "ABOUT EVALUATION".

METHODS

Remember this is an alpha version, so the API may change and that includes the methods documented here. So consider this section as implementation notes for a while.

In later versions, some of these information may be promoted to a public status. Others may be hidden or changed and even disappear without further notice.

new
$sh = Shell::Version->new;

The constructor.

run_with_args
Shell::Perl->run_with_args;

Starts the read-eval-print loop after reading options from @ARGV. It is a class method.

If an option -v or --version is provided, instead of starting the REPL, it prints the script identification and exits with 0.

$ pirl -v
This is pirl, version 0.0017 (bin/pirl, using Shell::Perl 0.0017)
run
$sh->run;

The same as run_with_args but with no code for interpreting command-line arguments. It is an instance method, so that Shell::Perl->run_with_args is kind of:

Shell::Perl->new->run;
eval
$answer = $sh->eval($exp);
@answer = $sh->eval($exp);

Evaluates the user input given in $exp as Perl code and returns the result. That is the 'eval' part of the read-eval-print loop.

print
$sh->print(@args);

Prints a list of args at the output stream currently used by the shell.

help
$sh->help;

Outputs the help as provided by the command ":help".

reset
$sh->reset;

Does nothing by now, but it will.

dump_history
$sh->dump_history();
$sh->dump_history($file);

Prints the readline history to STDOUT or the optional file. Used to implement experimental command ":dump history".

This is experimental code and should change in the future. More control should be added and integrated with other terminal features.

set_ctx
$sh->set_ctx($context);

Assigns to the current shell context. The argument must be one of ( 'scalar', 'list', 'void', 's', 'l', 'v', '$', '@', '_' ) .

set_package
$sh->set_package($package);

Changes current evaluation package. Doesn't change if the new package name is malformed.

set_perl_version
$sh->set_perl_version($version);

Changes perl version used to evaluate statements.

set_out
$sh->set_out($dumper);

Changes the current dumper used for printing the evaluation results. Actually must be one of "D" (for Data::Dump), "DD" (for Data::Dumper), "DDS" (for Data::Dump::Streamer), "Y" (for YAML) or "P" (for plain string interpolation).

prompt_title
$prompt = $sh->prompt_title;

Returns the current prompt which changes with executable name and context. For example, "pirl @>", "pirl $>", and "pirl >".

quit
$sh->quit;

This method is invoked when these commands and statements are parsed by the REPL:

:q
:quit
:x
:exit
quit
exit

It runs the shutdown procedures for a smooth termination of the shell. For example, it saves the terminal history file.

GORY DETAILS

ABOUT EVALUATION

When the statement read is evaluated, this is done at a different package, which is Shell::Perl::sandbox by default.

So:

$ perl -Mlib=lib bin/pirl
Welcome to the Perl shell. Type ':help' for more information

pirl @> $a = 2;
2

pirl @> :set out Y # output in YAML

pirl @> \%Shell::Perl::sandbox::
---
BEGIN: !!perl/glob:
  PACKAGE: Shell::Perl::sandbox
  NAME: BEGIN
a: !!perl/glob:
  PACKAGE: Shell::Perl::sandbox
  NAME: a
  SCALAR: 2

This package serves as an environment for the current shell session and :reset can wipe it away.

pirl @> :reset

pirl @> \%Shell::Perl::sandbox::
---
BEGIN: !!perl/glob:
  PACKAGE: Shell::Perl::sandbox
  NAME: BEGIN

TO DO

There is a lot to do, as always. Some of the top priority tasks are:

  • Accept multiline statements;.

  • Refactor the code to promote easy customization of features.

BUGS

It is a one-line evaluator by now.

I don't know what happens if you eval within an eval. I don't expect good things to come. (Lorn who prodded me about this will going to find it out and then I will tell you.)

There are some quirks with Term::Readline (at least on Windows).

There are more bugs. I am lazy to collect them all and list them now.

Please report bugs via Github https://github.com/aferreira/pirl/issues.

SEE ALSO

This project is hosted at Github:

https://github.com/aferreira/pirl

To know about interactive Perl interpreters, there are two FAQS contained in perlfaq3 which are good starting points. Those are

How can I use Perl interactively?
http://perldoc.perl.org/perlfaq3.html#How-can-I-use-Perl-interactively%3f

Is there a Perl shell?
http://perldoc.perl.org/perlfaq3.html#How-can-I-use-Perl-interactively%3f

Also:

AUTHORS

Adriano R. Ferreira, <ferreira@cpan.org>

Caio Marcelo, <cmarcelo@gmail.com>

Ron Savage, <ron@savage.net.au>

COPYRIGHT AND LICENSE

Copyright (C) 2007–2017 by Adriano R. Ferreira

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.