NAME

String::Formatter - build sprintf-like functions of your own

VERSION

version 1.235

SYNOPSIS

  use String::Formatter stringf => {
    -as   => 'str_rf',
    codes => {
      f => sub { $_ },
      b => sub { scalar reverse $_ },
      o => 'Okay?',
    },
  };

  print str_rf('This is %10f and this is %-15b, %o', 'forward', 'backward');

...prints...

  This is    forward and this is drawkcab       , okay?

DESCRIPTION

String::Formatter is a tool for building sprintf-like formatting routines. It supports named or positional formatting, custom conversions, fixed string interpolation, and simple width-matching out of the box. It is easy to alter its behavior to write new kinds of format string expanders. For most cases, it should be easy to build all sorts of formatters out of the options built into String::Formatter.

Normally, String::Formatter will be used to import a sprintf-like routine referred to as "stringf", but which can be given any name you like. This routine acts like sprintf in that it takes a string and some inputs and returns a new string:

  my $output = stringf "Some %a format %s for you to %u.\n", { ... };

This routine is actually a wrapper around a String::Formatter object created by importing stringf. In the following code, the entire hashref after "stringf" is passed to String::Formatter's constructor (the new method), save for the -as key and any other keys that start with a dash.

  use String::Formatter
    stringf => {
      -as => 'fmt_time',
      codes           => { ... },
      format_hunker   => ...,
      input_processor => ...,
    },
    stringf => {
      -as => 'fmt_date',
      codes           => { ... },
      string_replacer => ...,
      hunk_formatter  => ...,
    },
  ;

As you can see, this will generate two stringf routines, with different behaviors, which are installed with different names. Since the behavior of these routines is based on the format method of a String::Formatter object, the rest of the documentation will describe the way the object behaves.

There's also a named_stringf export, which behaves just like the stringf export, but defaults to the named_replace and require_named_input arguments. There's a method_stringf export, which defaults method_replace and require_single_input. Finally, a indexed_stringf, which defaults to indexed_replaced and require_arrayref_input. For more on these, keep reading, and check out the cookbook.

String::Formatter::Cookbook provides a number of recipes for ways to put String::Formatter to use.

PERL VERSION

This library should run on perls released even a long time ago. It should work on any version of perl released in the last five years.

Although it may work on older versions of perl, no guarantee is made that the minimum required version will not be increased. The version may be increased for any reason, and there is no promise that patches will be accepted to lower the minimum required perl.

METHODS

new

  my $formatter = String::Formatter->new({
    codes => { ... },
    format_hunker   => ...,
    input_processor => ...,
    string_replacer => ...,
    hunk_formatter  => ...,
  });

This returns a new formatter. The codes argument contains the formatting codes for the formatter in the form:

  codes => {
    s => 'fixed string',
    S => 'different string',
    c => sub { ... },
  }

Code values (or "conversions") should either be strings or coderefs. This hashref can be accessed later with the codes method.

The other four arguments change how the formatting occurs. Formatting happens in five phases:

  1. format_hunker - format string is broken down into fixed and %-code hunks

  2. input_processor - the other inputs are validated and processed

  3. string_replacer - replacement strings are generated by using conversions

  4. hunk_formatter - replacement strings in hunks are formatted

  5. all hunks, now strings, are recombined; this phase is just join

The defaults are found by calling default_WHATEVER for each helper that isn't given. Values must be either strings (which are interpreted as method names) or coderefs. The semantics for each method are described in the methods' sections, below.

format

  my $result = $formatter->format( $format_string, @input );

  print $formatter->format("My %h is full of %e.\n", 'hovercraft', 'eels');

This does the actual formatting, calling the methods described above, under "new" and returning the result.

format_hunker

Format hunkers are passed strings and return arrayrefs containing strings (for fixed content) and hashrefs (for formatting code sections).

The hashref hunks should contain at least two entries: conversion for the conversion code (the s, d, or u in %s, %d, or %u); and literal for the complete original text of the hunk. For example, a bare minimum hunker should turn the following:

  I would like to buy %d %s today.

...into...

  [
    'I would like to buy ',
    { conversion => 'd', literal => '%d' },
    ' ',
    { conversion => 's', literal => '%d' },
    ' today.',
  ]

Another common entry is argument. In the format strings expected by hunk_simply, for example, these are free strings inside of curly braces. These are used extensively other existing helpers for things liked accessing named arguments or providing method names.

hunk_simply

This is the default format hunker. It implements the format string semantics described above.

This hunker will produce argument and conversion and literal. Its other entries are not yet well-defined for public consumption.

input_processor

The input processor is responsible for inspecting the post-format-string arguments, validating them, and returning them in a possibly-transformed form. The processor is passed an arrayref containing the arguments and should return a scalar value to be used as the input going forward.

return_input

This input processor, the default, simply returns the input it was given with no validation or transformation.

require_named_input

This input processor will raise an exception unless there is exactly one post-format-string argument to the format call, and unless that argument is a hashref. It will also replace the arrayref with the given hashref so subsequent phases of the format can avoid lots of needless array dereferencing.

require_arrayref_input

This input processor will raise an exception unless there is exactly one post-format-string argument to the format call, and unless that argument is a arrayref. It will also replace the input with that single arrayref it found so subsequent phases of the format can avoid lots of needless array dereferencing.

require_single_input

This input processor will raise an exception if more than one input is given. After input processing, the single element in the input will be used as the input itself.

forbid_input

This input processor will raise an exception if any input is given. In other words, formatters with this input processor accept format strings and nothing else.

string_replacer

The string_replacer phase is responsible for adding a replacement entry to format code hunks. This should be a string-value entry that will be formatted and concatenated into the output string. String replacers can also replace the whole hunk with a string to avoid any subsequent formatting.

positional_replace

This replacer matches inputs to the hunk's position in the format string. This is the default replacer, used in the synopsis, above, which should make its behavior clear. At present, fixed-string conversions do not affect the position of arg matched, meaning that given the following:

  my $formatter = String::Formatter->new({
    codes => {
      f => 'fixed string',
      s => sub { ... },
    }
  });

  $formatter->format("%s %f %s", 1, 2);

The subroutine is called twice, once for the input 1 and once for the input 2. This behavior may change after some more experimental use.

named_replace

This replacer should be used with the require_named_input input processor. It expects the input to be a hashref and it finds values to be interpolated by looking in the hashref for the brace-enclosed name on each format code. Here's an example use:

  $formatter->format("This was the %{adj}s day in %{num}d weeks.", {
    adj => 'best',
    num => 6,
  });

indexed_replace

This replacer should be used with the require_arrayref_input input processor. It expects the input to be an arrayref and it finds values to be interpolated by looking in the arrayref for the brace-enclosed index on each format code. Here's an example use:

  $formatter->format("This was the %{1}s day in %{0}d weeks.", [ 6, 'best' ]);

method_replace

This string replacer method expects the input to be a single value on which methods can be called. If a value was given in braces to the format code, it is passed as an argument.

keyed_replace

This string replacer method expects the input to be a single hashref. Coderef code values are used as callbacks, but strings are used as hash keys. If a value was given in braces to the format code, it is ignored.

For example if the codes contain i => 'ident' then %i in the format string will be replaced with $input->{ident} in the output.

hunk_formatter

The hunk_formatter processes each the hashref hunks left after string replacement and returns a string. When it is called, it is passed a hunk hashref and must return a string.

format_simply

This is the default hunk formatter. It deals with minimum and maximum width cues as well as left and right alignment. Beyond that, it does no formatting of the replacement string.

FORMAT STRINGS

Format strings are generally assumed to look like Perl's sprintf's format strings:

  There's a bunch of normal strings and then %s format %1.4c with %% signs.

The exact semantics of the format codes are not totally settled yet -- and they can be replaced on a per-formatter basis. Right now, they're mostly a subset of Perl's astonishingly large and complex system. That subset looks like this:

  %    - a percent sign to begin the format
  ...  - (optional) various modifiers to the format like "-5" or "#" or "2$"
  {..} - (optional) a string inside braces
  s    - a short string (usually one character) identifying the conversion

Not all format modifiers found in Perl's sprintf are yet supported. Currently the only format modifiers must match:

    (-)?          # left-align, rather than right
    (\d*)?        # (optional) minimum field width
    (?:\.(\d*))?  # (optional) maximum field width

Some additional format semantics may be added, but probably nothing exotic. Even things like 2$ and * are probably not going to appear in String::Formatter's default behavior.

Another subtle difference, introduced intentionally, is in the handling of %%. With the default String::Formatter behavior, string %% is not interpreted as a formatting code. This is different from the behavior of Perl's sprintf, which interprets it as a special formatting character that doesn't consume input and always acts like the fixed string %. The upshot of this is:

  sprintf "%%";   # ==> returns "%"
  stringf "%%";   # ==> returns "%%"

  sprintf "%10%"; # ==> returns "         %"
  stringf "%10%"; # ==> dies: unknown format code %

HISTORY

String::Formatter is based on String::Format, written by Darren Chamberlain. For a history of the code, check the project's source code repository. All bugs should be reported to Ricardo Signes and String::Formatter. Very little of the original code remains.

AUTHORS

  • Ricardo Signes <cpan@semiotic.systems>

  • Darren Chamberlain <darren@cpan.org>

CONTRIBUTORS

  • Darren Chamberlain <dlc@sevenroot.org>

  • David Steinbrunner <dsteinbrunner@pobox.com>

  • dlc <dlc>

  • Ricardo Signes <rjbs@semiotic.systems>

COPYRIGHT AND LICENSE

This software is Copyright (c) 2022 by Ricardo Signes <cpan@semiotic.systems>.

This is free software, licensed under:

  The GNU General Public License, Version 2, June 1991