NAME

Template::Plugin::Gettext - Gettext Support For the Template Toolkit Version 2

SYNOPSIS

Load the plug-in in templates:

    [% USE Gettext('com.textdomain.my', 'fr', 'utf-8', 'DIRECTORIES'...) %]
    [% Gettext.gettext('Hello, world!') %]

    [% 'Hello, world!' | gettext %]

Or alias "Gettext":

    [% USE gtx = Gettext('com.textdomain.my', 'fr', 'utf-8', 'DIRECTORIES'...) %]
    [% gtx.gettext('Hello, world!') %]

Use method invocations:

    [% Gettext.gettext("Hello, world!") %]
    [% Gettext.xgettext("Hello, {name}!", name => 'John Doe') %]

Or filters (without the prefix):

    [% FILTER gettext %]
    Hello, world!
    [% END %]

    [% 'Hello, world!' | gettext %]

    [% FILTER xgettext(name => 'John Doe') %]
    Hello, {name}!
    [% END %]

You have a multitude of methods available:

    [% gtx.gettext("Hello, user!") %]
    [% gtx.xgettext("Hello, {user}!", user => 'John Doe') %]
    [% gtx.ngettext("One document deleted.",
                    "Multiple documents deleted."),
                    42) %]
    [% gtx.nxgettext("One document deleted.",
                     "{num} documents deleted."),
                     42,
                     num => 42) %]
    [% gtx.npgettext("context..."
                     "One document deleted.",
                     "Multiple documents deleted."),
                     42) %]
    [% gtx.npxgettext("context...",
                      "One document deleted.",
                      "{num} documents deleted."),
                      42,
                      num => 42) %]

DESCRIPTION

The Template::Plugin::Gettext plug-in makes the GNU gettext API available for documents using the Template Toolkit version 2. See https://github.com/gflohr/Template-Plugin-Gettext for an overall picture and the recommended tool-chain.

FUNCTIONS

The following methods produce translatable content:

[% gtx.gettext(STRING) %]

Retrieves the translation for STRING.

[% gtx.xgettext(STRING, PLACEHOLDER1 => VALUE1, PLACEHOLDER2 => VALUE2, ...) %]

Gets the translation for a string with placeholders and interpolates values into it. Placeholders have the format <{PLACEHOLDER}>. For a literal left curly brace you can use this hack:

    [% gtx.xgettext("String with {LBRACE}PLACEHOLDERS{RBRACE}",
                    LBRACE => "{", RBRACE => "}") %]
[% gtx.pgettext(CONTEXT, STRING) %]

Retrieves the translation for STRING in context CONTEXT. You should use message context to disambiguate identical strings that require different translations depending on the context. See this example for an explanation:

    [% gtx.gettext("State: ") %]
        [% gtx.gettext("Open")] | [% gtx.gettext("Close") %]
    
    [% gtx.gettext("Menu:") %]
        [% gtx.pgettext("menu", "Open")]
        [% gtx.gettext("Save")]
        [% gtx.gettext("Save As")]
        [% gtx.pgettext("menu", "Close")]

The strings "Open" and "Close" in line 2 are adjectives. As menu entries they are verb forms and will have a different translation in many languages.

 In doubt: Only use contexts if one of your translators complains
 about a message having multiple meanings.
[% gtx.pxgettext(CONTEXT, STRING, PLACEHOLDER1 => VALUE1, PLACEHOLDER2 => VALUE2) %]

Get the translation for STRING with placeholders in context CONTEXT. This is a mixture of xgettext() and pgettext() above.

[% gtx.nxgettext(SINGULAR, PLURAL, COUNT, PLACEHOLDER1 => VALUE1, PLACEHOLDER2 => VALUE2) %]

Retrieves the translation for the string with the singular form SINGULAR and the plural from PLURAL, both possibly containing placeholders. The correct form is picked based upon the third argument COUNT.

Example:

    [% gtx.nxgettext("One document deleted",
                     "{num} documents deleted"),
                     count,
                     num => count) %]

In English this will produce "42 documents deleted" if the variable count has the value 42. It will produce "One document deleted" if the variable count has the value 1.

In other languages, the rules for plural forms may be a lot simpler (for example Chinese, which has no plural) or a lot more complicated (for example Russian with two or Slovenian with even 3 plural forms). Using ngettext() gives your translators the chance to provide syntactically correct translations for these cases.

[% gtx.npxgettext(CONTEXT, SINGULAR, PLURAL, COUNT, PLACEHOLDER1 => VALUE1, PLACEHOLDER2 => VALUE2) %]

Putting it all together: For message context CONTEXT the translation for a message in SINGULAR and PLURAL is retrieved based on the argument COUNT. Possible placeholders are expanded.

The function is a mixture of xgettext(), ngettext(), and pgettext(), see above!

[% gtx.ngettext(SINGULAR, PLURAL, COUNT) %]

Useless function, provided for completeness. Use nxgettext() instead, so that you can interpolate the value of COUNT!

[% gtx.npgettext(CONTEXT, SINGULAR, PLURAL, COUNT) %]

Useless function, provided for completeness. Use npxgettext() instead, so that you can interpolate the value of COUNT!

In fact, you have also all the keywords used for "FILTERS" available but those not listed here have such an odd ordering of arguments that they are not listed here.

FILTERS

The entire gettext API is also exposed as a filter. There are two things to note here:

  • When used as filters, you don't prefix the method names. It' s "gettext" not "Gettext.gettext" or "gtx.gettext()".

  • The filters with message contexts have rather strange names, for example:

        [% FILTER gettextp("greeting") %]
        Hello, world!
        [% END %]
    
        or 100 % equivalent:
    
        [% 'Hello, world!' | gettextp("greeting") %]

    Why? The text between FILTER and END resp. the text in front of the pipe symbol | is always the first argument. This plug-in therefore tries to make the first argument the most significant one. Nobody stops you from writing the following:

        [% FILTER pgettext("Hello, world!") %]
        greeting
        [% END %]
    
        or again 100 % equivalent:
    
        [% 'greeting' | pgettext("Hello, world!") %]

    It produces exactly the same results as above but it looks a little bit odd, doesn't it?

    It would have been arguably better understandable to silently reorder the arguments, when using the plug-in as a filter. But it would break extraction of your strings with xgettext-tt2 (Locale::XGettext::TT2) because the string extractor would then confuse the arguments.

    But stay relaxed! Message contexts are rarely needed, and when you need them, you have to live with this little weirdness.

    In order to avoid confusion, those filters that would not have the translatable string (in the singular form) where one would expect it, are not documented here.

You can use the following filters:

[% STRING | gettext %]
[% FILTER gettext %]STRING[% FILTER %]

Retrieves the translation for STRING.

[% STRING | xgettext(PLACEHOLDER1 => VALUE1, PLACEHOLDER2 => VALUE2, ...) %]
[% FILTER xgettext(PLACEHOLDER1 => VALUE1, PLACEHOLDER2 => VALUE2, ...) %]STRING[% END %]

Gets the translation for a string with placeholders and interpolates values into it. Placeholders have the format <{PLACEHOLDER}>. For a literal left curly brace you can use this hack:

    [% "String with {LBRACE}PLACEHOLDERS{RBRACE}" | xgettext(LBRACE => "{", RBRACE => "}") %]
[% STRING | gettextp(CONTEXT) %]
[% FILTER gettextp(CONTEXT) %]STRING[% END %]

Retrieves the translation for STRING in context CONTEXT. You should use message context to disambiguate identical strings that require different translations depending on the context. See the docuemntation for pgettext() in "FUNCTIONS" above for more details!

[% STRING | xgettextp(CONTEXT, STRING, PLACEHOLDER1 => VALUE1, PLACEHOLDER2 => VALUE2) %]
[% FILTER xgettextp(CONTEXT, STRING, PLACEHOLDER1 => VALUE1, PLACEHOLDER2 => VALUE2) %]STRING[% END %]

Get the translation for STRING with placeholders in context CONTEXT. This is a mixture of xgettext() and pgettext() above.

[% SINGULAR | nxgettext(PLURAL, COUNT, PLACEHOLDER1 => VALUE1, PLACEHOLDER2 => VALUE2) %]
[% FILTER nxgettext(PLURAL, COUNT, PLACEHOLDER1 => VALUE1, PLACEHOLDER2 => VALUE2) %]SINGULAR[% END %]

Retrieves the translation for the string with the singular form SINGULAR and the plural from PLURAL, both possibly containing placeholders. The correct form is picked based upon the third argument COUNT.

Example:

    [% "One document deleted" | nxgettext("{num} documents deleted"),
                                          count,
                                          num => count) %]

In English this will produce "42 documents deleted" if the variable count has the value 42. It will produce "One document deleted" if the variable count has the value 1.

In other languages, the rules for plural forms may be a lot simpler (for example Chinese, which has no plural) or a lot more complicated (for example Russian with two or Slovenian with even 3 plural forms). Using ngettext() gives your translators the chance to provide syntactically correct translations for these cases.

[% SINGULAR | nxgettextp(PLURAL, COUNT, CONTEXT, PLACEHOLDER1 => VALUE1, PLACEHOLDER2 => VALUE2) %]
[% FILTER nxgettextp(PLURAL, COUNT, CONTEXT, PLACEHOLDER1 => VALUE1, PLACEHOLDER2 => VALUE2) %]SINGULAR[% END %]

Putting it all together: For message context CONTEXT the translation for a message in SINGULAR and PLURAL is retrieved based on the argument COUNT. Possible placeholders are expanded.

The filter is a mixture of xgettext(), ngettext(), and pgettext(), see above!

[% SINGULAR | ngettext(PLURAL, COUNT) %]
[% FILTER ngettext(PLURAL, COUNT) %]SINGULAR[% END %]

Useless filter, provided for completeness. Use nxgettext() instead, so that you can interpolate the value of COUNT!

[% SINGULAR | ngettextp[(PLURAL, COUNT, CONTEXT) %]
[% FILTER | ngettextp[(PLURAL, COUNT, CONTEXT) %]SINGULAR[% END %]

Useless filter, provided for completeness. Use nxgettextp() instead, so that you can interpolate the value of COUNT!

[% debug_locale %]

The plug-in implicitely calls web_set_locale() (see Locale::Util) if a language was specified in the USE statement. The function debug_locale() gives you the return value of the call for debugging purposes. Example:

    [% USE gtx = Gettext('com.mydomain.www', de') %]
    
    Using locale [% debug_locale() %].

This way, you can determine whether setting the specified locale actually worked.

CLASS METHODS

textdomains

Returns a hash where the keys are the textdomains found in templates that invoked the plug-in, and the values are the corresponding template file names.

The purpose of this method is to allow harvesting files that should be processed by xgettext-tt2.

If the template is either "input text" or "input file handle", the template variable gettext_filename - if existing - is assumed as the template name. Rationale: "input text" and "input file handle" are used by the Template Toolkit as aliases, when reading from a scalar or a file handle.

resetTextdomains

Resets the hash described above for textdomain() to its initial, empty state.

COPYRIGHT

Copyright (C) 2016-2018 Guido Flohr (http://www.guido-flohr.net/). License LGPLv3+: GNU General Public License version 3 or later. This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Copyright (C) 2016-2018 Guido Flohr <guido.flohr@cantanea.com>, all rights reserved.

SEE ALSO

Template, Template::Manual::Filters, xgettext-tt2, perl