The Format of PO Files¶
This is a condensed (and rather qooxdoo-slanted) version of the official PO file format description. Not all features described here are tested to work with current qooxdoo.
.po files are made up of entries. The first entry is usually used to maintain meta-data, the other entries contain message id's and their translations.
A simple example:
# A simple .po file msgid "" msgstr "" "Project-Id-Version: 1.0\n" "Report-Msgid-Bugs-To: firstname.lastname@example.org\n" "POT-Creation-Date: 2012-08-28 17:19+0100\n" "PO-Revision-Date: 2012-08-28 17:19+0100\n" "Last-Translator: you <email@example.com>\n" "Language-Team: Team <firstname.lastname@example.org>\n" "MIME-Version: 1.0\n" "Content-Type: text/plain; charset=utf-8\n" "Content-Transfer-Encoding: 8bit\n" msgid "This field is required." msgstr "Este campo es obligatorio."
Each entry has the following structure:
# translator-comments #. extracted-comments #: reference... #, flag... #| msgid previous-untranslated-string msgctxt context msgid untranslated-string msgstr translated-string
# an entry with plural forms #, range: 0..10 msgid untranslated-string-singular msgid_plural untranslated-string-plural msgstr translated-string-case-0 ... msgstr[N] translated-string-case-n
White-space and all comments are optional. Dangling white space or comments (after the last entry) should be avoided. Explanations of the various lines:
'# ' : (# and at least one white space). Comments like this are usually maintained by the person editing the po file, e.g the translator.
'#.': These are comments provided by developers in source code, usually aimed at the translator, and are automatically extracted by tools into the po file. (All comments starting with # and some non-white character are considered "automatic comments" which are in some way manipulated by tools).
'#:': Here appear comma-separated occurrences of this msgid in source code, e.g. like #: src/msgcmp.c:338 src/po-lex.c:699; extracted by tools.
'#,': flag... is a comma-separated list of a number of pre-defined keywords, like fuzzy, range, c-format, no-python-format, .... These flags can both be provided by humans as well as tools, and inform further tool processing (e.g. when checking validity of msgid and translations).
'#|' : Comments starting with this indicator keep previous values of this entry, e.g. "#| msgid previous-untranslated-entry" or "#| msgctxt previous-context". Only provided by tools under certain circumstances (e.g. fuzzy matching).
'msgid' : The message id, the key for this entry. msgid and one of the msgstr forms are the only mandatory lines for each entry. msgid's are extracted from source files by tools.
For the untranslated-string the following constraints apply (the same holds for the translated-strings, e.g. with msgstr):
The string follows C syntax. It has to start on the same line as the msgid keyword.
It is enclosed by double quotes.
It allows embedded escape ("control") sequences: \n, \t, \\, \"
The string can optionally be broken across multiple lines. These parts are concatenated without space to make up the actual value. The line with msgid must at least contain an empty part. This is independent of embedded \n escapes. Example:
msgid "" "This is the start " "and this is the end."
'msgid_plural' : The plural form of the message id.
'msgstr' : The translation of the message id.
'msgstr[<num>]' : Translations depending on singular/plural forms (some languages differentiate if the plural is actually exactly 2, 3, or more concrete numbers of the involved entities; e.g. the translation for "two houses" might differ from the translation of "three houses"). msgstr represents the singular translation. The source code needs to provide the desired number to pick the right translation at runtime. For these kinds of plural forms the "#, range: min..max" flag can be used to instruct the translator how big N can be.
'msgctxt' : A context word for the msgid, e.g. a module/package/component name. This is used to disambiguate entries with the same msgid within a single .po file. An empty string for context is different from a missing msgctxt line. Has to be provided by the developer of the source code. (This entry introduces extra complexity, so try to avoid).