Difference between revisions of "Translation"

From GnuCash
Jump to: navigation, search
m (Special characters and other tips: typo)
(Disambiguation Prefix: enhance sample:)
Line 670: Line 670:
 
msgstr "Reconciled:A"
 
msgstr "Reconciled:A"
 
</SyntaxHighlight>
 
</SyntaxHighlight>
* '''Sample text''' to determinate the size of the output cell. Here is no translation required, but the "worst case" of expected text, e.g. to make german business users happy:
+
* '''Sample text''' to determinate the size of the output cell. Here is no translation required, but the "worst case" of expected text. For the following example the german tranlators copied the longest path of account names from their business account templates as shown in the accounts tab of Gnucash:
 
<SyntaxHighlight lang="po" highlight="2">
 
<SyntaxHighlight lang="po" highlight="2">
 
msgid "sample:Expenses:Automobile:Gasoline"
 
msgid "sample:Expenses:Automobile:Gasoline"

Revision as of 18:16, 20 February 2019

HOWTO: Translating GnuCash

The concept of this document is to give you step-by-step instructions on how to update (or create if non-existant) language translations and other localization tasks for the gnucash project. See Translation Status for the current status of the project with respect to translation contributions.

Contents

Overview

GnuCash has several separate areas that need translations or localization, which are by priority:

  • A glossary file: A reference in form of message catalogs with roughly 200 terms that are commonly used throughout GnuCash. Preferably it should be translated first for each new language to define a consistent terminology for the other parts.
  • The program's message catalogs: These contain all text messages for the application's user interface. It currently consists of roughly 5000 strings. Not all are equally important though.
  • A Windows Installer with about 20 strings.
  • The account templates: A set of ready to use account chart snippets for personal users that can be mixed and matched into a full chart of accounts during the creation of a new a new GnuCash file with currently 115 translatable account names.
  • If your government or other authorities offer specific templates for business users, you can create them, too.
  • The documentation: The Help Manual and the Tutorial and Concepts Guide. These documents explain how to use GnuCash. They are written in DocBook, an XML variant.
  • The website, while mostly written in PHP/HTML, uses message catalogs, too. Currently it has 430 messages.
  • Income Tax Tables require some knowledge of your regional tax rules and are related to account templates.
  • Optionally you can translate doc/README-* files, too.

GnuCash uses the translation of the ISO 4217 currency codes and names. This are maintained by the ISO Codes Team. If your language is missing or has issues contact them.

Available resources

There are many resources to support you at gnucash.org and other places. Don't hesitate to use them.

gnucash.org

We have collected a bunch of useful information for you here in this wiki -

navigate from the main page or use the search function -

and on the website http://www.gnucash.org.

And we have several channels of communication and a few other useful tools:

IRC

The fastest way to get help is using the internet relay chat IRC, as many of the developers hang out there and are eager to help.

Mailing lists

Translators will probably find two or three gnucash mailing lists of interest. If you can not find the answers to your questions in their Mailing_Lists#Mailing List Archives, feel free to ask them.

  • For language and region specific questions, where you should discuss terms and ask for proofreading of your work, use your languages mailing list. Currently we have
    gnucash-[pt_]br, de, es, fr, it, nl.
If none exists for your language or nobody has an answer there, use the english lists:
  • General use questions and answers are found on the gnucash-users mailing list,
  • specific development questions go to the gnucash-devel list.
  • If you dislike the heavy traffic on the above lists, you should at least subscribe the gnucash-announce list to get informed about new releases.

To subscribe or view archives of these lists follow the links to the Mailing Lists.

translationproject.org

The Translation Project coordinates the message cataloges of several programs; you can see which languages they host on their GnuCash page and learn about their process here.

If you want to help with an existing Translation Project translation or add a new translation under their aegis, please contact the language team using the information on their team page.

Coordinating Message Catalog Translations

The Translation Project requires that you use either an existing translation or gnucash.pot from the latest release as the basis for your translation.

Coordinating Account Templates and Documentation

The Translation Project handles only message catalogs, so if you have glossaries, account templates or documentation translations you'll submit them directly to us.

Direct Contribution

If you want to work on a translation that's not coordinated by the Translation Project or a new translation and you want to submit it directly to GnuCash, read on. If it's an existing translation please try to contact the last translator first using the contact information in the message catalog.

If you're starting a new translation it's much easier to begin with a tarball, which will include po/gnucash.pot; if you start from a Git clone you'll have to build Gnucash yourself first to generate gnucash.pot.

On Other Sites

General Translations Guidelines and Tips

General information about how to approach the translation of software can be found here:

Other Web Based Translation Tools

Several services for translation coordination exist:

Pootle
Weblate
Zanata
and several more

As we have no experience with them contact the gnucash-devel mailing list before using them.

Other Useful Sites

  • There are many online dictionaries, use a search engine like google, to find the best fitting.
  • An index of on-line dictionaries and a lot of other resources can be found at www.yourdictionary.com.
  • Comparing articles on wikipedia in your language and english can be helpful.
  • In special cases the terminology database of the European Union Inter Active Terminology for Europe can be used, too.

How to submit changes directly to GnuCash

Once you've got a translation ready for submission, there are three options:

Pull Requests at GitHub

If you already are a github user you can publish your work as a separate branch on your gnucash[-{[ht]docs|on-windows}] fork and send a pull request to the GnuCash team. If you wish to continue your work before the request was completed, simlply create a new working branch. After the pull request was completed you can delete the related branch again. See Git for details.

Patches at Bugzilla

The easiest is to use Bugzilla:

containing the diff between the old files - if any - and your new files:
  • If you have Git installed, the easiest way to create the patch is git format-patch origin/maint..maint.
  • Else you can use diffutils: diff -u[r] <original path> <modified path> > <name>.patch. See info diff for details.

Mailing lists

You can simply email your completed message catalog (.po file) to the developers at the gnucash-devel mailing list. We'd really rather you use Bugzilla or Github because those are much less likely to get lost or forgotten, but if you really find those too hard we'll try to accommodate you on the list.

Translating the Program

To begin a translation you need either the message template file (gnucash.pot) if you're starting a new translation or the existing message catalog (xx.po or xx_YY.po, where xx is the ISO code for your language and YY the ISO code for a language-variant locale, see Naming Convention). These files are in the gnucash sources, which you can obtain either from our git repository or by downloading the latest release tarball from Sourceforge.

Get the source from Git

GnuCash uses Git as a version control system. Click here for an introduction to Git.

The first thing to do is usually, to download the latest STABLE branch, called maint, of gnucash from git and get it to compile. See Translation Status to choose which branch to use for translations. Do not use the master branch unless specifically mentioned in Translation Status. Because the text in the master branch changes so much it would be a waste of time to translate it. Do not worry, when the master branch becomes stable the existing translations in the STABLE branch will be merged. Your work will not be lost.

Normally checkout the current stable branch:

git clone https://github.com/Gnucash/gnucash ${SOURCEDIR}
cd ${SOURCEDIR}
git checkout maint

The argument ${SOURCEDIR} above can be whatever you want your local directory to be called, and is optional in the first line. If you leave it out, you'll have a directory called gnucash created containing all the source code below your current working dir.

Checkout the documentation (optional, but recommended):

git clone https://github.com/Gnucash/gnucash-docs gnucash-docs
cd gnucash-docs
git checkout maint

Update your repository

After this initial git checkout, you can later update your local repository using

git pull --rebase

from anywhere in the tree of ${SOURCEDIR}. We recommend to do it daily when you start your work.

Get packages, which are used while building

  • Dependencies contains the list of packages, which are used while building GnuCash. If some are missing the configuration by cmake or the build by make or ninja will fail.
  • For working on .po files we use several commands starting with msg. This are part of gettext-tools.

Under Linux use your package manager and install them.

Steps in the Build System

GnuCash uses the cmake build generator to control the build of all components. Details are described in Building. The following short-form instructions work on Unix systems and assume that you have already set up the dependencies according to the FAQ. If you are building the unstable or master branches you'll need to use your package manager to install two more development packages, boost-all-devel and googletest.

Configuration

It's generally preferred to use a separate build directory. For translators it's convenient to make this a hidden directory in the source directory as that provides the least complicated paths in the translation template; hiding the directory prevents the extraction tool from including build products in the files searched for translatable strings. We'll call the directory .build. You can call it anything you like, just remember to start the name with a dot so that it's hidden.

Change directories to your build directory and run cmake to configure the build. In most cases translators will not need to install GnuCash and it will build any arguments:
cd ${BUILDDIR} # e.g. .build
cmake ..

tells cmake to configure the current directory using the CMakeLists.txt in the parent directory. If configuration fails or if you want to refine your build consult Building.

Optional Compile

It is suggested that you compile the gnucash source code. It is a good idea to actually run gnucash with your new translations because it is quite helpful to see the phrases in the context of the running program.

make
Note
Advanced users might consider to use Ninja instead of Make, but that is behind the scope of this page.

Next, compile gnucash in the build directory:

cd ${BUILDDIR} # e.g. .build
make

GnuCash can be run from the build directory as follows:

cd ${BUILDDIR} # e.g. .build
GNC_UNINSTALLED=1 GNC_BUILDDIR=`pwd` bin/gnucash

It is a good idea to use absolute paths like this to insure you run the proper gnucash executable. To run your distributions pre-installed version of gnucash, under Linux usually you can type:

/usr/bin/gnucash

In either case, you can easily switch between the various languages the gnucash has available by placing the LANG env var before the call to the executable. You can find details in Locale Settings.

Documentation

The GnuCash Help Manual and the Tutorial and Concepts Guide use a different build system, Autotools. To make a browsable documentation set,

  1. git clone gnucash-docs or unpack a gnucash-docs tarball,
  2. create a hidden build directory just as you did for GnuCash itself, and
  3. change directory to the build directory, and run the following commands:
    ../autogen.sh
    ../configure
    make html
    

The new documentation home pages will be

guide/<locale>/gnucash-guide/index.html and
help/<locale>/gnucash-help/help.html,

where <locale> is one of C, de, it, ja, or pt (English, German, Italian, Japanese, or Portuguese respectively). You can use the file: URL scheme to point your browser to one of them; the links will work from there.

Naming Convention

The language code is of the following form:

<language>[_<REGION>][.<Charset>][@modifier]
using the well known ISO codes 639-1 for languages and 3166-1 for regions, which are in most cases countries .
  • Only if no 2 letter code exists for your language in ISO_639-1, use the 3 letter code from ISO_639-2.
  • If you are the first translator for your language, you should name your files and directories only with your language code. So all people using your language despite of their region will benefit from your work.
  • If there are parts, which are specific for your region, e.g. business account templates respecting local law, then they should be in a <language>_<REGION> directory.
  • The common charset in gnucash is utf8.
  • In the rare case different scripts are used for the same language like kyrilic and latin add for the less used the modifier e.g. sr@latin.

In the following parts of this document 'XXXX' and 'LL' refer to your language code.

Contact the maintainer of your language

To find out who is the last person to work on your language, look near the top of the po/XXXX.po file which corresponds to your language. If your language does not have a .po file available, see the next section.

The beginning of your .po file should look something similar to this:

# Localization for Portuguese-Brazil
# Copyright (C) 2003 Free Software Foundation, Inc.
# Jon Lapham <lapham@extracta.com.br>, 2003
# Jose Carlos Nascimento - <joseca@psabs.com>, 2001.
# 
msgid ""
msgstr ""
"Project-Id-Version: GnuCash 1.8.3\n"
"Report-Msgid-Bugs-To: https://bugs.gnucash.org/enter_bug.cgi?"
"product=GnuCash&component=Translations\n"
"POT-Creation-Date: 2003-05-16 16:42-0300\n"
"PO-Revision-Date: 2003-06-02 12:00-0300\n"
"Last-Translator: Jon Lapham <lapham@extracta.com.br>\n"
"Language-Team: NONE \n"
"Language: pt_BR\n"
"MIME-Version: 1.0\n"
"Content-Type: text/plain; charset=UTF-8\n"
"Content-Transfer-Encoding: 8bit\n"
"Plural-Forms: nplurals=2; plural=n != 1;\n"

Look to see who the "Last-translator" was, and send an email to that person and ask what you can do to help. This is important because if there already is an active maintainer of the translation file, you should interact directly with him or her. If there is no Last-Translator, or that person is not maintaining the file actively (and tells you to take over or does not answer in reasonable time), you will become the maintainer and you should change the "Last-Translator" to your name and email address.

The glossary file

Inside the po/glossary/ directory should be a "glossary" file for your language. This file contains a bunch of commonly used terms found in GnuCash - and the explanation of a few very specific for GnuCash like the disambiguation between bill, invoice and voucher. It is recommended that you get this file translated first, and use it as a guide when translating the real .po file or the documentation. Please keep in mind that this file will never be visible to a user! The glossary file is only a tool for you, the translator! I repeat: The strings from the glossary file will never be user-visible, they are only used by you, the translator.

Go into the glossary directory and rebuild your language's glossary file:

cd po/glossary/
./txt-to-pot.sh gnc-glossary.txt > gnc-glossary.pot
  • If your .po glossary file does not exist, use this gnc-glossary.pot file to
create it either with
msginit # add -l<locale> if different from your settings
or
cp gnc-glossary.pot XXXX.po
  • If your .po glossary file does exist, use the msgmerge program to update it:
/usr/bin/msgmerge -o XXXX.po XXXX.po gnc-glossary.pot;

Now, open your language's glossary file and translate it completely.

Terms missing or inadequate in the glossary file

If you detect an important term which is missing or could be better explained, open a bug and add a patch for the source file gnc-glossary.txt. The - with the exception of the headline - alphabetical sorted lines of this file have the form:

Msgid<TAB>Comment
Reminder
Does your editor enter <TAB>s or will it replace them with <Space>s?

To test it, run the following steps after you changed gnc-glossary.txt:

  • ./txt-to-pot.sh gnc-glossary.txt > gnc-glossary.pot will generate a new gnc-glossary.pot,
  • run above msgmerge command for your XXXX.po and check it for the new msg,
  • if successful, update all *.po files with:
find *.po -exec /usr/bin/msgmerge -o '{}' '{}' gnc-glossary.pot \;
  • check and translate the new string in your new glossary file.

Get a fresh template

The Portable Object Template (.pot file) is a collection of all translatable strings.

It is important, to repeat this step e.g. after you asked a developer to change an insufficient string or you got an announcement about changed strings.

  • If your repository is no fresh checkout, you should first #Update your repository:
    git pull --rebase
    
  • Then in gnucash, a specific command at the top-level of the build tree (or source tree, if you do not use a separate build directory) will perform all necessary steps for this:
    cd ${BUILDDIR} # adjust ${BUILDDIR}
    make pot
    
Note
If make complains
make: *** No rule to make target `<path/to/missing-source-file>', needed by `gnucash.pot'.  Stop.
, you should run
make clean
and then retry "make pot". That can happen because we sometimes remove obsolete files.
Now go in the po directory:
cd ${SOURCEDIR}/po # adjust ${SOURCEDIR}

If your language file already exists, continue with #Updating an existing .po file.

Build a new .po file

If there is no .po file for your language, then you can start a new one. Start by

msginit [-lLL[_RR]] [-i<path/to/>gnucash.pot]

where LL is your language and RR your region code, if there is already a different LL.po like e.g. pt and pt_BR.

-i<path/to/>gnucash.pot is required, if you use a separate build dir,
-l... is only required, if your target is different from your environments current language.

This will initialize the meta information with values from your user environment.

Note
msginit 0.19.3 is querying an obsolete address of the translation project for language teams, but that doesn't matter.

If that does not work you can copy the file gnucash.pot into a work file named LL.po and just edit this file.

Adjust the header and special strings

The top of the .po file should be edited somewhat. The comments at the top of the file should be changed to be current:

# Messages in Deutsch fuer GnuCash
# Copyright (C) 1999 Free Software Foundation, Inc.
# Jan-Uwe Finck <Jan-Uwe.Finck@bigfoot.de>, 1999.

Expand the copyright comment by a list of <Translator name> <email address>, year[ range] in reverse historical order. Then you can later copy the list in the translator_credit.

Make sure that the header of your .po file contains an adjusted form, e.g. for slavian languages set nplurals=3, of this line:

"Plural-Forms: nplurals=2; plural=n != 1;\n"
See Gettext Manual: Translating Plural Forms for details.


...and that you change the "Last-translator" string to your name.

The first lines as shown in the previous section will contain several capitalized entries. Replace all the words in capitals with something appropriate. In this case, you will be the first author of the translation, and also the

  • Last-Translator of it.
  • PO-Revision-Date: Editors with a specific po mode should update it on saving. If you use a normal editor, you will have to do it manually.
  • CHARSET should be UTF-8 today, and
  • ENCODING then is 8bit.

See the full explanation in Gettext Manual: Header Entry.

  • Remove the `#, fuzzy' line once you have specified the items in capitals, because once this is done the header entry is no longer fuzzy.

There is currently one special string:

#: ../src/gnome-utils/gnc-main-window.c:4389
msgid "translator_credits"
msgstr "Christian Stimming <christian@cstimming.de> et. al."

This will appear in Help->About->Credits->Translation. So you should enter your name or that of your team and an email address where users can contact you for typos and gifts.

Tip
After some time more persons would have worked on the translation. Then you can expand it from your header comments to:
msgstr ""
"<current translator>, <years>\n"
"<previous translator>, <years>\n"
:
"<first translator>, <years>"
But we suggest to remove the email address of individuals for data protection reasons.

Prepopulate your file with translations from goffice

Gnucash has borrowed a couple of source files from goffice. Those files contain a number of translatable strings. The goffice translation teams have already put effort in translating those in many languages. To reduce our translation effort, the script linked in #Borrowing Code can be used to import these translations into our own po files.

If the goffice part for your language is incomplete, you might consider to offer them to update their file with your work.

Adjust po/CMakeLists.txt

Also include your language code into the NEW_LINGUAS variable in the CMakeLists.txt file in the po folder of your source directory:

# Set of available languages:
# * managed at the Translation Project (TP):
set (TP_LINGUAS az ca cs da eu fa ja nl rw sk sr sv tr uk zh_CN)
# * already marked as external at TP:
set (GC_LINGUAS ar bg de el en_GB es fi fr gu he hi hu it kn ko lt lv mr nb ne pl pt pt_BR ro ru ta te ur vi zh_TW)
# * New or unmarked: The release manager should announce them to TP 
# and when listed there move in the respective group above.
set (NEW_LINGUAS as brx doi es_NI kok kok@latin ks mai mni mni@bengali)

set (ALL_LINGUAS ${TP_LINGUAS} ${GC_LINGUAS} ${NEW_LINGUAS})

CMakeLists.txt is a file, which controls the configuration of the build process. So after changing it, you have to rerun cmake. Some IDEs like Eclipse will do it automagical for you.

As part of the build your LL.gmo file is generated in the po directory of your build tree. Finally make install will copy it to the place, where it will be found at runtime. If you forget one of the steps, your translated language will not appear.

Continue with #Translating the .po file.

Updating an existing .po file

Before you begin actual translation work, you should update the gnucash.pot file and use this to update your .po file. This process will insure that you have the latest translatable strings.

If your language file already exists, update it using the msgmerge program. This will move the old translations of unchanged strings in the new file:
msgmerge -o LL.po LL.po gnucash.pot
Note
If you had choosen a separate build directory, e.g. .build, adjust the path in above command:
cd ${SRCDIR}/po # adjust ${SRCDDIR}
msgmerge -o LL.po LL.po ../.build/po/gnucash.pot
Separate commits for "noise" and real work
You should now run git commit or create a patch "Merge recent template in <locale>.po".
This will contain only the "noise" from the updated pot file as usually many line numbers change.
Diff example after msgmerge
 #. Business options
-#: ../src/app-utils/app-utils.scm:303
-#: ../src/business/business-gnome/gncmod-business-gnome.c:117
+#: ../src/app-utils/app-utils.scm:322
+#: ../src/business/business-gnome/gncmod-business-gnome.c:119
 msgid "Business"
 msgstr "Geschäft"
Hiding your real changes in hundreds of such sections will make it really hard for your coworkers to find them.

You should also do it, if your current translation tool uses other format settings like line breaks as the previous tool. In this case just open the file, save it and git commit or create a patch "Preparing patch: reformating <locale>.po".

Example in KBabel format
#: ../src/app-utils/business-prefs.scm:33
msgid "The format string to use for generating customer numbers. This is a printf-style format string."
msgstr "Используемая строка форматирования для создания номеров клиентов. Её формат соответствует printf."
Example in PoEDit format
#: ../src/app-utils/business-prefs.scm:33
msgid ""
"The format string to use for generating customer numbers. This is a printf-"
"style format string."
msgstr ""
"Используемая строка форматирования для создания номеров клиентов. Её формат "
"соответствует printf."
Review the file header
  • Project-Id-Version should be GnuCash 5.5 and
<tt>POT-Creation-Date should be recent. If they are not, you probably forgot #Get a fresh template or #Updating an existing .po file.
  • The PO-Revision-Date should be >= POT-Creation-Date.

Some, but not all tools will do this reliable for you.

Translating the .po file

Finally. You are ready to do some translating!

The .po source files are plain text files.

Tools

Some plain text editors offer a specific syntax highlighting for .po files, but there are also specific tools you can use:

  • Emacs has the po-mode to edit po files.
  • Geany, an editor with sytax highlighting and a little bit more
  • GTranslator is another tool but we recommend not to use it because the version of 2006 doesn't support all of the interesting elements inside the po file. Update this, if you know it is fixed.
  • KBabel was another recommended tool, see Lokalize.
  • Lokalize is the successor of KBabel in KDE4.
  • Poedit ( get it here: http://www.poedit.net/download.php ) to finish the PO file edit and build.
Version 2.1.1 had issues

(feel free to add more tools here)

Gettext source (.po) file format

A record in a po file has the following form:

<empty or only white-space>
#  translator-comments
#. extracted-comments
#: reference…
#, flag…
#| msgid previous-untranslated-string
msgid untranslated-string
msgstr translated-string
  • The empty or only white-space line is the record separator.
  • In translator-comments you can put your own notes.
  • The extracted-comments are notes from the programmers for you.
  • One or more references tell you, where the message appears in the sources.
  • The most important flags #The fuzzy flag and source language format flag will be explained below.
  • The msg* should explain themself above.
Example
Here is an example of translating some text into German:
Before:
#: messages-i18n.c:11
msgid ""
"The GnuCash personal finance manager.\n"
"The GNU way to manage your money!"
msgstr ""

After, the translation in the de.po file:

#: messages-i18n.c:11
msgid ""
"The GnuCash personal finance manager.\n"
"The GNU way to manage your money!"
msgstr ""
"GnuCash: Ihr persoenlicher Finanzmanager.\n"
"Der GNU-Weg, ihr Geld zu verwalten!"

You should read through every translation in the .po file at least once.

The fuzzy flag

If you translate a string that has the phrase "#, fuzzy" in the comments above it, remove the word fuzzy. A fuzzy translation means that there is no translation. The computer took some guess about what the translation might be, but as long as it is marked "#, fuzzy", this guess will simply be ignored and this string will stay untranslated. A string marked as "fuzzy" means it will not be translated in the program.

After you finish translating, you should not have any "#, fuzzy" strings left. Remember, a string marked as "fuzzy" means it will not be translated in the program.

For example:

#: messages-i18n.c:35
#, fuzzy, c-format
msgid ""
"There was an error writing the file\n"
"     %s\n"
"\n"
"%s"
msgstr ""
"Es gab einen Fehler beim Oeffnen der Datei. \n"
"     %s."

Here the msgid was changed from "opening" to "writing". You need to correct the translated string and remove the 'fuzzy' keyword, because only then the translation will actually appear in the program. Remember, a string marked as "fuzzy" means it will not be translated in the program. For example:

#: messages-i18n.c:35
#, c-format
msgid ""
"There was an error writing the file\n"
"     %s\n"
"\n"
"%s"
msgstr ""
"Es gab einen Fehler beim Schreiben der Datei. \n"
"     %s."

Notice that the comment "c-format" was not removed. That is correct, you should leave that.

When you see the comment "c-format", it means that the format codes in the translatable string are referring to C formatting codes. So, '%s' means text, '%d' means an integer, etc...

Translate the strings

Each message to translate is then given in turn in the PO file. For example, an untranslated entry might be:

#: lib/error.c:88
msgid "Unknown system error"
msgstr ""

The empty msgstr string has to be filled with the translation for the string shown after msgid. If you were a German speaker, say, the entry once translated might look like:

#: lib/error.c:88
msgid "Unknown system error"
msgstr "Unbekannter Systemfehler"

You just produce a translation for all entries in the PO file, one after another, respecting the overall file format and the quoting needed for special characters, when needed. Observation and intuition may allow you to grasp what the format should be; the precise rules for PO files are given in the GNU gettext manual. The msgfmt program is helpful for pinpointing formatting errors.

Special characters and other tips

Depending on the context a few characters have a special meaning and need some special treatment:

"_" (underline)
In menu and dialog entries the following character will become the accelerator, mnemonic or hotkey, which can be used together with a superkey [ctrl] or [alt] to jump to the entry. While in GTK2 they have always been visible, in GTK3 they appear only after holding the superkey. More specific under Linux you reach a main menu entry with [alt]+[key] and its submenus and other menu entries with [key]. In dialogs always use [alt]+[key].
The terminology is not unique here: while
msgfmt has a --check-accelerator option and
DocBook uses <accel> tag for a letter used with a meta key and <shortcut> for a key combination, but
GTK+ distinguishes the underline marked character of the label as mnemonic and the hotkeys like F1 for Help as accelerator.
So the key should be unique in its context and you should control it by #Running GnuCash with your file after #Compile & Install.
wrong:
"do _this" # Hotkey: t
"do _that" # Hotkey: t => not directly reachable
right:
"do th_is" # Hotkey: i
"do th_at" # Hotkey: a
That is one of the reasons why you should run the program with your translation: to see duplicate accelerators.
Characters to avoid
Already used on buttons like in English: Close, Help. Others depend on the context.
Characters breaking the baseline like in latin script: j,p,q,y. At least in some fonts the underline becomes invisible - leaving the user clueless.
In the headers of exported files
here it is used to link multiple words to one identifier together. Example:
#: ../src/import-export/csv-exp/csv-tree-export.c:155
msgid "full_name"
msgstr "Vollständige_Bezeichnung"
"\" (backslash)
It is the escape character in many programming languags. The following character has a special meaning like e.g.:
"\n" (New line)
The most often used special character in our strings. If msgid contains "\n" keep the layout and add them to msgstr too - at the same places.
"some \"quoted\" text"
Because " terminates the messages, you must precede it by a backslash inside of the messages or use other quoting characters like:
  • "some 'quoted' text"
  • "some »quoted« text"
  • "some „quoted” text"
You are free to use the conventions which are common or suggested in your language, but stay consistent. For that purpose you should add a translator comment about the convention at the beginning of the file for better cooperation.
Use "\\" to print a backslash.
"%" (percent)
In C based programming languages "%" marks the beginning of a format specifier, e.g. "%d6" means insert the next variable here in decimal format with 6 digits. Such format specifiers should be copied into your translated message, at the appropriate spot for your language. Boost's list of format flags for date and time is available at https://www.boost.org/doc/libs/release/doc/html/date_time/date_time_io.html#date_time.format_flags.
Note
If a string is marked with c-format and has a % mark that does not start a format specifier, file a bugreport and tell the developers the location of the string (the lines above the msgid). The developers should fix this in the code. One way to do so is to insert a comment containing xgettext:no-c-format before the gettect call.
In order to continue without having to wait for the developers' fix to propagate you can remove the c-format flag from the #, comment line above. If no other flags are in the line, simply remove the line.
  • To output "%" if a string contains format specifiers, use "%%" in your string.
Reordering parameters
Assume a string "In %d cases the result will be %d.", but in your language you would prefer to write "%d will be the result in %d cases." Now you would get the wrong numbers.
Solution
Insert the ordinal number of the parameter, followed by "$" in the format specifier: "%2$d will be the result in %1$d cases.".
"~" (tilde)
like "%", but for scheme-format.
"{num,optional other specifiers}"
this is a format specifier for C++ code. You cat either copy it verbatim somewhere in your translated message or you may adapt it to your language.
"&" (ampersand)
is the starting character of HTML encoding, which is used in some reports. E.g.
&nbsp;
means NonBreakableSpace.
"<" (less)
is the starting charcter of tags in several markup languages. E.g.
<b>Text</b>
results in bold Text.
See also
GUI Guidelines.

Check syntax and statistics of your .po file

Required
msgfmt -c --statistics LL.po

If that program reports one or more fatal errors for your language, you should review the criticized lines of your file.

Desired

In a second run you might wish to see, where you forgot to add an accelerator:

msgfmt -c --check-accelerators="_" --statistics LL.po

The users will love you if you fix this, too.

Note
There are currently 3 "false positives" which you should ignore:
gnc_numeric
translator_credits
place_holder

Current status

See Translation Status for the current status of the project with respect to translation contributions.

To see some over all statistics how many strings are translated, you can run

for i in *.po; do echo -n "$i:"; msgfmt -c --statistics $i;done

in your po directory.

Priority

As you might have noticed, the po file for GnuCash is rather big by containing more than 4000 strings. It is somewhat helpful to look into different priorities for the different strings there. However, unfortunately the po file format doesn't enable any easy way for the GnuCash project to make the import strings differently from the less important ones. However, for a few files the developer team can give you some guidelines for the general priorities.

The source code location of a string gives you some hint about its priority. In particular, the following file suffixes or file locations imply a certain priority:

  • High Priority: Strings from all *.glade.h files will appear in the GnuCash user interface (UI) somewhere for the user. This may imply a somewhat higher priority to those strings - however, some UI elements are rarely used, so the actual location of the .glade file has an influence as well, but there isn't an easy rule for that anymore.
Tip
If you wish to see the string from a *.glade.h file in it's context, but don't know, where it is hidden in gnucash, you can open the file gtkbuilder/*.glade with Gnomes interface builder Glade.
  • Low Priority: Strings from all *.schemas.h files will not appear in the gnucash UI. Instead, they are shown only inside the
  • Linux: dconf-editor,
  • MacOs: defaults or
  • Windows: regedit program when the user wants to set particular GSettings keys of GnuCash. You can safely consider the .schemas.h strings with lower priority than the others.
    • Register2 feature is a stalled developement. The strings are only visible, if you #configure --enable-register2. Some, but not all strings are marked with a "Register2 feature" comment.

Unfortunately, there is no such simple rule for strings from *.c files or the single large intl-scm/guile-string.c file. The strings from there may be of higher or lower priority, but this isn't easily visible. We can only recommend to start GnuCash on your own with your updated translation, and then check for strings which you see but are not yet translated. Those are for sure of high priority.

Two additional source code locations can be noted as low priority:

  • Everything from the file src/bin/gnucash-bin.c is of low priority because those are the command line options when running gnucash with different options from the command line. This is a use case that is performed only by highly experienced users which are accustomed to using the English-language command line commands. Therefore, translations here are of lower priority.
  • Everything from the folder src/tax or from the intl-scm/guile-string.c file with a comment indicating some file in src/reports/locale-specific is related to tax preparation in the U.S. or in Germany. Hence, those strings are uninteresting for anyone living in different countries, and you can safely consider those with very low priority.

Disambiguation Prefix

At some places, GnuCash uses "disambiguation prefixes" in translatable strings. Here is an old explanation: https://lists.gnucash.org/pipermail/gnucash-devel/2005-October/014236.html ; more explanation is also here: http://live.gnome.org/GnomeI18nDeveloperTips .

There are at least 2 use cases:

  • Abbreviations, e.g.:
msgid "Reconciled"
msgstr "Abgeglichen"
:
msgid "Reconciled:R"
msgstr "Reconciled:A"
  • Sample text to determinate the size of the output cell. Here is no translation required, but the "worst case" of expected text. For the following example the german tranlators copied the longest path of account names from their business account templates as shown in the accounts tab of Gnucash:
msgid "sample:Expenses:Automobile:Gasoline"
msgstr "sample:Aufwendungen 2/4:Reparatur/Instandhaltung:4805 Reparatur u. Instandh. von Anlagen/Maschinen u. Betriebs- u. Geschäftsausst."

@Developers: In the normal source code of a gtk-based program, this feature is available by simply using the Q_( ) macro instead of the _( ) macro.

This would be used e.g. as follows

  /* Translators: This string has a disambiguation prefix */
  func(Q_("Noun|Transfer"));

where the string that shows up in the GUI is simply "Transfer", but the translators need the prefix to be able to choose the correct translation of the noun vs. the verb.

In the latest gettext versions, this disambiguation is solved in yet another way by introducing an additional "context string" in addition to the translatable string, see below. As this is still a relatively new feature, we don't use this in GnuCash yet.

Context by Particular Gettext

A more recent method is "particular gettext" pgettext().

Example:

 pgettext("account status: not cleared", "n")

would result in a msgid like:

#: src/app-utils/gnc-ui-util.c:656
msgctxt "account status: not cleared"
msgid "n"
msgstr ""

See Gettext Manual: Contexts for details.

Testing and submitting your translations

You must check that your new translations are programatically correct (ie: that there are no unclosed quotes, etc). To do this, use the msgfmt program

msgfmt -c --statistics LL.po

Above call will report most important errors in your .po file and must not fail, while the call below

msgfmt -c --check-accelerators="_" --statistics LL.po

whill additionally check if you missed some keyboard accellerators aka hotkeys. But this call formally will always fail as there is at least 1 string, where the underline is no accelerator marker: msgid "translator_credits" from #Adjust the header and special strings. So just eleminate the other warnings and ignore that about "translator_credits".

Note for developers
Another python based tool is i18nspector - checking tool for gettext POT, PO and MO files. It is stricter and shows also warnings about expected entries which our .pot file is currently missing. That can be discussed e.g. in #Background.

If you want to see your translations within a running version of gnucash, simply place your .po file in the po/ directory of your local gnucash repository (which you have previously installed) and from within the po/ directory type (you may need to be root to do this):

make install

Now you can run gnucash with your new translations:

LANG=XXXX /opt/gnucash-git/bin/gnucash

To review the schema strings you can use

since 2.6
dconf-editor
until 2.4
gconf-editor.


Alternative Way with Poedit

If you use Poedit while working on .po file, you would notice that it saves file in .mo format, the very format that GnuCash uses itself. So one would just need to copy this LL.mo file into appropriate subdirectory under /opt/gnucash-git:

 cp LL.mo /opt/gnucash-git/share/locale/LL/LC_MESSAGES/gnucash.mo

The only thing that needs to be changed here is LL which stands for your language code (ka Georgian, de German etc). The rest is as in above method:

LANG=XXXX /opt/gnucash-git/bin/gnucash

Submitting

Only if

  • you want to send your file to the Translation Project or
  • it did not exist before and you want to send it via mailing list, create a gzipped version of it:
gzip XXXX.po
Do the same with your #The glossary file.

In all other cases just push your commit (Git#Pull Requests) or upload the Git#Patches of the files.

Then follow #How to submit changes directly to GnuCash.

Alternatively, you could submit the finished translation via the GNU Translation Project. See the instructions on http://translationproject.org/ and the first section of this document about how to do that.

See Translation Status for the current status of the project with respect to translation contributions.

See also http://lists.gnucash.org/pipermail/gnucash-devel/2009-January/024700.html [FIXME: obsolete?]

Committers

Notify the statistics for the commit message.
gnucash 	2.6.5 	0 / 4628
If you there click on 2.6.5 you will get a fresh gnucash-2.6.5.pot, while we have already an ar.po with 2031 translated and 4 fuzzy messages. By contrast on the as external marked http://translationproject.org/team/de.html, you see:
gnucash 	2.6.5 	unknown 	external
and will not get a new pot file.
But we can bundle this notes with the release announcement, if you add the language to Release Process#TP Status changes.
  • On committing add the name and email address of the last translator from the file as author, e.g.
 git commit --author="Lieutenant Worf <lt.worf@starfleet.org>" --message="Update of tlh.po
 4677 translated messages." # << The msgfmt statistics
  • If you check in a new language, remove a language or see the status of the language changes from partial to (almost) complete, update the numbers in <sect2 id="oview-featuresintl2"> in file guide/C/ch_oview.xml of gnucash-docs.

Problems

intl/Makefile is already registered

When you see this error message during autogen.sh run:

configure.ac:1219: error: `intl/Makefile' is already registered with
AC_CONFIG_FILES.
autoconf/status.m4:844: AC_CONFIG_FILES is expanded from...
configure.ac:1219: the top level
autom4te: /usr/bin/m4 failed with exit status: 1
autoheader: /usr/bin/autom4te failed with exit status: 1
**Error**: autoheader failed.

Then you need to revert the configure.ac script to the original form that is in git:

git checkout configure.ac

Gtk-CRITICAL messages

Note: Fixing this problem is quite difficult. We keep the explanation here in case some developers need it in the future, but if you have no idea what we are talking about in the next sentences, you should rather ignore these Gtk-CRITICAL messages and our explanation here!

If you see any "Gtk-CRITICAL" messages while running gnucash, it is probably because you translated a string differently than how it exists in some other gnome library. You must discover which string you translated differently, and change the translation to exactly match that of the gnome libraries.

To do this, you need to run gnucash under gdb:

LANG=XXXX /opt/gnucash-git/bin/gnucash-env gdb /usr/bin/guile

Then, from within gdb, issue:

run -e main -s /opt/gnucash-git/libexec/gnucash/overrides/gnucash --g-fatal-warnings

Eventually, gnucash should crash (because of the --g-fatal-warnings directive), when it does, issue from within gdb:

backtrace

You should see some output that looks like this:

#0  0xffffe002 in ?? ()
#1  0x42028a73 in abort () from /lib/tls/libc.so.6
#2  0x4019d3d8 in g_logv () from /usr/lib/libglib-1.2.so.0
#3  0x4019d414 in g_log () from /usr/lib/libglib-1.2.so.0
#4  0x40500fdd in gtk_type_check_object_cast () from 
/usr/lib/libgtk-1.2.so.0
#5  0x407292e5 in gnc_mdi_tweak_menus (mc=0x825adb0) at gnc-mdi-utils.c:574
#6  0x40729d13 in gnc_mdi_child_changed_cb (mdi=0x8266fd8, prev_child=0x0,
     data=0x8265fd8) at gnc-mdi-utils.c:861

Notice position #5 which has "gnc_mdi_tweak_menus at gnc-mdi-utils.c:574"? Open that source file and find line 574:

573:  widget = gnc_mdi_child_find_menu_item(mc, "_View/_Toolbar");
574:  gtk_signal_handler_block_by_data(GTK_OBJECT(widget), info);

So, the problem is with the translation of "_View/_Toolbar". The "/" is a menu seperator, so you now know that the problem is with either the translation of "_View" or "_Toolbar". By switching to an English gnucash (LANG=C) and looking through your .po file, you should be able to find out the problem. Change the offending translation to whatever you see in the gnucash app. Remember that the translations must contain the proper underscores.

Watch File accesses

To follow gnucash as it access files,

strace /opt/gnucash-git/bin/gnucash

Thank You

Thanks so very much to all the translators for their hard effort and excellent work.

How to translate the Windows Installer

Follow the instructions in the this thread.

Check doc/README-*

The following filepatterns in doc/ are of interest:

  • README-<language>
  • README-<LL>.win32-bin.txt

This plain text files are often forgotten and become easily outdated. But if gnucash-docs are not installed or the default help browser is for some reason not working, they are the only documentation, which the users can find in your language on their computer.

Q: Shouldn't it be possible by some make-magic, to adjust current-, last-stable-, next-stable- release and release-date - e.g. like target gnucash.1 in makefile? --Fell 20:21, 16 September 2010 (UTC)

Translating the GnuCash Guide and Help

This section describes the actions needed to translate the manuals. There are currently 2 parts:

  • Tutorial and Concept Guide: This is an introduction in the basic principles of double entry accounting and their application in GnuCash. This text is really useful for new users.
  • Help: This is the context sensitive help system.

If you are interested in translating these documents, please decide whether you want only to translate the existing text or whether you want also to improve and cross-check the text with the actual status in the program in your language. It is less effort only to translate the text, but you also run into the risk of doing unnecessary work if you translate explanations which are no longer correct. It might be more effort to create a new text in your language, using the English text only as an inspiration, but it will surely lead to more useable and more helpful documentation. We, the programmers, encourage you to do the latter and create a new text in your language. As you are the one doing the work, the decision is up to you.

Both documents consist of

  • text files in an XML format called DocBook, where the complete document is split into a collection of xml files by chapter, and
  • images: Those are mostly screenshots stored as Portable Network Graphics (png) files in the subdirectory figures/. The xml files will contain links to the png files where they should appear in the text.
See also
Documentation Update Instructions#Screenshots and Images

At the beginning you might wish to concentrate on the text. If desired, you could link the english pictures from C/figures.

You can find an introduction to DocBook in Creating DocBook Documents.

Please use a unified terminology: The translator of the program messages should have created #The glossary file for the translations of the key wordings in the program. Please make sure to use the same terms in the documents, too.

You can check the nightly builds - in particular after your updates were committed.

Prerequisites

Linux

In addition to the software for running autogen.sh, the following additional software is needed for the development on Linux. From the gnucash-docs/README file as of 2016-01-23:

  • libxml2
  • libxslt [Debian packed the required xsltproc in a separate package,
which depends on libxslt]
  • docbook-xsl
  • rarian or
scrollkeeper >=0.3.4 (its ancestor)
  • yelp (for viewing)
Optional:
  • gnome-doc-utils (contains xml2po for the use of po editors like in the it translation)
Additional Requirements for Generating PDF:
  • Apache fop >= 0.95
For the Japanese PDF, Japanese fonts are included. If you want to use
other Japanese fonts you can use the --with-japanese-fonts-dir,
--with-japanese-mincho-ttf, and --with-japanese-gothic-ttf configure
options to select them. fop's TTFReader can't, as of version 1.1
anyway, handle OpenType fonts.
Additional Requirements for Generating Mobipocket:


Note for wiki editors: Keep above section in sync with gnucash-docs/README

Other Operating Systems

Wm says the section below is being edited and not yet complete, see the thread "Help with help" in gnucash-devel for more info


MacOSX

At present we do not know of a suitable DocBook file toolset for MacOSX though a *nix on Mac approach may / should work. We encourage solutions so if someone could document this we would be grateful [Wm: aside is this redundant? Don't most modern Macs include *nix so instructions similar to those I have provided for Win might apply?]

Windows

One way of being able to edit and produce DocBook files on Windows is to use the *nix instructions under cygwin. Instructions for obtaining a suitable toolset follow

  • Get cygwin by going to http://www.cygwin.com and pressing the link for the setup-*.exe file suited to your system
  • save it, virus scan it, etc. as you see fit then run it
  • type "libxslt" into the search box
  • expand Libs and press Skip so that it shows a version for all 3 GNOME XSLT Library options, you want the most recent which it will offer by default, cygwin will work out the gndependencies
  • you may follow similar steps to obtain git, i.e. type "git" into the search box, expand and tick the obvious candidates
  • Note: because these are *nix type text based utilities you are downloading hundreds of kiloBytes not megaBytes; relax, this isn't Windows Update, a few extra progs won't kill your system, promise
  • click next at the bottom rhs and let cygwin do its stuff; this may take some time depending on connection speeds at both ends
  • Aside: if you get a msg about "Package: Unknown package / inetutils.sh exit code 1" you may ignore it, after reading the full blurb, of course, my advice is not to try and track it down unless you have a lot of spare time on your hands
  • When it is done fire up cygwin and do stuff as per the *nix instructions

Further details:

  • don't look for
  • yelp (the viewer)
  • on cygwin: it isn't there; it is an old prog (last updated in 2007, I think
wrong, see http://ftp.gnome.org/pub/GNOME/sources/yelp/. --Fell 17:21, 26 October 2013 (EDT)
) and should probably not be regarded as current though it may be useful if you have it on another OS
[Wm:
  • I wonder how many current *nix users have it?
AFAIK it is still part of the GNOME desktop: google search for yelp+gnome --Fell 17:21, 26 October 2013 (EDT)
Should ref to yelp be made less prominent?
It is the only way I know to get some context sensitive help from inside GnuCash. --Fell 17:21, 26 October 2013 (EDT)
]

Wm says the section above is being edited and not yet complete, see the thread "Help with help" in gnucash-devel for more info

Java (all OS)

Eclipse has

  • several Git plugins,
  • a C Development Tool with autotools support,
  • a feature rich, builtin XML editor and a few plugins like XML WYSIWYG editors

or in other words, all what is needed here, available.

The Procedure

First, you have to download the most recent version of the gnucash-docs package. This is done as follows:

  1. Checkout the documentation in a directory gnucash-docs:
    git clone https://github.com/Gnucash/gnucash-docs -b maint gnucash-docs
  2. Create your working branch. If you plan to use pull requests, you should use unique names like guide-LL-chapter-x-part-n instead of working-branch.
    git checkout -b working-branch
  3. Create a new directory (if it doesn't already exist) in guide/<locale> where <locale> is of the form <language>[_<region>] e.g. es, en_GB, or pt_PT. See #Naming Conventions.
    See Locale_Settings#IETF_language_tags for details.
    If your translation is the first for your language do not add a region code. So also other regions benefit from your translation.
    And copy the files from guide/C into this directory:
    cp -r guide/C guide/LL
    Not in latin letters written languages
    They need additinal files to process printing to pdf. See ja or ru as examples.
  4. Add the directory you created to [gnucash-docs/]configure.ac (as a new line guide/<language>[_<region>]/Makefile under AC_OUTPUT) and [gnucash-docs/]guide/Makefile.am (at the end of SUBDIRS).
  5. Notify git about the new directory and files. "*" is quoted by "\".
    git add guide/LL/\*.\*
  6. Now the real work:
    1. Edit all the xml files (see #DocBook xml editors below for suitable editor programs) and translate them into your language.
      • In the event the english documents get updated, it might be useful to have a comment in the header of each chapter in the form:
      <!-- Translation based on maint/master commit 1234567 2014-01-01 10:00 UTC -->
      
      Then you can later view a diff between the current and that version of the english file in one window to locate the changes and adjust the translation in another window. Some translators like to do also the translation in a graphical diff program like KDiff3 between english and their language because the <sectN id>s will synchronize the text.
    2. Adjust gnucash-{guide|help}-<lang>.omf. This file follows Open Metadata Framework (OMF), a subset of the Dublin Core description for metadata. See Writing ScrollKeeper OMF Files for details like mandatory standard elements. Without this file your document will not show up in yelps index page. At some point in the future we might extract this data automaticly from the docs using Gnome Doc Utils.
      Currently it is also necessary to replace in {guide|help}/<lang>/Makefile.am the "C" in both occurrences of $(docname)-C.omf by <lang>.
    3. There are some general headers, which do not appear in the xml-files in your locale directory. But they can be translated by adding a section in xsl/l10n.xml. Obey the comment at the beginning. If you use non-ASCII characters, you should run recode -d <input_encoding>..h0 l10n.xml, where input_encoding might be u8 for UTF8, to get them right encoded.
    4. In addition to the text, you need to recreate the image files in guide/C/figures so that they are appropriate to your language. For details (size, style,...) see Documentation_Update_Instructions#Screenshots_and_Images.
      1. Most of them are screenshots of a gnucash session - save them as png files because this will use a lossless compression. In theory some of the files in gnucash/doc/examples could be a starting point therefor, but they are currently - 2010-09-17 - broken.
      2. A few figures are in scalable vector graphic (svg) format. Here you can edit the files e.g. in Libre/OpenOffice and translate the strings and adjust the size.
    5. Watch out: The documentation itself is probably outdated in many places, as it has been written for the 1.8.0 version of gnucash. You can watch the current progress in Concept_Guide#Ongoing_work. If you encounter any description that is wrong for the current version of gnucash, do not hesitate to ignore the english original text and instead describe the situation of the current version of gnucash in your translation. It would be even better if you have the time to change the english original text as well, or at least file a bug against it, but even if you can't do that, feel free to describe the actual state of gnucash in your translation and simply ignore the original english text.
      If you file a patch against the english text, you can also file a patch, which adds your name and email address to AUTHORS, so you will become famous. ;-) Ask to apply this latter patch also against gnucash/DOCUMENTERS because both files should have the same content. More details about editing the english text can be found in the Documentation Update Instructions.
  7. Test your xml files for syntax errors by running xmllint on the main file gnucash-{guide|help}.xml, e.g.:
    xmllint --valid --noout gnucash-guide.xml
    The program xmllint is part of the package libxml.
    Tip: Some xml aware editors have a menu entry like validate to run this test.
  8. Optional: test your work with yelp - there are autotools files, to support you:
    1. After you copied the directory, change to the 'root' of the gnucash-docs package and run once
      ./autogen.sh.
    2. If you want to change some default values like installation paths, run
      ./configure --help,
      to see the available options.
    3. There are 2 variants of configuration: with or without a separate build directory. The downside of a separate build dir is you will have to cd to the builddir to run the make commands and then again cd to your sources to edit them again - at least if you work only with command line tools. The benefit is your source directory will not be cluttered. In most cases developers prefer to use a separate build directory.
      1. with build directory run
        md build ## only once required
        cd build ## every time required, you want to run make or configure
        ../configure with your desired options,
      2. without build directory just run
        ./configure with your desired options.
      Tip for Windows and MacOSX, which have no Yelp: use --disable-scrollkeeper
    4. After your editing of the files, run
      make
      for some additional processing. The default target are xml files for yelp. Other targets like html, pdf, ... must be given explicit:
      make html
    5. To check the syntax of your files, run
      make check
      This will run the xmllint command above over the full diretory structure.
    6. If you had choosen something with "$HOME" as prefix in your configuration, run
      make install,
      otherwise run
      sudo make install.
    7. Run
      yelp <path/filename>
      with the path, to where you installed the files. This path consists of $PREFIX/share/gnome/help/gnucash-guide/LL/gnucash-guide.xml or gnucash-help.xml.
      Tip
      export XDG_DATA_DIRS=/home/<yourname>/<your-install-dir>/gnucash-docs/share:${XDG_DATA_DIRS}:/usr/local/share:/usr/share
      will tell yelp the path to your version.
    8. Repeat steps 4-7 until everything is right.
  9. Update the branches of your local repo often:
    git checkout maint # switch to maint
    git pull --rebase # update it
    git rebase maint working-branch # update working-branch
    git checkout working-branch # switch back to working-branch
  10. Follow #Submitting your work direct to GnuCash.

To translate the help files, repeat steps 2-10 but replace the "guide" directory with "help".

Committers: If it builds fine, add the directory changes from 2. also to gnucash/maint/packaging/win32/install-impl.sh and merge the change to master, if required.

DocBook xml editors

For editing these DocBook xml files, various editors can be used: http://en.wikipedia.org/wiki/DocBook might contain pointers to some, or also http://www.tldp.org/LDP/LDP-Author-Guide/html/tools-edit.html . Some developers use Eclipse.

People have said that AbiWord and OpenOffice / LibreOffice have support for DocBook available, in which case you could directly edit gnucash's xml files with those. However, those editors will probably not work with the current multi-file setup where each chapter is in a separate XML file and the main XML file contains points to the chapters' files. As a workaround, you can copy the relevant parts of the main XML file into the chapter's file by a plain text editor so that it "looks like" one single DocBook document. This can be opened with OpenOffice/LibreOffice and edited as normal. After that, the added parts from the main file need to be removed again, and then you have the edited chapter text.

Using po editors

Some translators are more familiar with using po file editors like poedit, Kbabel, Gtranslator etc. For those people, it is possible to convert the content of the help documents (the DocBook XML files) into po files, translate the messages in the po file, and convert the result back into xml. This section describes how to use this approach for the gnucash help documents.

The downside of this approach
It will be much harder for the documenters to fix e.g. broken links in times where you are inactive.

In order to use po files you have first to convert xml files to po files, translate them and then convert back to xml.

First you have to install a fresh version of gnome-doc-utils. For example in a Fedora 13 system

 yum install gnome-doc-utils.
  1. Do the conversion with the following commands (xx is your language code, example: el for Greek):
    xml2po -o helpfile.xml.pot helpfile.xml
    mv helpfile.xml.pot helpfile.xml.xx.po
  2. Translate helpfile.xml.xx.po using your favorite po editor.
  3. Convert back to xml:
    xml2po -e -p helpfile.xml.xx.po -o xx_helpfile.xml helpfile.xml
  4. Test your translated xml file as mention above:
    xmllint --valid --noout xx.helpfile.xml
  5. Remove xx from xx_helpfile.xml and you are ready to commit it in your language help directory

Workaround for the broken OpenSuse help system

As in OpenSuse 10.x and 11.x the help system is broken, there is a workaround, to review the result of your work. Locate gnucash-guide.xml in your system, probably /usr/local/share/gnome/help/gnucash/C. Then from that directory do

xmlto -o /path-to-where-you-want-to-store-docs/ html gnucash-guide.xml

which will make an html copy of the docs.

Copy over the 'figures' directory with contents to your html directory above and set a bookmark in browser to the index.html so that you get the docs any time you need them. The context sensitive facility is lost, but for general reading it works.[1]

GnuCash Maintainer Tasks

For new languages the core developers have some additional tasks:

New Script

Check make pdf! Do we need additional TTFs?

First Nightly

Ask the webmaster of code.gnucash.org to create the new directory.

First Release

Adjust htdocs ...

How to translate the files containing the new account hierarchies

This section describes the actions needed to translate the files containing the new account hierarchies.

Preliminary Considerations

However, please take a moment to think about the intention of the account hierarchy templates. The intention of those files is much more language-related than any other of the files inside gnucash. A string-by-string translation is not the best thing to do here. Instead, you can and should find out an actual recommended account structure which makes sense in your language, and implement that one in your language. By this, I mean several of the english accounts make sense only in the U.S., but probably not in other countries. Hence, your translated account template should not contain them. Also, you can add additional parts of the account templates for your language as well, if the users are likely to need it.
For example, in the U.S., users are likely to own a car, and for this reason there is an account structure template for "Car". If in another hypothetical region users are likely to own, say, a spaceship instead of a car, you should remove the "Car" template and instead create a new account structure that represents all accounts related to the ownership of a spaceship, and offer this as additional "Spaceship" template.
In other words, you should feel free to create a completely new account template structure that is most suited to your language. The english-language templates are just a proposal, but will need further adaption and not a string-by-string translation.

Having said that, here is how you would start for a direct translation of the English templates:

In this section always replace <locale> with your language code and optional your region code, if there are multiple countries with the same language and your templates are fitting only to one of them.

Prerequisite

Initialize accounts/<locale>

  • If it does not already exists, execute the following steps to initialize your accounts/<locale> directory:
  1. Copy the directory accounts/C to accounts/<locale>
  2. In the directory accounts/ change the file CMakeLists.txt and add your <locale> to both alphabetical sorted lists:
    1. add_subdirectory(C)
      :
      add_subdirectory(<locale>)
      :
      
    2. set(accounts_DIST ${C_DIST} ...  ${<locale>_DIST} ... ${dist_list} PARENT_SCOPE)
      
  3. In accounts/<locale>/CMakeLists.txt adjust the <locale>:
    :
    set_dist_list(<locale>_DIST ${dist_account_DATA} CMakeLists.txt)
    
    install(FILES ${dist_account_DATA} DESTINATION ${ACCOUNTS_INSTALL_DIR}/<locale>)
    file(COPY ${dist_account_DATA} DESTINATION ${ACCOUNTS_BUILD_DIR}/<locale>)
    
    Note
    The next 2 steps are are discussed in #Localize the Account Charts below.
  4. Localize acctchrt_full.gnucash-xea. This is only a helper file where you have all accounts in their context.
  5. Now create the real modules by merging the respective parts from acctchrt_full.gnucash-xe in the other acctchrt_*.gnucash-xea files.
  • If you remove or create new files, adjust in accounts/<locale>/CMakeLists.txt the list
    set(dist_account_DATA
      ...)
    
  • Whenever you changed one or more CMakeLists.txt files, you have to rerun
    cd ${SOURCEDIR}
    cmake <your options>
    cd ${BUILDDIR} # e.g. .build
    make # or ninja
    

Localize the Account Charts

Tip: C/acctchrt_full.gnucash-xea is a helper file. Because we prefer modularization it usually is not part of the distribution. It contains the full tree of the personal en_US accounts. The other acctchrt_*.gnucash-xea files contain single branches of that. So after translating it you can copy&paste the respective parts in the other files to ensure consistency between them.

If you modularize the templates you should start with acctchrt_common.gnucash-xea - it is the most basic.

For each desired acctchrt_* file in that directory:

  1. Change the gnc-act:title, gnc-act:short-description, and gnc-act:long-description to contain appropriately translated text. Do not add any newlines in the long description except at the end and begining of the string.
  2. For each gnc:account in the file
    1. translate the act:name, and act:description fields.
    2. Optionally you can
      1. assign an account number <act:code>1234</act:code> after <act:commodity-scu> or
      2. add a note as i.e. in
           <act:slots>
            <slot>
              <slot:key>hidden</slot:key>
              <slot:value type="string">true</slot:value>
            </slot>
            <slot>
              <slot:key>notes</slot:key>
              <slot:value type="string">Some additional text about the usage of this account</slot:value>
            </slot>
            <slot>
              <slot:key>placeholder</slot:key>
              <slot:value type="string">true</slot:value>
            </slot>
          </act:slots>
        
    Please do not translate any other fields or the internally used ROOT account.
  3. To avoid typos, run the AccountHierarchyTemplate#Syntax Check.
  4. New files to be shown to the user must be added to the account_DATA section,
    helper files like acctchrt_full.gnucash-xea to EXTRA_DIST in Makefile.am.

Note again: You absolutely don't need to translate all of the files from accounts/C. A subset of those are fine as well. Probably several of them will not apply to your local legislative/economic system anyway. For a really customized account hierarchy you might better create a new account hierarchy file in GnuCash, and then, by hand-editing the xml code, split it up into several files and cut&paste the appropriate tags from the accounts/C/acctchrt_* files.

If you wish to add new accounts manually, you will need some additional guids, those funny random numbers. To get them you can use:

uuidgen | sed -e 's/-//g'

or use an online uuid generator like this one (any other one will do as well). Be sure to untick "Hyphens" to generate gnucash compatible guids. If you forget or the site you use doesn't offer that option, simply remove the hyphens yourself.

Dependencies
uuidgen is in a package 'util-linux', sed has its own package.

AccountHierarchyTemplate describes, how to create new templates e.g. for business purpose according local rules.

Test your Work

To test your work under real conditions, you should finally Compile & Install it. But before you need an adjusted Makefile.am in your accounts/<locale> directory. If none exists, copy that from accounts/c and adjust it:

  • 1. Line (accountdir): adjust the locale.
  • If you removed accountcharts, remove their line from account_DATA,
  • if you added new files, insert lines with their names and "\" (the line continuation symbol).

How to create localized Income Tax Tables

Note
This section is under developement and mostly untested!

As of the beginning of 2015 there is an nice US tax module with a report and export in TXF format. Additional there is a small de_DE derivate for the monthly VAT declaration. you can compare them to see, how you can tweak them to get something else.

Tax Table

From https://github.com/Gnucash/gnucash/blob/maint/src/tax/us get the files

  • txf.scm - the tables, and
  • txf-help.scm - the descriptive help strings for each TXF code. This is also used to get a table in gnucash-docs/help/

They have relative simple table structures:

(define txf-tax-entity-types 

contains a list of forms for different entities (person, company,...).

(define txf-income-categories ...
(define txf-expense-categories ...

contain for each of the above lists the entries. If you compare this file with the respective *_DE file, you can see an example localization with additional categories:

(define txf-asset-categories ...
(define txf-liab-eq-categories ...

It should be possible to adapt the lists for your tax forms. Then you might submit them to an enhancement request in buzilla, product:Gnucash, component:TXF Export.

Report and Export

Probably you need to have other export reports. That are in https://github.com/Gnucash/gnucash/tree/maint/src/report/locale-specific/us like

  • txf-export.scm
  • us.scm is unused

Here you might probably need some help from a programmer.

How to translate the website

Note about OS
These instructions are written for linux. They will probably work on other unix-like systems as well, but not on Windows. On Windows you may be able to do this if you set up a linux-like environment (with cygwin or MinGW[-w64]), but that's untested so far. Please report your findings here.
  • Get a checkout of https://github.com/Gnucash/gnucash-htdocs
  • Run the command
    make pot
  • Then depending on whether or not a translation for you language exists already (complete or not):
    Existing translation

    Run the command
    msgmerge po/LL.po po/gnucash-htdocs.pot -o po/LL.po
    where LL is your language code, see #Naming Convention above.
    Alternative
    make msgmerge
    will update all .po files
    New translation

    In the po directory run
    msginit
    If that fails copy the newly created file po/gnucash-htdocs.pot to po/LL.po, where LL is your language code, see #Build a new .po file earlier.
  • Translate the po/LL.po file as described in #Translating the .po file.
    Priority
High
Startpage: {index|externals/*}.phtml
Low
sizing.phtml (outdated) search/templates/NMZ.* (unused)
Medium
the rest depends on you.
  • Run
    msgfmt -c --statistics po/LL.po
    to see your success.
  • Optionally run
    recode -d <input_encoding>..h0 po/LL.po
    
    to get special characters HTML quoted.
  • Send your LL.po either as
    • GitHub Pull Request or
    • attachment of a Bugzilla enhancement request to the mainteiner team.

Maintainers Task

For new languages adjust the appropriate

  • git add po/LL.po
  • Makefile,
  • lang.php and
  • externals/header.phtml rules.

To make the changes visible, you will have to

  • run
msgfmt -c --statistics po/<ll>.po -o locale/<ll>/LC_MESSAGES/gnucash-htdocs.mo
or to compile all translations
make mos
  • and include the mo file(s) in your commit.

Tips for Developers

This section collects some notes for developers/programmers on how to correctly prepare their code for translations.

How to make strings in code translatable

Strings in Glade files

In Glade files translatable strings have the form:

   <property name="label" translatable="yes">Some translatable text with _mnemonic</property>
   <property name="visible">True</property>
   <property name="can_focus">True</property>
   <property name="receives_default">False</property>
   <property name="use_underline">True</property>

For the mnemonic property use_underline see #Special_characters_and_other_tips.

Strings in C files

Normally, strings in C code just need to be enclosed with the _( ) function. Well, actually it is a macro expanding gettext() into a function, but this is usually just an implementation detail (For more macros search your IncludeDir for Glib and that for gi18n.h). For example,

 func("A translatable string!");

should instead be written as

 func(_("A translatable string!"));

However, it is important to keep in mind that _( ) is a function; this means that in certain situations it cannot be used. For example,

 gchar* array_of_strings[] = {_("first string"), _("second string"), _("third string")};

would cause a compiler error. Instead, these strings should be wrapped with the N_( ) macro, which is declared as #define N_(String) gettext_noop(String). Then, whenever one of the strings is actually used (as opposed to declared), it should be wrapped with the _( ) function again:

 gchar* array_of_strings[] = {N_("first string"), N_("second string"), N_("third string")};
 func(_(array_of_strings[0]));

Another use case for gettext_noop are strings which should be stored untranslated in some backend (configuration or GnuCash data), but displayed translated to the user.

See also: #Disambiguation Prefix

Strings in C Preprocessor Macros

Xgettext doesn't run cpp and can't see inside macros. Call gettext in the macro, not on the macro.

Wrong
#define EXPLANATION "The list below shows ..."
/* snip */
gnc_ui_object_references_show( _(EXPLANATION), list);
Right
#define EXPLANATION  _("The list below shows ...")
/* snip */
gnc_ui_object_references_show(EXPLANATION, list);

Strings in C++ files

For strings in C++ files the same techniques as for C files can be used. The drawback of this is one could only work with C-style strings (char*), which means one would not be able to benefit from the many advantages of C++.

A better alternative is to use boost::locale's message translation functionality.

In order to use this, one needs to add the following near the top of the C++ source file:

#include <boost/locale.hpp>
// Optionally:
namespace bl = boost::locale;
// to be able to use bl::translate and bl::format instead of boost::locale::translate and boost::locale::format

After that using boost::locale essentially means using boost::locale::translate() instead of gettext(). This function is designed to be used in a stream and work with the locale imbued in the stream, so one would usually

  • create a stringstream
  • imbue the desired locale
  • stream the result of boost::locale::translate into this stream
  • when completely done, copy the string from the stringstream.

The link above gives some more elaborate examples and you can find some examples in the gnucash code already as well, like in filepathutils.cpp.

The biggest hurdle at the time of this writing is imbuing a locale in the stream. For most of the uses in gnucash this should simply be the global locale. However we don't set this yet in the C++ environment (where to do this still has to be determined). So the few spots in the code where we use this feature are generating a locale on the spot to imbue and that continues to be required until we set a global C++ locale.

Formatted strings

Some strings come have placeholders in that can only be completed at runtime, like a number of transactions in a selection, or the name of a file to be opened. The C world has a whole set of functions that work with such format strings (printf and all variants).

For C++ a more elegant method exists as well in the form of boost's localized text formatting. These have several advantages over the C based format functions, so in C++ it's recommended to use the boost::locale::format features.

Plural Forms

Not all languages have the same simple kind of plural forms as english. See gettexts Plural-forms for details.

Empty Strings

There is no need, to mark an empty strings as translatable. The gettext tools use the empty string for their message catalog header. So it will confuse translators to have your preceding comment in the header of their .po file.

When to use N_() and when to use _()

For developers, both N_() and _() seem to be used. What is the difference?

Simple English:

N_("s")
will mark string "s" as translatable, but will not translate in code.
_("s")
will mark string "s" as translatable, and will be translated on the spot.

However, some code may be of the form...

 x = "A description";
 alert(_(x));

In this case, the x, although is meant to be translated, will not be successfully translated because the engine had no idea the "A description" string is meant to be translated.

The gettext solution is as follows:

 x = N_("A description");
 alert(_(x));

where the first statement assigns string (marked as translatable to x), and the second statement takes the string, passes it as a parameter to _() which translates it, and is then passed it as a parameter to alert().

How to give the translators a clue

Sometimes it is useful to give the translators some additional information about the meaning of a string. To achieve this effect, you can insert a section of comment lines direct before the related string, where the first comment starts with "Translators:". There must not be any control statement between comment and string.

In C based files

Example:
 /* Translators: the following string deals about:
  * The Answer to Life, the Universe and Everything
  * Source:
  * http://en.wikipedia.org/wiki/Phrases_from_The_Hitchhiker%27s_Guide_to_the_Galaxy */
 func(_("42"));


In the pot and po files, this comment will show up as follows:

#. Translators: the following string deals about:
#. The Answer to Life, the Universe and Everything
#. Source:
#. http://en.wikipedia.org/wiki/Phrases_from_The_Hitchhiker%27s_Guide_to_the_Galaxy
#: foo/bar.c:123
msgid "42"
msgstr ""


Note: An empty comment line will end the process. If the string " Source:" were missing, the URL were not part of the output.

In SCM files

If the first expression has a translatable string and the file has a long header comment, split the line before the string and insert a translator comment. Otherwise the POT file will be flooded with file headers.

Example:
;; Boston, MA  02110-1301,  USA       gnu@gnu.org
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;


(define GNC_COMMODITY_NS_CURRENCY
;; Translators: Namespaces of commodities
   (gettext "CURRENCY"))

In glade files

In glade 3.16, using the edit button for a label, one can enter translatable, translation context and comment. From intltool fails to extract comments from glade-type files when not last attributes:

 <property name="label" translatable="yes" context="infinitive" comments="my comment">
should work.

Introducing new terms

If you introduce new terms which are more than once used, you should include them to #The_glossary_file. The instruction is in #Terms missing or inadequate in the glossary file.

Borrowing Code

Ideally they would use their own text domain like in our use of ISOcode.

If not - like in 2.7 /borrowed/goffice - we should consider some msgcat magic with their po files to

  • update our existing po files
  • create new po files.

At present a bash script was written to take care of the first part: import-goffice-translations.sh


This script can be used to import translations from the goffice source directory into our po files. Note this script will run over all of our existing po files in one go, so it can only be used to update all po files at once.

There is no code to create new po files. I think this can be done in a few steps as well, as in

  • create the new po file as we would normally do for gnucash (see msginit elsewhere on this page)
  • run the script mentioned above to import goffice translations
  • this may change more files than needed, use git checkout to undo the changes to files you didn't want to alter
  • continue as usual with translation work.

Further Reading

If you want to read more about this topic, GnomeI18nDeveloperTips might be a good starting point.

Follow the rules of the Gnome Human Interface Guide like Writing style and Typography.

More technical and historical details can be found in Gettext Manual: The Programmer’s View and Guile Reference Manual: Support for Internationalization.

Project Maintainers

Some notes for Maintainers:

When to msgmerge

Some translators complain, it is so hard to update .po files. When should we do it?

From https://translationproject.org/html/robot.html
The Translation Project robot is not allowed to update the registry with new information about maintainers, languages, or translators, nor is it allowed to process POT files. These things are still handled by hand. When a new POT file is being registered by a project coordinator, the registering script calls msgmerge when previous translations exist, so translators are notified of a PO file which is up-to-date. New POT files are announced automatically to a team's mailing list, if the team has asked for this feature.

So for $TP_LINGUAS it is done on each release. Probably we should do it, too?

Merging with divergent source trees

The source tree in GnuCash 2.7 was thoroughly restructured, so it is now very different from 2.6. One side effect of this is that the MsgIds in our 2.7 *.po files are in a completely different order compared to 2.6. This means git merge of maint in unstable will break the 2.7 *.po files. So when merging from 2.6 to 2.7 we have to ignore all changes in the maint po files at this point. To rescue at least parts of the work, try this.

# Set your language in LL, e.g.: export LL=de
# while you are in the maint branch save your .po file outside of the repo e.g. as maint."$LL".po
git checkout unstable
git merge maint                                       # resolve the full LL.po merge by using the unstable version of LL.po (other files can be merged as normal)
cd ../build                                           # assuming ../build is your build directory
make pot                                              # update the template with changes from maint
cd ../src/po                                          # assuming src is your gnucash git repo
msgmerge "$LL".po ../build/po/gnucash.pot -o "$LL".po # update the po of your $LL
msgcat --use-first "$LL".po ../<path_to>/maint."$LL".po -o "$LL".new.po

For the best result you might test different msgcat parameters. Finally

mv "$LL".new.po "$LL".po

Background

This section was started to get some overview, work in progress! Beyond it helped to get rid of the outdated #IntlTool in version 2.7.6. For details see the GetText Manual.

Suggestions for Testing

Because it is hard for 'normal' translators to handle all the different tools, we collect here tools, whicht might be easier to use in the one or other component:

Packages

In our build process we use beneath self written scripts the following packages:

  • glib2-devel: glib-gettextize
  • gettext, probably broken in:
    • -runtime: msgfmt, ...
    • -tools: (autopoint), gettextize, msg*, xgettext, ...
    • -runtime-tools-doc: manuals etc.
  • (only up to 2.7.5) intltool: Intltool*

IntlTool

From freedesktop Software/intltools intro, last edited Sat 18 May 2013 05:42:11 PM PDT:

intltool is a set of tools to centralize translation of many different file formats using GNU gettext-compatible PO files.
The intltool collection can be used to do these things:
  • Extract translatable strings from various source files (.xml.in,glade, .desktop.in, .server.in, .oaf.in).
  • Collect the extracted strings together with messages from traditional source files (.c, .h) in po/$(PACKAGE).pot.
  • Merge back the translations from .po files into .xml, .desktop and oaf files. This merge step will happen at build resp. installation time.

And in the more recent Launchpad Internationalization Tool Collection site it seems to concentrate on xml.

Conclusio: In case we remove it

  • from gnucash.git (only up to 2.7.5): see #configure.ac,
  • from gnucash-docs: one translation is using xml2po instead,
  • htdocs: not used.

Googles first hit for "Intltool manual" is Localize using gettext and intltool, which starts with

This page consists of some parts from Christian Rose's L10N Guidelines for Developers document that could not be moved anywhere else. It does not provide enough information on localization your application, and may be outdated. Please look for additional resources for developers at TranslationProject wiki pages.
Ideally, this will be integrated together with Malcolm Treddinick's Internationalising GNOME application document and GnomeI18nDeveloperTips wiki page.

Our Build Process

The build process goes through the following scripts:

autogen.sh
Note
Autotools were only used up to 2.6 branch. 2.7 and later use CMake instead.
glib-gettextize
helps to prepare a source package for being internationalized through gettext. It is a variant of the gettextize that ships with gettext.
glib-gettextize differs from gettextize in that it doesn't create an intl/ subdirectory and doesn't modify po/ChangeLog (note that newer versions of gettextize behave like this when called with the --no-changelog option).

»Note that on GNU systems, you don't need to link with libintl because the gettext library functions are already contained in GNU libc.« (glibc) [2]

intltoolize
copy intltool related files to software package
configure.ac
  • Sets
ALL_LINGUAS= ...

In recent versions it can be substituted by LINGUAS files in the po directories. We should consider this to distinguish TP vs. GC maintained translations.

GETTEXT_PACKAGE=gnucash

This is intltool style. Gettext uses PACKAGE=...

AC_SUBST(GETTEXT_PACKAGE)
AC_DEFINE_UNQUOTED(GETTEXT_PACKAGE, "$GETTEXT_PACKAGE",
   [GetText version number])

AM_GNU_GETTEXT_REQUIRE_VERSION|AM_GNU_GETTEXT_VERSION(x.yy.zz) can be used to require a version. We should use it to avoid ...

  • TP expects > 0.11.
  • intltool-update checks for GNU gettext >= 0.12
  • gnulib expects >= 0.17
  • RHEL7 offers 0.18.2 (RHEL6 0.17)
  • Debian Stable offers 0.18.1 and in backport 0.19.3
  • Glade2 msgctxt and Glade3 were implemented in 0.18.3 - July 2013
  • GSettings and Desktop entries are implemented in 0.19 - June 2014
  • and got bugfixes until Version 0.19.3 - October 2014
  • Scheme format strings got a fix in 0.19.4 - December 2014
  • AppData support: gettext-0.19.6 released - 2015-09
  •  ? custom XML formats in 0.19.7
<gjanssens> GtkUIManager files don't contain translatable strings as far as I know. They're not in POTFIILES.in either.

This setting then can be used by autoPoInt. Watch the caveaets from https://www.gnu.org/software/gnulib/manual/html_node/gettextize-and-autopoint.html:

On the other hand, if your package is not as concerned with compliance to the latest standards, but instead favors development on stable environments, the steps are:
  1. Determine the oldest version of gettext that you intend to support during development (at this time, gnulib recommends going no older than version 0.17). Run autopoint (not gettextize) to copy infrastructure into place (newer versions of gettext will install the older infrastructure that you requested).
  2. Invoke gnulib-tool, and import the gettext-h module. # Fixme: Meaning?
Regardless of which approach you used to get the infrastructure in place, the following steps must then be used to preserve that infrastructure (gnulib’s bootstrap script follows these rules):
  1. When a script of yours run autopoint, invoke gnulib-tool afterwards.
  2. When you invoke autoreconf after gnulib-tool, make sure to not invoke autopoint a second time, by setting the AUTOPOINT environment variable, like this:
$ env AUTOPOINT=true autoreconf --install
AM_GLIB_GNU_GETTEXT          # FIXME: Meaning?
  • runs checks with settings
    • AC_PROG_INTLTOOL
    • AC_CHECK_HEADERS(ltdl.h, ...
dnl Make sure we have a proper gettext installed
AC_MSG_CHECKING(for a valid gettext/gmsgfmt installation)
if test "$gt_cv_have_gettext" != "yes" || test "x$GMSGFMT" = "x"; then
  AC_MSG_RESULT(no)
  AC_MSG_ERROR([Cannot find Glib Gettext.  Maybe you need to install the gettext package?])
else
  AC_MSG_RESULT(yes - $GMSGFMT)
fi
makefile.am

has the following section:

.PHONY: pot
pot: Makefile po/POTFILES.in
	rm -f po/$(PACKAGE).pot
	${MAKE} -C po $(PACKAGE).pot


$(srcdir)/po/POTFILES.in: make-gnucash-potfiles .potfiles
	if test -w $(srcdir)/po/POTFILES.in ; then ./make-gnucash-potfiles > $(srcdir)/po/POTFILES.in ; fi

# Creation rules so that po/gnucash.pot can always be created for
# make dist.
po/gnucash.pot: po/POTFILES.in
	${MAKE} -C po gnucash.pot

.potfiles:
make-gnucash-potfile.in

This is a self written perl script from the time as gettext had several issues. It creates po/potfiles.in, the "list of files which contain translatable strings". This script is not used in a cmake based build environment for gnucash. As gnucash 2.7.4 and more recent is cmake only, this file has been removed starting from that release.

po/makefile.in.in
This is always copied from your installed intltool package. It contains
MSGMERGE = INTLTOOL_EXTRACT="$(INTLTOOL_EXTRACT)" XGETTEXT="$(XGETTEXT)" srcdir=$(srcdir) $(INTLTOOL_UPDATE) --gettext-package $(GETTEXT_PACKAGE) --dist
GENPOT   = INTLTOOL_EXTRACT="$(INTLTOOL_EXTRACT)" XGETTEXT="$(XGETTEXT)" srcdir=$(srcdir) $(INTLTOOL_UPDATE) --gettext-package $(GETTEXT_PACKAGE) --pot

So it calls in both cases intltool-update with different parameters.

There are several commands in the packages with overlapping funtionality:

intltool-prepare
Extracts all localized strings in .desktop style files (including ".desktop", ".soundlist", ".keys" and ".directory") into corresponding po files. See man:/intltool-prepare(8)
intltool-extract
extracts strings in the specified XML/INI type SOURCE_FILE and writes them into a C header file. Currently supported types are:
"gettext/glade" (.glade, .glade2)
"gettext/gsettings" (.gschema.xml)
"gettext/ini" (Generic INI file)
"gettext/keys" (.keys)
"gettext/rfc822deb" (RFC 822 format file) Debian configuration files
"gettext/quoted" (all strings within "")
"gettext/schemas" (.schemas)
"gettext/scheme" (.scm)
"gettext/xml" (Generic XML file)
xgettext
Extract translatable strings from given input files. See --language for supported list:
--language=NAME
recognise the specified language (C, C++, ObjectiveC, PO, Shell, Python, Lisp, EmacsLisp, librep, Scheme, Smalltalk, Java, JavaProperties, C#, awk, YCP, Tcl, Perl, PHP, GCC-source, NXStringTable, RST, Glade, Lua, JavaScript, Vala, Desktop)
Options, which we should check:
--copyright-holder=STRING
set copyright holder in output
--foreign-user
omit FSF copyright in output for foreign user
--msgid-bugs-address=EMAIL@ADDRESS
set report address for msgid bugs
intltool-update
updates PO template file and merge translations with it. You must change working directory to the subdirectory containing translations (usually "po/") before running intltool-update. Special options:
--maintain
Search for left out files, which should have been listed in POTFILES.in or POTFILES.skip. A list of all these files are written into another file called "missing".
--headers
Executes intltool-extract(8) to extract strings inside XML/INI style files listed in POTFILES.in, and writes the extracted strings into header files, so that the strings can be recognised by xgettext(1).
--report
Display a status report for all translations in the software.
Warning
Usually you will not want to commit the changed .po files!
intltool-merge
merge translated strings into various types of file.