git/po/README
Ævar Arnfjörð Bjarmason 1ff750b128 tests: make GIT_TEST_GETTEXT_POISON a boolean
Change the GIT_TEST_GETTEXT_POISON variable from being "non-empty?" to
being a more standard boolean variable.

Since it needed to be checked in both C code and shellscript (via test
-n) it was one of the remaining shellscript-like variables. Now that
we have "env--helper" we can change that.

There's a couple of tricky edge cases that arise because we're using
git_env_bool() early, and the config-reading "env--helper".

If GIT_TEST_GETTEXT_POISON is set to an invalid value die_bad_number()
will die, but to do so it would usually call gettext(). Let's detect
the special case of GIT_TEST_GETTEXT_POISON and always emit that
message in the C locale, lest we infinitely loop.

As seen in the updated tests in t0017-env-helper.sh there's also a
caveat related to "env--helper" needing to read the config for trace2
purposes.

Since the C_LOCALE_OUTPUT prerequisite is lazy and relies on
"env--helper" we could get invalid results if we failed to read the
config (e.g. because we'd loop on includes) when combined with
e.g. "test_i18ngrep" wanting to check with "env--helper" if
GIT_TEST_GETTEXT_POISON was true or not.

I'm crossing my fingers and hoping that a test similar to the one I
removed in the earlier "config tests: simplify include cycle test"
change in this series won't happen again, and testing for this
explicitly in "env--helper"'s own tests.

This change breaks existing uses of
e.g. GIT_TEST_GETTEXT_POISON=YesPlease, which we've documented in
po/README and other places. As noted in [1] we might want to consider
also accepting "YesPlease" in "env--helper" as a special-case.

But as the lack of uproar over 6cdccfce1e ("i18n: make GETTEXT_POISON
a runtime option", 2018-11-08) demonstrates the audience for this
option is a really narrow set of git developers, who shouldn't have
much trouble modifying their test scripts, so I think it's better to
deal with that minor headache now and make all the relevant GIT_TEST_*
variables boolean in the same way than carry the "YesPlease"
special-case forward.

1. https://public-inbox.org/git/xmqqtvckm3h8.fsf@gitster-ct.c.googlers.com/

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-06-21 09:42:49 -07:00

307 lines
10 KiB
Plaintext

Core GIT Translations
=====================
This directory holds the translations for the core of Git. This document
describes how you can contribute to the effort of enhancing the language
coverage and maintaining the translation.
The localization (l10n) coordinator, Jiang Xin <worldhello.net@gmail.com>,
coordinates our localization effort in the l10 coordinator repository:
https://github.com/git-l10n/git-po/
The two character language translation codes are defined by ISO_639-1, as
stated in the gettext(1) full manual, appendix A.1, Usual Language Codes.
Contributing to an existing translation
---------------------------------------
As a contributor for a language XX, you should first check TEAMS file in
this directory to see whether a dedicated repository for your language XX
exists. Fork the dedicated repository and start to work if it exists.
Sometime, contributors may find that the translations of their Git
distributions are quite different with the translations of the
corresponding version from Git official. This is because some Git
distributions (such as from Ubuntu, etc.) have their own l10n workflow.
For this case, wrong translations should be reported and fixed through
their workflows.
Creating a new language translation
-----------------------------------
If you are the first contributor for the language XX, please fork this
repository, prepare and/or update the translated message file po/XX.po
(described later), and ask the l10n coordinator to pull your work.
If there are multiple contributors for the same language, please first
coordinate among yourselves and nominate the team leader for your
language, so that the l10n coordinator only needs to interact with one
person per language.
Translation Process Flow
------------------------
The overall data-flow looks like this:
+-------------------+ +------------------+
| Git source code | ---(1)---> | L10n coordinator |
| repository | <---(4)--- | repository |
+-------------------+ +------------------+
| ^
(2) (3)
V |
+------------------+
| Language Team XX |
+------------------+
* Translatable strings are marked in the source file.
* L10n coordinator pulls from the source (1)
* L10n coordinator updates the message template po/git.pot
* Language team pulls from L10n coordinator (2)
* Language team updates the message file po/XX.po
* L10n coordinator pulls from Language team (3)
* L10n coordinator asks the result to be pulled (4).
Maintaining the po/git.pot file
-------------------------------
(This is done by the l10n coordinator).
The po/git.pot file contains a message catalog extracted from Git's
sources. The l10n coordinator maintains it by adding new translations with
msginit(1), or update existing ones with msgmerge(1). In order to update
the Git sources to extract the messages from, the l10n coordinator is
expected to pull from the main git repository at strategic point in
history (e.g. when a major release and release candidates are tagged),
and then run "make pot" at the top-level directory.
Language contributors use this file to prepare translations for their
language, but they are not expected to modify it.
Initializing a XX.po file
-------------------------
(This is done by the language teams).
If your language XX does not have translated message file po/XX.po yet,
you add a translation for the first time by running:
msginit --locale=XX
in the po/ directory, where XX is the locale, e.g. "de", "is", "pt_BR",
"zh_CN", etc.
Then edit the automatically generated copyright info in your new XX.po
to be correct, e.g. for Icelandic:
@@ -1,6 +1,6 @@
-# Icelandic translations for PACKAGE package.
-# Copyright (C) 2010 THE PACKAGE'S COPYRIGHT HOLDER
-# This file is distributed under the same license as the PACKAGE package.
+# Icelandic translations for Git.
+# Copyright (C) 2010 Ævar Arnfjörð Bjarmason <avarab@gmail.com>
+# This file is distributed under the same license as the Git package.
# Ævar Arnfjörð Bjarmason <avarab@gmail.com>, 2010.
And change references to PACKAGE VERSION in the PO Header Entry to
just "Git":
perl -pi -e 's/(?<="Project-Id-Version: )PACKAGE VERSION/Git/' XX.po
Once you are done testing the translation (see below), commit the result
and ask the l10n coordinator to pull from you.
Updating a XX.po file
---------------------
(This is done by the language teams).
If you are replacing translation strings in an existing XX.po file to
improve the translation, just edit the file.
If there's an existing XX.po file for your language, but the repository
of the l10n coordinator has newer po/git.pot file, you would need to first
pull from the l10n coordinator (see the beginning of this document for its
URL), and then update the existing translation by running:
msgmerge --add-location --backup=off -U XX.po git.pot
in the po/ directory, where XX.po is the file you want to update.
Once you are done testing the translation (see below), commit the result
and ask the l10n coordinator to pull from you.
Testing your changes
--------------------
(This is done by the language teams, after creating or updating XX.po file).
Before you submit your changes go back to the top-level and do:
make
On systems with GNU gettext (i.e. not Solaris) this will compile your
changed PO file with `msgfmt --check`, the --check option flags many
common errors, e.g. missing printf format strings, or translated
messages that deviate from the originals in whether they begin/end
with a newline or not.
Marking strings for translation
-------------------------------
(This is done by the core developers).
Before strings can be translated they first have to be marked for
translation.
Git uses an internationalization interface that wraps the system's
gettext library, so most of the advice in your gettext documentation
(on GNU systems `info gettext` in a terminal) applies.
General advice:
- Don't mark everything for translation, only strings which will be
read by humans (the porcelain interface) should be translated.
The output from Git's plumbing utilities will primarily be read by
programs and would break scripts under non-C locales if it was
translated. Plumbing strings should not be translated, since
they're part of Git's API.
- Adjust the strings so that they're easy to translate. Most of the
advice in `info '(gettext)Preparing Strings'` applies here.
- If something is unclear or ambiguous you can use a "TRANSLATORS"
comment to tell the translators what to make of it. These will be
extracted by xgettext(1) and put in the po/*.po files, e.g. from
git-am.sh:
# TRANSLATORS: Make sure to include [y], [n], [e], [v] and [a]
# in your translation. The program will only accept English
# input at this point.
gettext "Apply? [y]es/[n]o/[e]dit/[v]iew patch/[a]ccept all "
Or in C, from builtin/revert.c:
/* TRANSLATORS: %s will be "revert" or "cherry-pick" */
die(_("%s: Unable to write new index file"), action_name(opts));
We provide wrappers for C, Shell and Perl programs. Here's how they're
used:
C:
- Include builtin.h at the top, it'll pull in gettext.h, which
defines the gettext interface. Consult with the list if you need to
use gettext.h directly.
- The C interface is a subset of the normal GNU gettext
interface. We currently export these functions:
- _()
Mark and translate a string. E.g.:
printf(_("HEAD is now at %s"), hex);
- Q_()
Mark and translate a plural string. E.g.:
printf(Q_("%d commit", "%d commits", number_of_commits));
This is just a wrapper for the ngettext() function.
- N_()
A no-op pass-through macro for marking strings inside static
initializations, e.g.:
static const char *reset_type_names[] = {
N_("mixed"), N_("soft"), N_("hard"), N_("merge"), N_("keep"), NULL
};
And then, later:
die(_("%s reset is not allowed in a bare repository"),
_(reset_type_names[reset_type]));
Here _() couldn't have statically determined what the translation
string will be, but since it was already marked for translation
with N_() the look-up in the message catalog will succeed.
Shell:
- The Git gettext shell interface is just a wrapper for
gettext.sh. Import it right after git-sh-setup like this:
. git-sh-setup
. git-sh-i18n
And then use the gettext or eval_gettext functions:
# For constant interface messages:
gettext "A message for the user"; echo
# To interpolate variables:
details="oh noes"
eval_gettext "An error occurred: \$details"; echo
In addition we have wrappers for messages that end with a trailing
newline. I.e. you could write the above as:
# For constant interface messages:
gettextln "A message for the user"
# To interpolate variables:
details="oh noes"
eval_gettextln "An error occurred: \$details"
More documentation about the interface is available in the GNU info
page: `info '(gettext)sh'`. Looking at git-am.sh (the first shell
command to be translated) for examples is also useful:
git log --reverse -p --grep=i18n git-am.sh
Perl:
- The Git::I18N module provides a limited subset of the
Locale::Messages functionality, e.g.:
use Git::I18N;
print __("Welcome to Git!\n");
printf __("The following error occurred: %s\n"), $error;
Run `perldoc perl/Git/I18N.pm` for more info.
Testing marked strings
----------------------
Even if you've correctly marked porcelain strings for translation
something in the test suite might still depend on the US English
version of the strings, e.g. to grep some error message or other
output.
To smoke out issues like these, Git tested with a translation mode that
emits gibberish on every call to gettext. To use it run the test suite
with it, e.g.:
cd t && GIT_TEST_GETTEXT_POISON=true prove -j 9 ./t[0-9]*.sh
If tests break with it you should inspect them manually and see if
what you're translating is sane, i.e. that you're not translating
plumbing output.
If not you should replace calls to grep with test_i18ngrep, or
test_cmp calls with test_i18ncmp. If that's not enough you can skip
the whole test by making it depend on the C_LOCALE_OUTPUT
prerequisite. See existing test files with this prerequisite for
examples.