1 files changed, 81 insertions, 0 deletions
diff --git a/po/README.txt b/po/README.txt
new file mode 100644
index 00000000000..cadcb7194e8
--- /dev/null
+++ b/po/README.txt
@@ -0,0 +1,81 @@
+Blender translation HOWTO
+=========================
+
+I'll try briefly explain how translation works and how to update translation files.
+
+1. How it works
+---------------
+
+This folder contains source files for translation system. These source files have
+got .po extension and they've got pretty simple syntax:
+
+msgid "some message id"
+msgstr "translation for this message"
+
+This means when string "some message id" is used as operator name, tooltip, menu
+and so it'll be displayed on the screen as "translation for this message".
+Pretty simple.
+
+This source files are pre-compiled into ../release/bin/.blender/locale/<language>/LC_MESSAGES/blender.mo,
+so they aren't getting compiled every time Blender is compiling to save some time and prevent
+failure on systems which don't have needed tools for compiling .po files.
+
+2. How to update translations
+-----------------------------
+
+It's also pretty simple. If you can find string you want to translate in <language>.po
+file as msgid, just write correct msgstr string for it. If msgid is marked as fuzzy,
+i.e.
+
+#, fuzzy
+msgid "some message id"
+msgstr "translation for this message"
+
+it means translation used to exist for this message, but message was changed, so translation
+also have to be updated (it's easier to make new translation based on previous translation).
+When translation was updated, remove line with '#, fuzzy' and it'll work.
+
+If there's no message in .po file you want to translate, probably .po file should be updated.
+Use the following steps for this:
+- With newly compiled blender run:
+  `blender --background --factory-startup --python update_msg.py`
+  to update messages.txt file (this file contains strings collected
+  automatically from RNA system and python UI scripts)
+- Run update_pot.py script which will update blender.pot file. This file contains all
+  strings which should be transated.
+- Run update_po.py script to merge all .po files with blender.pot (so all .po files
+  will contain all msgid-s declared in blender.pot) or update_po.py <language> to
+  update only needed .po file(s) to save time when you're busy with translation.
+  But before commit all .po files better be updated.
+
+When you've finished with translation, you should re-compile .po file into .mo file.
+It's also pretty simple: just run update_mo.py script to recompile all languages or
+just update_mo.py <language> to re-compile only needed language(s).
+
+NOTE: msgfmt, msgmerge and xgettext tools should be available in your PATH.
+
+These steps to update template, translation files and compile them can be made in "batch" mode
+using GNUMakefile:
+
+make -f GNUMakefile translations
+
+NOTE: Blender has to be compiled using GNUMakefile first.
+
+
+3. Note for Windows users
+-------------------------
+You can find compiled builds of gettext in the lib folder under "binaries\gettext\" for both windows and win64.
+In order to run the scripts you will need to replace the location of the GETTEXT_..._EXECUTABLE.
+
+For example in update_pot.py:
+-GETTEXT_XGETTEXT_EXECUTABLE = "xgettext"
++GETTEXT_XGETTEXT_EXECUTABLE = "C:\\Blender\\lib\\\windows\\\binaries\\\gettext\\xgettext.exe"
+
+4. Other scripts
+----------------
+
+- check_po.py: this script checks if all messages declared in blender.pot exists in.po files
+               and that no extra messages are declared in .po files
+- clean_po.py: this script removes all commented messages which aren't required by .pot file anymore.
+- merge_po.py: this script accepts two files as arguments and copies translations from second file
+               into first file.