Developer Documentation 👩🔧
Contents
- Reporting bugs 🐞
- Translation 🇺🇳
- Mailing list 📬
- git repository 🐱
- Releases
- Building and installing from source 🛠️
- Running from a source tree 🌳🏃♀️➡️
- Distribution specific stuff for building, installing, and running
Reporting bugs 🐞
You can report bugs and suggest improvements in the github issue tracker: https://github.com/mike-fabian/ibus-typing-booster/issues
Translation 🇺🇳
You can contribute translations for ibus-typing-booster using the online translation platform weblate: https://translate.fedoraproject.org/projects/ibus-typing-booster/.
If you cannot use weblate, you can also create a pull request for new or updated translations.
Mailing list 📬
https://lists.fedorahosted.org/admin/lists/ibus-typing-booster.lists.fedorahosted.org/
git repository 🐱
You can browse the source code at https://github.com/mike-fabian/ibus-typing-booster.
Anonymous git clone
$ git clone git://github.com/mike-fabian/ibus-typing-booster.git
Authorized git clone
$ git clone git@github.com:mike-fabian/ibus-typing-booster.git
Releases
Release tarballs are available at https://github.com/mike-fabian/ibus-typing-booster/releases.
Building and installing from source 🛠️
When using a git checkout, go to the directory of the checkout and use autogen.sh:
$ cd ibus-typing-booster
$ ./autogen.sh
autogen.sh automatically runs configure and you can give the same arguments to autogen.sh you can give to configure. See below for the right arguments for your distribution.
When using the source from a release tarball, unpack the tarball and go into the source directory:
$ tar xvf ibus-typing-booster-2.27.53.tar.gz
$ cd ibus-typing-booster-2.27.53
Now continue like this:
$ ./configure --prefix=/usr --libexecdir=/usr/libexec/ # on Fedora
# ./configure --prefix=/usr --libexecdir=/usr/lib/ibus/ # on most other distributions
$ make
$ make check # optional
$ make install
For all distributions: please use --prefix=/usr and not the
default /usr/local, installing into /usr/local will usually not
work!
And use the correct --libexecdir option for your distribution, the
default is --libexecdir=/usr/libexec which is correct for Fedora,
but on openSUSE and Debian based distributions and Alpine Linux it is
--libexecdir=/usr/lib/ibus!
Build requirements
To build ibus-typing-booster, you will need at least (Package names may differ depending on your distribution!):
- autoconf
- automake
- gettext-devel
- python3-devel >= 3.6
- ibus-devel >= 1.5.3
If you build from a git checkout and not from a release tarball, you also need:
- autopoint (might already be in the gettext-devel package)
If you also want to run the test-suite with “make check”, you also need the following packages:
- m17n-lib
- m17n-db-extras
- python3-enchant or pyhunspell-python3
- hunspell-cs
- hunspell-de
- hunspell-en
- hunspell-es
- hunspell-it
- hunspell-ko
Now you can build and install.
Runtime requirements
To run ibus-typing-booster, the following software is required or
optional for additional features (Package names may differ depending
on your distribution!):
Required
- ibus >= 1.5.3
- m17n-lib
- m17n-db
- python3 >= 3.6
- python3-dbus
- python3-pyxdg
- python3-packaging
Optional
python3-enchant or pyhunspell-python3: Needed if you want spell-checking suggestions to work. Both packages work equally well, it doesn’t matter which one you choose.
hunspell-*: Hunspell dictionaries for the languages you want to type. If they are not there, ibus-typing-booster will still work but only learn from user input. With the hunspell dictionaries and the above Python modules to use them one will also get spell-checking suggestions and if no good suggestions can be found in the data gathered from user input, the word lists from the hunspell dictionaries will be used as a fallback to offer at least some suggestions.
python3-rapidfuzz: Makes the matching of emoji and Unicode characters faster.
Fonts: Install fonts for all language you need. Installing a nice font for emoji is also recommended.
m17n-db-extras: for some additional Japanese and Chinese input methods.
m17n-lib-anthy: if you want to type Japanese using
ja-anthy.pykakasi: Only needed if you want to input emoji by typing their Japanese names in romaji (i.e. Latin transliteration). That is an alternative to using the Japanese input method “ja-anthy” to type the names of the emoji in hiragana. There is no “pykakashi” package for Fedora, if you want this feature you need to install it from source.
pinyin: A Python module to convert Chinese text into pinyin. This is only needed if you want to type emoji by typing the Chinese names of the emoji in pinyin Latin transliteration. There is no “pinyin” for Fedora, if you want this feature you need to install it from source.
Running from a source tree 🌳🏃♀️➡️
To run Typing Booster directly from the source directory (unpacked tar or git clone), follow these steps:
Build
Build as usual, just omit the make install at the end.
Check and Kill Any Running Instances
Before starting, ensure no other Typing Booster process is active. You can check and terminate it using one of these methods:
Using pgrep and pkill:
$ pgrep -u ${USER} -a -f ibus-typing-booster/engine/main.py
2803020 /usr/bin/python3 /usr/share/ibus-typing-booster/engine/main.py --ibus
$ pkill -u ${USER} -f ibus-typing-booster/engine/main.py
Using ps and kill:
$ ps aux | grep ibus-typing-booster/engine/main.py
mfabian 2803020 8.1 1.4 2617344 474388 ? SLl 23:33 0:04 /usr/bin/python3 /usr/share/ibus-typing-booster/engine/main.py --ibus
$ kill 2803020
Start Typing Booster from Source
From the top-level source directory, run:
$ python3 ./engine/main.py --profile
(Alternatively, python3 ../ibus-typing-booster/engine/main.py --profile works too. Using the full path makes it easier to grep for
the process later.)
Switch to Typing Booster in IBus
First, switch to any other IBus engine (e.g., a plain keyboard layout) using
Super+spaceor the system panel.Then, switch back to Typing Booster. Since you already started it manually, IBus will reuse your local instance instead of launching the system-wide version.
Notes:
--profileflag (optional): Enables profiling, logging performance data to~/.local/share/ibus-typing-booster/debug.logwhen the process exits.Debug logging: To monitor logs in real-time:
$ tail -F ~/.local/share/ibus-typing-booster/debug.logFilter output for specific functions (e.g., key events or preedit updates):
$ tail -F ~/.local/share/ibus-typing-booster/debug.log | grep -E 'do_process_key_event|_update_preedit'Debug level: Adjust verbosity in the setup tool—higher levels log more details.
New gsettings options: If you added any new gsettings options (rare), the schema file needs to be installed and compiled, otherwise Typing Booster and the setup tool trying to use the new options will fail:
$ mkdir -p ~/.local/share/glib-2.0/schemas $ cp org.freedesktop.ibus.engine.typing-booster.gschema.xml ~/.local/share/glib-2.0/schemas $ glib-compile-schemas ~/.local/share/glib-2.0/schemas
Distribution specific stuff for building, installing, and running
Fedora
Build requirements
To build ibus-typing-booster, you will need at least:
- autoconf
- automake
- gettext-devel
- python3-devel >= 3.6
- ibus-devel >= 1.5.3
If you build from a git checkout and not from a release tarball, you also need:
- autopoint (is already in the gettext-devel package)
If you also want to run the test-suite with “make check”, you also need the following packages:
- m17n-lib
- m17n-db-extras
- python3-enchant or pyhunspell-python3
- hunspell-cs
- hunspell-de
- hunspell-en
- hunspell-es
- hunspell-it
- hunspell-ko
Now you can build and install.
Runtime requirements
Required
- ibus >= 1.5.3
- m17n-lib
- m17n-db
- python3 >= 3.6
- python3-dbus
- python3-pyxdg
- python3-packaging
Optional
python3-enchant or pyhunspell-python3: Needed if you want spell-checking suggestions to work. Both packages work equally well, it doesn’t matter which one you choose.
hunspell-*: Hunspell dictionaries for the languages you want to type. If they are not there, ibus-typing-booster will still work but only learn from user input. With the hunspell dictionaries and the above Python modules to use them one will also get spell-checking suggestions and if no good suggestions can be found in the data gathered from user input, the word lists from the hunspell dictionaries will be used as a fallback to offer at least some suggestions.
python3-rapidfuzz: Makes the matching of emoji and Unicode characters faster.
Fonts: Install fonts for all language you need. Installing a nice font for emoji is also recommended.
m17n-db-extras: for some additional Japanese and Chinese input methods.
m17n-lib-anthy: if you want to type Japanese using
ja-anthy.pykakasi: Only needed if you want to input emoji by typing their Japanese names in romaji (i.e. Latin transliteration). That is an alternative to using the Japanese input method “ja-anthy” to type the names of the emoji in hiragana. There is no “pykakashi” package for Fedora, if you want this feature you need to install it from source.
pinyin: A Python module to convert Chinese text into pinyin. This is only needed if you want to type emoji by typing the Chinese names of the emoji in pinyin Latin transliteration. There is no “pinyin” for Fedora, if you want this feature you need to install it from source.
openSUSE
Build requirements
To build ibus-typing-booster, you will need at least:
- autoconf
- automake
- python3-devel >= 3.6
- ibus-devel >= 1.5.3
If you build from a git checkout and not from a release tarball, you also need:
- autopoint
If you also want to run the test-suite with “make check”, you also need the following packages:
- m17n-lib
- m17n-db
- python3-pyenchant
- myspell-cs_CZ
- myspell-de
- myspell-de_DE
- myspell-en
- myspell-en_US
- myspell-es
- myspell-es_ES
- myspell-it_IT
- python3-gobject
- dbus-1-x11
Now you can build and install
but remember to use --libexecdir=/usr/lib/ibus/ when running ./configure or
./autogen.sh!
Runtime requirements
Required
- ibus >= 1.5.3
- m17n-lib
- m17n-db
- python3 >= 3.6
- dbus-1-python3
- python3-pyxdg
- python3-packaging
Optional
python3-pyenchant: Needed if you want spell-checking suggestions to work.
myspell-*: Hunspell dictionaries for the languages you want to type. If they are not there, ibus-typing-booster will still work but only learn from user input. With the hunspell dictionaries and the above Python modules to use them one will also get spell-checking suggestions and if no good suggestions can be found in the data gathered from user input, the word lists from the hunspell dictionaries will be used as a fallback to offer at least some suggestions.
Fonts: Install fonts for all language you need. Installing a nice font for emoji is also recommended.
pykakasi: Only needed if you want to input emoji by typing their Japanese names in romaji (i.e. Latin transliteration). That is an alternative to using the Japanese input method “ja-anthy” from m17n-db-extras to type the names of the emoji in hiragana. There is no “pykakashi” package for Fedora, if you want this feature you need to install it from source.
pinyin: A Python module to convert Chinese text into pinyin. This is only needed if you want to type emoji by typing the Chinese names of the emoji in pinyin Latin transliteration. There is no “pinyin” for Fedora, if you want this feature you need to install it from source.
Ubuntu
Build requirements
To build ibus-typing-booster, you will need at least:
- autoconf
- automake
- python3 >= 3.6
- ibus >= 1.5.3
- libibus-1.0-dev >= 1.5.3
- make
- gcc
If you build from a git checkout and not from a release tarball, you also need:
- autopoint
If you also want to run the test-suite with “make check”, you also need the following packages:
- libm17n-0
- m17n-db
- python3-enchant
- hunspell-de-de
- hunspell-cs
- hunspell-es
- hunspell-it
- hunspell-ko
Now you can build and install
but use --libexecdir=/usr/lib/ibus/ when running ./configure or
./autogen.sh!
Runtime requirements
Required
- ibus >= 1.5.3
- libm17n-0
- m17n-db
- python3 >= 3.6
- python3-dbus
- python3-xdg
- python3-packaging
Optional
python3-enchant : Needed if you want spell-checking suggestions to work.
hunspell-*: Hunspell dictionaries for the languages you want to type. If they are not there, ibus-typing-booster will still work but only learn from user input. With the hunspell dictionaries and the above Python modules to use them one will also get spell-checking suggestions and if no good suggestions can be found in the data gathered from user input, the word lists from the hunspell dictionaries will be used as a fallback to offer at least some suggestions.
Fonts: Install fonts for all language you need. Installing a nice font for emoji is also recommended.
pykakasi: Only needed if you want to input emoji by typing their Japanese names in romaji (i.e. Latin transliteration). That is an alternative to using the Japanese input method “ja-anthy” from m17n-db-extras to type the names of the emoji in hiragana. There is no “pykakashi” package for Fedora, if you want this feature you need to install it from source.
pinyin: A Python module to convert Chinese text into pinyin. This is only needed if you want to type emoji by typing the Chinese names of the emoji in pinyin Latin transliteration. There is no “pinyin” for Fedora, if you want this feature you need to install it from source.
Alpine Linux
Build requirements:
To build ibus-typing-booster, you will need at least:
- autoconf
- automake
- python3 >= 3.6
- gettext
- gettext-dev
- ibus
- ibus-dev
- make
- gcc
If you build from a git checkout and not from a release tarball, you also need:
- autopoint (is already in the gettext package)
If you also want to run the test-suite with “make check”, you also need the following packages:
- libm17n-core
- libm17n-flt
- m17n-db
- m17n-db-dev
- m17n-lib
- m17n-lib-dev
- hunspell-en
- hunspell-de-de
- py3-enchant
Now you can build and install
but use --libexecdir=/usr/lib/ibus/ when running ./configure or
./autogen.sh!
Runtime requirements
Required
- ibus >= 1.5.3
- libm17n-core
- libm17n-flt
- m17n-db
- python3 >= 3.6
- py3-dbus
- py3-xdg
- py3-packaging
Optional
py3-enchant : Needed if you want spell-checking suggestions to work.
hunspell-*: Hunspell dictionaries for the languages you want to type. If they are not there, ibus-typing-booster will still work but only learn from user input. With the hunspell dictionaries and the above Python modules to use them one will also get spell-checking suggestions and if no good suggestions can be found in the data gathered from user input, the word lists from the hunspell dictionaries will be used as a fallback to offer at least some suggestions.
Fonts: Install fonts for all language you need. Installing a nice font for emoji is also recommended.
pykakasi: Only needed if you want to input emoji by typing their Japanese names in romaji (i.e. Latin transliteration). That is an alternative to using the Japanese input method “ja-anthy” from m17n-db-extras to type the names of the emoji. There is no “pykakashi” package for Alpine Linux, if you want this feature you need to install it from source or using pip.
pinyin: A Python module to convert Chinese text into pinyin. This is only needed if you want to type emoji by typing the Chinese names of the emoji in pinyin Latin transliteration. There is no “pinyin” for Fedora, if you want this feature you need to install it from source or using pip.