Enchant - enchant language tags, ordering files and personal word lists
Enchant identifies dictionaries by their language tag. A language tag is typically an IETF BCP 47 language tag of the form LANGUAGE_COUNTRY; for example, en_US or zh_SG. Multiple dictionaries may be used together by giving a comma-separated list; for example en_GB,fr_FR. This can be useful for checking multi-lingual text, or for using specialised word lists for a particular subject, project or document (these might have names such as en-medical or en-my-novel).
Enchant uses global and per-user ordering files named enchant.ordering to decide which spelling provider to use for particular languages. The per-user file takes precedence.
The ordering file takes the form language_tag:<comma-separated list of spelling providers>. To see what dictionaries are available, run enchant-lsmod-2. ’*’ is used to mean “use this ordering for all languages, unless instructed otherwise.” For example:
*:aspell,hunspell,nuspell
en:aspell,hunspell,nuspell
en_GB:hunspell,nuspell,aspell
fr:hunspell,nuspell,aspell
Personal word lists are simple plain text files with one word per line. The name of the file starts with the language tag and ends .dic. Each personal word list has a corresponding exclude file, ending in .exc, which lists words that are found in the dictionary but that the user wants to be considered invalid. The files are stored in an Enchant configuration directory; see FILES AND DIRECTORIES below. Lines starting with a hash sign ’#’ are ignored.
SHARING
PERSONAL WORD LISTS BETWEEN SPELL-CHECKERS
It is possible, and usually safe, to share Enchant’s
personal word lists with other spelling checkers that use
the same format (note that other spell-checkers may not
support comments!). The spell-checkers known to be
compatible are Hunspell, Nuspell and Ispell. (Although
Enchant does not support Ispell as a provider, it’s
still fine to share word lists with it.) Other
spell-checkers supported by Enchant are either incompatible,
or have no personal word list mechanism. There may well be
yet other spell-checkers, unknown to Enchant, that use the
same format.
Some applications use Hunspell or Nuspell, but store the personal word list under another name or in another location; Firefox and Thunderbird do this. Firefox also seems to reorder its word list when updating it; again, this is OK, as the result is still in the same format.
To share word lists with Enchant, find the other spelling checker’s word list file, e.g. ~/.hunspell_fr_FR or ~/.config/nuspell/fr_FR, and merge it with the corresponding Enchant file, in this case ~/.config/enchant/fr_FR.dic. Use the following command, replacing ENCHANT-DICT and OTHER-DICT with the corresponding dictionary file names:
cat ENCHANT-DICT OTHER-DICT | sort -u > merged.txt
Take a look at merged.txt to check the merge has worked, then
mv merged.txt
ENCHANT-DICT
rm OTHER-DICT
ln -s OTHER-DICT ENCHANT-DICT
to replace the other dictionary file with a link to the Enchant dictionary, again filling in the name of the dictionary files.
The following
variables affect the behavior of Enchant:
ENCHANT_CONFIG_DIR
A directory in which Enchant should look for configuration files. See below.
G_MESSAGES_DEBUG
Enchant uses GLib’s log functions, with the domain libenchant, to output messages useful for debugging. Setting G_MESSAGES_DEBUG to libenchant will cause Enchant to output debugging messages to standard error. See the GLib documentation for more details.
Enchant looks
in the following places for user files, in decreasing order
of precedence:
ENCHANT_CONFIG_DIR
(If the environment variable is set.)
XDG_CONFIG_HOME/enchant (non-Windows systems)
Default: ~/.config/enchant
CSIDL_LOCAL_APPDATA\enchant (Windows systems)
Default: C:\Documents and Settings\username\Local Settings\Application Data\enchant
Dictionaries for some providers are looked for in a subdirectory with the same name as the provider, for example ~/.config/enchant/hunspell. Currently this works for Hspell, Hunspell, Nuspell and Voikko.
Providers also look in specific system directories, and in some cases and user directories, for their dictionaries; see the documentation for each provider.
In addition,
Enchant looks in the following systems directories for
ordering files:
/home/rrt/.local/etc/enchant-2
(Or the equivalent location relative to the enchant library for a relocatable build.)
/home/rrt/.local/share/enchant-2
(Or the equivalent location relative to the enchant library for a relocatable build.)
enchant-2(1), enchant-lsmod-2(1)
Written by Dom Lachowicz and Reuben Thomas.