On 03/08/2017 02:39 PM, Andrew Toskin wrote:
Repost of my question on Ask Fedora --
https://ask.fedoraproject.org/en/question/101943/how-to-add-words-to-syst...
I'm trying to figure out how to add words to a spelling dictionary so that I get the
most impact for the effort --- some way to have it apply to all applications, *or at least
as many of the ones I actually use* as possible.
If I have to target specific spell checkers for specific applications, then I'm
mostly interested in adding words to the spell checking dictionaries used by the
applications I most often compose text with: Firefox for web forms, Evolution for email,
Atom and gedit for code and plain text, and LibreOffice for office documents. Each of
these applications has a way to add words to their own spelling dictionary, but those
dictionaries are application-specific, and I would have to add a given word to all of
them. So I've been hunting to see If they share a single spell-check system, or even
if there's 2 or 3 systems that covers all of those. Here's what I've found out
so far:
I notice my system in particular has these packages which match "spell" in the
name, plus enchant:
* aspell - 12:0.60.6.1-14.fc25
* gnome-python2-gtkspell - 2.25.3-48.fc25
* gspell - 1.2.3-1.fc25
* gtkspell - 2.0.16-11.fc24
* gtkspell3 - 3.0.9-1.fc25
* hunspell - 1.4.1-1.fc25
* hunspell-en - 0.20140811.1-5.fc24
* hunspell-en-GB - 0.20140811.1-5.fc24
* hunspell-en-US - 0.20140811.1-5.fc24
* enchant - 1:1.6.0-14.fc25
* python3-enchant - 1.6.8-1.fc25
Wikipedia says hunspell replaces myspell and is widely used. `rpm --query --requires`
says that Firefox and LibreOffice both depend on hunspell. gedit uses gspell, which relies
on enchant, which can use myspell as a backend --- and the directory `/usr/share/myspell/`
is actually owned by hunspell. Atom's spell-checking package searches in the myspell
directory too, so it seems like I *should* have a good chance of doing everything I want
to do from the hunspell configuration. So I tried creating a user addon dictionary by
creating the file `~/.hunspell_en_US` containing a few line-separated words (all lower
case except for proper nouns, no /flags or anything). However, those words are still
marked as a spelling error in all the applications I'm testing on, *except* for
hunspell itself, via the command line tool.
Why isn't it working?
According to the man page for hunspell (specifically the bit describing
the "-p" option), the default user dictionary should be called
"$HOME/.hunspell_default" unless you are using hunspell from the
command line and specifying "-p", "-d" or setting the DICTIONARY
environment variable.
My guess is that your dictionary isn't being referenced. Try renaming
it as shown above and try again. Also, "hunspell -D" will show you which
directories are searched, which dictionaries are available and which are
actually being used (based on your locale).
----------------------------------------------------------------------
- Rick Stevens, Systems Engineer, AllDigital ricks(a)alldigital.com -
- AIM/Skype: therps2 ICQ: 226437340 Yahoo: origrps2 -
- -
-"Jimmie crack corn and I don't care." What kind of a lousy attitude -
- is THAT to have, huh? -- Dennis Miller -
----------------------------------------------------------------------