Closed szepeviktor closed 2 weeks ago
Yeah I had a branch with aspell in make check at some point but it gave soo many false positive that I never merged it.
if hunspell is open source, packaged in Debian and we have a way to limit false positive then I am all for it adding it back to the makefile.
Language is en-US-UTF-8
Thanks
@cedounet You can get dictionaries from LibreOffice/dictionaries GitHub repo,
Hunspell from apt-get install hunspell
,
then hunspell -d path/to/libreoffice-dictionaries/en/en_US -l
with optional local dictionary: -p hunspell.txt
(1 word per line).
Thanks.
how many false positive when ran accross all tex files and eventually comments in code ?
I've used typos
in #375.
cat $(git ls-files '*.tex')|sed 's/\\[a-z{}]\+//g'|../hunspell-en_US|sort -f|uniq -i|grep -v '^[0-9]'
300 words including proper nouns, non-English words, and British spelling.
I have been using Grammarly in my IDE.
Please use a language tool when writing text for a book. e.g. DeepL, Hunspell or Grammarly