Currently we have: * An initial digit marks the whole word as not checked: "1eror" is not underlined * An ending digit doesn't, and the resulting word is always wrong: "error1" is underlined
I would like the spellchecker to split the words on digits, to have them behave like punctuation (like ".eror" and "error.")
Sounds reasonable. Though after having implemened special handling for ' (apostrophe) and underlines, I think we should just make all this configurable and also provide an option to configure the wordchars for spell checking at all.
Will do but might take some time.
github-comments@lists.geany.org