In the latest beta (18.104.22.168), it seems the Check Spelling as You Type feature is closer to being fully functional than previously. Keep up the good work! However, there are a couple of strange behaviours I’ve noticed:
First, it doesn’t recognise the word “reminds” which leads me to think that the parser that tries to understand different forms of words isn’t working fully. I haven’t noticed this behaviour with other words but presumably there will be some.
Second, and more importantly, it is not possible to Learn Spelling for some words flagged as misspelled that end apostrophe “s” (e.g. in my document: Monty’s, Tussaud’s, Eliot’s). When I right-click any of these words the suggested spelling is the word with the apostrophe as a question mark (e.g. Monty?s). Although Scrivener will not learn the spelling–when you click “Learn Spelling” it clearly doesn’t learn it and closes the pop-up–you can make it “Ignore Spelling” successfully. It behaves normally for words already in the dictionary (e.g. child’s) but not words whose base form have been added to custom spellings (e.g. I’ve added the name Monty to my custom dictionary but Monty’s still causes the behaviour).
I should say that I am using the UK English dictionary and smart quotes are on.
Thanks for the reports, Arthur! The issue with Learn Spelling not working with the smart apostrophes seems to be an unfortunate new flavour of a bug we’ve already filed regarding some special characters showing in the suggestions list (and subsequently being used for the replacement). Using straight apostrophes instead of smart will fix this, although it means when you select the word from the suggestions, it will use a straight mark instead of curly. If you’re not needing to do that often, though, this might still be the best solution for present, as it will keep the correctly typed words from being underlined even when they’re using a smart quote.
The easiest way to do this may be to just open the personal word list from the Corrections pane in Options and add all the words you need there via the “+” button; this will use straight apostrophes, and once you’ve added them, you should find the misspelling underlines removed in your editor. If you already have many added, you could instead open the wordlist.ini file in a plain-text editor (Notepad will do) and use a find/replace to swap out your curly apostrophes for straight ones, then resave the file. You’ll find the wordlist in %LOCALAPPDATA%\LiteratureAndLatte\Scrivener. (You can just copy and paste that into the address bar of File Explorer.)
As for the limitations of the UK English dictionary, we’re looking into our options here. The dictionary itself we don’t create but download from a set of Hunspell dictionaries; however, we may be able to replace it with a different one or, in the long term, provide a way for users to add their own dictionaries. There are a lot of technical hurdles with this, so even if it ends up something we’re able to do, it may not be for a while yet, but we are investigating.
It would be nice if there were a user-made dictionary option inside Scrivener itself, but if it’s not as simple as it sounds (and I’ve programmined with QT a little; I understand it’s often not), that’s okay. I think Python could do the necessary juggling. The process of replacing a dictionary (fooling Scrivener, essentially) is really not complex, and the fiddly bits you have to get right when replacing dictionaries a well-written script could handle without issue.
Just noticed that the issue with smart quotes/apostrophes is still present in beta 10. You probably realise that, but I it’s on your “to do” list, as it’s a very distracting bug (despite being very minor). Thanks.