Page 38 of 38

Posted: Wed Jun 22, 2022 10:13 am
by Diamen
I made a little video (now deleted).
Light underlining disappears later carriage return but not for "รจ" accented. Also accented problem in contextual menu suggestion.
Ty.

Posted: Wed Jun 22, 2022 10:40 am
by random6_2020
Diamen: I see.
First try to check what encodings are? What is the dictionary encoding? (Should be utf-8 as the plugin doesn't check what it actually is currently.)
Does anything change if you change the document encoding to utf-8?

Posted: Wed Jun 22, 2022 11:01 am
by Diamen
I have txt file ANSI encode, and also *.dic and *.aff.
Switch txt file to UTF-8 solve it.
Do it is not possible to check accented in ANSI text?

Other problem if I have more dictionary, change in setting not save and also check not adapt.
Also manually set "Lang=" in "Options" of "SpellCheck.ini" not work.
Plugin set every the last dictionary in alphabetic order.

Posted: Wed Jun 22, 2022 5:19 pm
by random6_2020
Here is version that tries to check dictionary encoding, allow switching dictionaries an re-enabled other underline styles:
SpellChecker 1.3.1 x64 | Source | RU+EN merged dictionary

Posted: Wed Jun 22, 2022 5:53 pm
by Diamen
Now it save lang but...
Spelling not work in Background, Document, Selection or Word.
"Word not found."
"No missspelling found."

Also Hunspell field in setting is empty but "CorePath" in SpellCheck.ini have path.

Posted: Thu Jun 23, 2022 9:44 am
by random6_2020
Some original code considers that all dictionaries are loaded at once. So dynamic dict loading on GetDict won't work, I guess.
Hunspell supports add_dict something. But it doesn't load affixes for that additional dict... strange.
Also it's probably possible to ignore words in non-present languages using character list defined in affixes file somehow.

"Word not found" is just renamed "Error" message from before - can be disabled in Settings.

But for now just use a single dictionary and utf-8.

Posted: Thu Jun 23, 2022 4:10 pm
by Diamen
random6_2020 wrote:But for now just use a single dictionary and utf-8.
it no longer works with a single dictionary and utf-8.
I suspect path not loading with Hunspell filed empty in setting.

Posted: Thu Jun 23, 2022 4:29 pm
by random6_2020
Works for me so can't confirm, try deleting SpellCheck.ini or regedit settings and reconfigure the plugin from its Settings dialog. The .ini should mention your dictionary name without extension after that.

Posted: Thu Jun 23, 2022 4:41 pm
by Diamen
If I delete SpellCheck.ini it work only with single dictionary.
When check document if I click accept it delete word instead of replace.

Posted: Fri Jun 24, 2022 10:45 am
by random6_2020
I have updated the plugin in the post above. Should work properly with multiple dictionaries now.
Manual document checking is still done only in single language configured in Settings, though.

IDK what's with the encoding, can't localize the problem.
The same goes for bug where pressing Accept removes the word.

Posted: Fri Jun 24, 2022 3:35 pm
by Diamen
Multi language selection work.
When check document or selection if I click accept it delete word instead of replace. No problem instead with "word".

Posted: Mon Jun 27, 2022 6:09 pm
by random6_2020
Moved downloads to a separate topic.

In that version words starting from extended Unicode chars (Japanese/Chinese/etc) are ignored and it should properly check the dictionary encoding now.

The other stuff is still unidentifiable.