Using more than one dictionary at the same time

[KevinH]

For those that are interested and build from source themselves, I am pushing most of this work to my own github site:

https://github.com/kevinhendricks/Sigil

As of right now the following has been implemented (roughly)

1. User can now specify a secondary dictionary for inline red squiggley spellchecking and that should work.

2. Ignore word lists are now cross language and no longer added to any dictionary

3. The SpellCheckEditor dialog, should now happily spellcheck words in multiple languages as determined by the lang and xml:lang attributes on html tags. It does use a simplistic gui where language code is used as a word prefix. That will most likely be changed before any release to use a separate language column. But for now, the simple gui is being used.

4. Updating words from suggestions should now work in the SpellcheckEditor dialog as well.

5. All UserDictionaries have been temporarily disabled until we figure out how to use them in a multi language environment more effectively.

Feel free to check it out and play with it and let me know of any bugs/issues/crashes that happens.

Once we figure out the best way to handle UserDictionaries, I will push everything to Sigil master for testers to play with the Travis and Appveyor CI builds if they so desire.

[KevinH]

An alternative idea that might be supported with very few changes is to make Sigil load any .dic_delta files automagically when it loads the dictionary for that language (that match the language region code) and have this be unrelated to UserDictionaries concept completely.

So UserDictionaries stay as they are now and are only added to the Primary Language dictionary (like now) as people may want to use these in a different way.

But more official .dic_delta files (one per language) would automatically be used if they exist.

It might be a good compromise approach.

[Doitsu]

Originally Posted by KevinH

An alternative idea that might be supported with very few changes is to make Sigil load any .dic_delta files automagically when it loads the dictionary for that language (that match the language region code) and have this be unrelated to UserDictionaries concept completely.

IMHO, this is a good compromise.

[KevinH]

Should the .dic_delta files need to be right beside the .aff and .dic files? I assume that is how Chrome needs them?

[Doitsu]

Originally Posted by KevinH

Should the .dic_delta files need to be right beside the .aff and .dic files? I assume that is how Chrome needs them?

Ideally, yes, but since the main dictionaries are in the program folder, it might be difficult to gain write access.

[KevinH]

So first look right beside the hunspell .aff and .dic and if not found look for something with the exact same basename with the extension .dic_delta in the Sigil Preferences folder in "/hunspell_dictionaries/";

That should work right?

[Doitsu]

Originally Posted by KevinH

So first look right beside the hunspell .aff and .dic and if not found look for something with the exact same basename with the extension .dic_delta in the Sigil Preferences folder in "/hunspell_dictionaries/";

That should work right?

Yes. That should work.

[KevinH]

I gave preference to the Sigil Prefs version so the user can overwrite any system level .dic_delta files.

I have now re-enabled UserDictionaries and added support for .dic_delta file loading.

I just now pushed all of this to master. So both travis and appveyor CI automated builds should have these features soon.

Please note, these new features are in quite rough form with little to no testing and many things could be broken. All testing and feedback is welcome.

Once we get the major bugs out, I can start work on improving the gui elements in the SpellcheckEditor.

Thanks

[KevinH]

I had some time so Sigil master and appveyor/travis should soon/now have an updated gui for the SpellcheckEditor dialog as well.

[KevinH]

Has anyone interested in multi-language spellchecking had a chance to either build their own or use one of the continuous improvement test builds created by Appveyor (Windows) or Travis (Mac)?

If so, I would love to hear feedback, bug reports, etc.

If someone is willing to test and report back on these features, please let me know and I can PM you with a download link for your platform.

ps.
I added some screen shots from my build so people can see how things look in master now.

[1v4n0]

Originally Posted by Tex2002ans

Do you recall Post #31 (and my Post #32)?

Getting amnesia? It was only a few weeks ago!

Gave those posts a quick look. Sorry, but it seems to me those solutions won't work for my needs. I can't bulk-tag all the words the spellcheck doesn't recognize, because although the majority on the positives are indeed just correct words in another language, some of them are errors, and the whole purpose of scanning the list of the words which the spellcheck doesn't recognize is indeed to identify the errors. And I need to be able to scan the list more times, so I can't just browse it once and tag all the words, because that way I'll end up adding words that are indeed errors. The example for merging UK and US English makes a bit more sense, because the words that appear in both lists will provide, if nothing else, a shorter list to check for real-negatives. But:
1) How do I quickly identify those words?
2) It remains a very long and complicated process.

Thank you for your answers anyway