Hunspell Guarani

l10n
#1

Hello everyone

I am currently developing a Guarani autocorrect in Hunspell. At the moment
I am having one final issue.

When testing in Hunspell, it won't detect words with g̃. Olivier filed a
bug for me and I also filed one, LO can't handle g̃ as well.

https://bugs.documentfoundation.org/show_bug.cgi?id=104195

I know this is a Character Set issue but I can't seem to solve it.

Adobe Latin-4 supports this

https://adobe-type-tools.github.io/adobe-latin-charsets/adobe-latin-4.html

https://en.wikipedia.org/wiki/G̃

If anyone wants to test it, you can download the files from here
https://mega.nz/#!v5pSAZaZ!Ec-sv4MIj3WQmzN1Y-rdsVVSVJRVfwItXceYb-PUH4Y

-Gio

#2

I noticed that there is no spell checker for Esperanto in LO.
(I use the extension mentioned at: https://wiki.documentfoundation.org/Language_support_of_LibreOffice).
It would be good if users could enable Esperanto spell checking in the same way that one can enable many other language.

What is the relationship between (Hunspell) spell checkers and extensions in the same language?
Esperanto is not mentioned on the Development/Dictionaries page: https://wiki.documentfoundation.org/Development/Dictionaries
or in the linked page:
https://cgit.freedesktop.org/libreoffice/dictionaries/tree/

I can supply a pair of Hunspell files (.aff and .dic) for Esperanto based on the Esperanto thesaurus at PuzzleFoundry.com produced in collaboration with Kris van der Merwe. The .dic file contains over 26000 words.

How should I proceed?

Donald

#3

Hi Donald,

I noticed that there is no spell checker for Esperanto in LO.
(I use the extension mentioned at:
https://wiki.documentfoundation.org/Language_support_of_LibreOffice).
It would be good if users could enable Esperanto spell checking in the
same way that one can enable many other language.

What is the relationship between (Hunspell) spell checkers and
extensions in the same language?
Esperanto is not mentioned on the Development/Dictionaries page:
https://wiki.documentfoundation.org/Development/Dictionaries
or in the linked page:
https://cgit.freedesktop.org/libreoffice/dictionaries/tree/

I can supply a pair of Hunspell files (.aff and .dic) for Esperanto
based on the Esperanto thesaurus at PuzzleFoundry.com produced in
collaboration with Kris van der Merwe. The .dic file contains over 26000
words.

How should I proceed?

You are right, there are many more dictionary extensions on various sites,
than "internal" dictionaries in LibreOffice source code.

For the internal dictionaries, the situation is different case by case.
Sometimes the dictionary has an upstream repository, and maintainer send
patch to LibreOffice or ask a developer to create a patch based on a new
release. In other cases the dictionary was committed once and it is
unmaintained in LibreOffice.

In case of Esperanto, I see at least two upstream versions on
https://wiki.documentfoundation.org/Language_support_of_LibreOffice
Did I understand you well, that you would like to submit a 3rd one?

Best regards,
Andras Timar

#4

Andras, thanks for your reply.

When we initiated the creation of the Esperanto and Afrikaans spelling data we were aware of one Esperanto implementation that was not actively maintained.
We am now aware of three Esperanto implementations.
So at this stage I will refrain from submitting a fourth Esperanto implementation.

Seeing that Esperanto has regular conjugations, there can be various approaches to the creation of the ".aff" conjugation rules.
A common flaw of the Esperanto spellcheck implementations is that they tend to produce false positives, especially when the ".aff" rules are complicated.
Maybe some evaluation of the various Esperanto spellcheck implementations is needed ...

Could you please help our participation by answering a few questions:
1) What is the "standard" way for an end-user to add/install a spellcheck language?
2) (How) Would an "upstream repository" become part of the "standard" install for an end-user?
3) What is a reasonable interval between the submission of spellcheck file updates?
4) Can these updates be submitted through the mailing list?

Regards
Donald

#5

Hi Donald,

1) What is the "standard" way for an end-user to add/install a spellcheck
language?

2) (How) Would an "upstream repository" become part of the "standard" install for an end-user?
3) What is a reasonable interval between the submission of spellcheck file updates?
4) Can these updates be submitted through the mailing list?

1) If a spell checker is part of LibreOffice, then it can be installed as

optional component (on Windows at least). Otherwise, users can always
install extensions themselves.
2) Basically if you submit .dic and .aff files, our point us to your
upstream repository, we can take the files from there.
3) LibreOffice bugfix releases come ~every month. More frequent updates
probably make little sense.
4) We prefer submission through gerrit. See
https://wiki.documentfoundation.org/Development/GetInvolved

Best regards,
Andras