[libreoffice-l10n] Understanding the hunspell

Hi,

I'm not sure if I'm doing it right by sending this mail do l10n, but 
hopping someone will point me to right direction. Even tried to 
subscribe to hunspell mailing list, but with no success.

I've done reading Hunspell help/documentation pdf from the internet but 
still have few questions. From what stands in that document I can't be 
certain in some specific things which are actually very important when 
we talk about morphology of highly flective language like Croatian.

Here we go:

Let's say that .aff contain:

FLAG long
SFX Y A1 1
SFX Y 0 a .

FLAG long
SFX Y A2 1
SFX Y A2 0 e .

Now, A1 and A2 needs to be applied to word "jezik" (eng. language). So 
if .dic file contain "jezik/A1" I should expect forms like "jezik" and 
"jezika". If .dic file contain entery like "jezik/A2" expected forms are 
"jezik", "jezike", but if the entery is

    jezik/A1A2

I can expect forms like "jezik", "jezika", "jezike", but will hunspell 
try to combine A1 and A2 together like "jezikae" or "jezikea"? Form 
"jezikae" and "jezikea" are not valid in Croatian. I know it can be 
prevented by doing

FLAG long
SFX Y A2 1
SFX Y A2 0 e [^a]

but question is - will hunspell combine (aglutinate) suffixes if suffix 
statements doesn't prevent it explicitly (and allows it implicitly)? 
Wasn't able to solve this by reading the manual from 
http://hunspell.sourceforge.net nor by looking into other .aff files 
from git repository.

Also if .dic file contains only form "PDF" will the word "pdf" in LO be 
underlined as misspelled? Can .dic file contain two words in one line?

Further more - have trouble with understanding the REP section in .aff 
file. Can I do:

REP 1
Plitvička_Jezera Plitvička_jezera

will that automatically correct capital letter in word "Jezera"?

I hope someone will give me a hand here...

Kruno

--
To unsubscribe e-mail to: l10n+unsubscribe@global.libreoffice.org
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/l10n/
All messages sent to this list will be publicly archived and cannot be deleted