Assume an 8-char word checked by Hunspell.
The first and last characters of this word are vowels with diacritics which the user has not entered. Therefore, the expected first correct suggestion would be the word with its two missing diacritics. ATTENTION: When one diacritic is missing, hunspell works fine. Problems of this nature appear when TWO chars in the same word are misspelled.
Here is how Hunspell works:
It first finds and suggests words that use the word entered as is (i.e. without the missing diacritics) in various transformations (compound words based on the word checked, inflected forms produced by adding extra characters at the beginning or the end of the word being checked, additions of characters AND replacements of more than one characters, etc). Only when these suggestions are finished does Hunspell suggest the simple obvious one, i.e., the word with its missing diacritics. Sometimes the useless words found are so many, that the correct word does not even appear at all. This is a problem.
I tried giving the priority to the settings of diacritics via REP statements right at the beginning of the REP list. Nothing happened.
Is there any way to instruct Hunspell to first search for correct words by replacing NOT JUST ONE, BUT TWO CHARACTERS in the misspelled word and then proceed to its usual processing? This would solve this problem and others that I have seen reported at vaious instances over time.
Thanks a lot for your attention.
Two distant mutations are not detected whatsoever. Better is to add the multiple REP.