Consider the following reduced testcase:

test.aff:

SFX E Y 1SFX E a u aSFX H Y 1SFX H 0 u m

test.dic

2Lima/ELiM/H

The first stemLima/E in the "test.dic" file is expected to produce the"Limu" form. But in reality the second stemLiM/H appears to somehow cancel it:

$ cat test.affSFX E Y 1SFX E a u aSFX H Y 1SFX H 0 u m$ cat test.dic2Lima/ELiM/H$ hunspell -d testHunspell 1.7.2Lima*Limu# Limu 0LiM*Lim& Lim 1 0: LiMLiMu& LiMu 1 0: LiM

Removing the second stem makes this problem disappear:

$ cat test.affSFX E Y 1SFX E a u aSFX H Y 1SFX H 0 u m$ cat test.dic1Lima/E$ hunspell -d testHunspell 1.7.2Limu+ Lima

Is this how Hunspell is supposed to work? Looks like a camel-case word cancels spellchecking of another capitalized word.