mfabian@ari:/usr/share/myspell $ rpm -qf hu_HU.aff hunspell-hu-1.6.1-4.fc18.noarch mfabian@ari:/usr/share/myspell $ The file command already tells use that there is something wrong with the encoding: mfabian@ari:/usr/share/myspell $ file hu_HU.aff hu_HU.aff: Non-ISO extended-ASCII text, with very long lines The file claims to be UTF-8 encoded: mfabian@ari:/usr/share/myspell $ grep ^SET hu_HU.aff SET UTF-8 There are indeed some lines which are UTF-8 encoded: mfabian@ari:/usr/share/myspell $ grep ^TRY hu_HU.aff TRY íóútaeslzánorhgkiédmyőpvöbucfjüűxwq-.à mfabian@ari:/usr/share/myspell $ grep ^KEY hu_HU.aff KEY öüó|qwertzuiopőú|asdfghjkléáű|íyxcvbnm But there are other lines in ISO-8859-2 encoding: mfabian@ari:/usr/share/myspell $ grep ^NAME hu_HU.aff NAME Magyar Ispell helyes�r�si sz�t�r mfabian@ari:/usr/share/myspell $ grep ^NAME hu_HU.aff | iconv -f iso-8859-2 -t utf-8 NAME Magyar Ispell helyesírási szótár mfabian@ari:/usr/share/myspell $ grep 2002-2010 hu_HU.aff # 2002-2010 (c) L�szl� N�meth <nemeth> and Ferenc God� mfabian@ari:/usr/share/myspell $ grep 2002-2010 hu_HU.aff | iconv -f iso-8859-2 -t utf-8 # 2002-2010 (c) László Németh <nemeth> and Ferenc Godó There are many lines at the top of the file which are neither in UTF-8 nor in ISO-8859-2 encoding: mfabian@ari:/usr/share/myspell $ head -n 2 hu_HU.aff AF 1262 AF V˯j�Ln�����TtYc��l # 1 mfabian@ari:/usr/share/myspell $ head -n 2 hu_HU.aff | iconv -f iso-8859-2 -t utf-8 AF 1262 AF VËŻj×LnÓéčłÄäTtYc¸źl # 1 mfabian@ari:/usr/share/myspell $ head -n 2 hu_HU.aff | iconv -f iso-8859-1 -t utf-8 AF 1262 AF V˯j×LnÓéè³ÄäTtYc¸¼l # 1 mfabian@ari:/usr/share/myspell $
This is *not* coming from our .spec file, the problem exists like that in the original upstream tarball hu_HU-1.6.1.tar.gz already.
I think this is the result of running "makealias" over the originals to make a more compact format where the stuff after "AF" is basically raw binary
I'm not touching this with a barge pole, it's apparently deliberate according to nemeth