The problem with new ICS LatinIME I noticed is that it does not recognize
unicode characters from languages like Czech or Croatian properly.
Gingerbread LatinIME worked fine.
In examples I'm refering to LatinIME versions in which I included Croatian
dictionary
First, examples of expected behaviour:
1. ICS English - Start typing: "hom"
IME offers:
him, home, homogeneous, hometown, homemade, homosexual, homeland, homicide...
2. Gingerbread Croatian - Start typing: "šuš"
IME offers:
šuška, šušku, šuški, šušte, šuškom...
Here is what happens in CM9 ICS:
1. Croatian - One starts to type: "šuš"
IME offers:
usput, sluša, šuma, slušanja, šumi, slušaš, usprokos, uskrsnuće, slušao,
usluge...
IME should offer (from dictionary with frequency):
...
<w word="šuška" f="48"></w>
<w word="šuškanje" f="44"></w>
<w word="šuškalo" f="44"></w>
<w word="šuša" f="37"></w>
<w word="šuškanja" f="29"></w>
<w word="šuše" f="28"></w>
<w word="šuštanja" f="26"></w>
...
2. Czech
One starts to type: "čer"
IME offers: čele, včetně, ČSSR, českou, dcera, čemu, česky, erik, českým...
IME should offer (from dictionary with frequency):
...
<w word="červ" f="255"></w>
<w word="červen" f="254"></w>
<w word="čele" f="170"></w>
<w word="černá" f="138"></w>
<w word="červenec" f="127"></w>
<w word="červená" f="103"></w>
<w word="června" f="96"></w>
...
I believe that the problem begins with native/src/char_utils.cpp but changes
are needed elsewhere too. It definitely misses the following:
{ 0x0106, 0x0107 }, // LATIN CAPITAL LETTER C WITH ACUTE
{ 0x010C, 0x010D }, // LATIN CAPITAL LETTER C WITH CARON
{ 0x017D, 0x017E }, // LATIN CAPITAL LETTER Z WITH CARON
{ 0x0160, 0x0161 }, // LATIN CAPITAL LETTER S WITH CARON
Possibly other chars for other languages...
Also noticed that character definitions are missing in basechars.h list. But
after I added both of these and recompiled LatinIME, it still didn't help.
It's obvious that prediction engine doesn't see undefined characters as valid
ones. Why - that's something I'm trying to figure out and would appreciate a
hint or two where to look for it. It seems to me that the problem could be in
/LatinIME/native/src but I don't see it. :(
I'm lost... :(
--
You received this message because you are subscribed to the Google
Groups "Android Developers" group.
To post to this group, send email to android-developers@googlegroups.com
To unsubscribe from this group, send email to
android-developers+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-developers?hl=en
