When LC_CTYPE it set to "C", neither textarea or textfield take ISO-Latin-1 supplmentary character input from keyboard, while it is set to a western european locale such as "fr", "de", they do.

The attacheed test program creates one textfield and one textarea object. When inputting <Compose>, <a> and <e> succeesively either of the objects, it simply input 'a' and 'e'. As is done in "fr" or "de", this operation is supposed to yield the single string \u00e6 in Unicode encoding, which looks alike combination character of <a> and <e>.

Looking at the sequence of keycodes in the case above thorough KeyDown method, they were none, 96, 101. while the correct result should have been none, 96 and 230, in Solaris platform.