IMHO, this is a bug in the VM in that it forces you to decompose
the unicode characters. It is a nice fact to know that the
filesystem stores decomposed UTF-8 file names, but why in the world
did that creep up into the semantic contract of the File
class????? In the unicode standard, the decomposed and composed
characters are intended to carry the same meaning. While there may
be nothing that states that both should work equivalently within a
Unicode application (maybe there is, I am not sure?), it makes life
difficult to have to know which form is required. If Apple chooses
to require the decomposed form in the C APIs, fine, but have the
JAVA VM do the decomposition so the WORA contract of the core JAVA
File class may be maintained. Logically, if the non-UI File class
becomes semantically platform specific, the entire WORA model of
the VM is compromised!!
There's a previous discussion on this topic: http://lists.apple.com/
archives/Java-dev/2004/Nov/msg00472.html
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Java-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/java-dev/email@hidden
