Re: PATCH: Constant-time interface dispatch, type checking, inlining...

[Per Bothner <per at bothner dot com>]

Tom Tromey <tromey@cygnus.com> writes:

> * I notice the JvNew<prim>Array functions no longer have C linkage.
>   Does anybody know why they had C linkage in the first place?
>   (This probably isn't a real problem.)

Memory is getting weak, but it may have had something to do with JNI,
which needs to work in C code.  For normal JNI it shouldn't matter (we
can write the JNI *implementation* in C++, as long it is usable by C code).
However, it might be possible to optimize JNI or use tricks to write
code that is JNI-compatible but optimized for CNI.  (Though in that case
the code would be C++.)

A perhaps better reason:  We do want to have the compiler emit calls to
_Jv_New<prim>Array, thus those should be extern C.  However, I don't
think it matters if JvNew<prim>Array are extern C.

In other words:  It is fine to change them to C++ linkage.  If we find
a reason they should have C linkage (unlikely), we can change it back.
-- 
	--Per Bothner
per@bothner.com   http://www.bothner.com/~per/