class initialization check overhead

Tom Tromey
Sat Dec 14 12:15:00 GMT 2002

>>>>> "Jeff" == Jeff Sturm <> writes:

>> (BTW does gcj currently know that vtable slots are immutable?  In
>> the past I think I looked at this and decided it didn't know and
>> was therefore missing some potential optimizations.  But I haven't
>> looked in quite some time.)

Jeff> I noticed that too.  I guess the backend considers all memory to
Jeff> be clobbered by a call, even expressions marked with
Jeff> TREE_READONLY or TREE_CONSTANT.  There must be some way to tell
Jeff> the backend that the value at some address will never change.

I know this came up for g++.  There was a discussion about it, but I
didn't follow it to its conclusion.  But there is an answer out there
somewhere.  I'll submit a PR a bit later.


More information about the Java mailing list