Playing with latest [lib]gcj

Sun Jun 4 07:58:00 GMT 2000

Alexandre Petit-Bianco wrote:
>> `Jeff Sturm writes:
>> > I think this is the old array initializer problem (didn't you fix
> > that once?).
>> I don't remember. I'll have to look at the source of the test
> case. Or, could you file a PR with the involved source code?

I was thinking of PR/214. We fixed array initializers then for int, and
discussed alignment for double arrays. I didn't find a suitable
regression test, so I've just entered PR/252.
> > Replacing the double array declaration with element-at-a-time
> > initializers seems to work, as does compiling from bytecode.
>> Good to hear we can have it to run.

It runs about as expected too... a little slower than the C code due to
array bounds checking and class initialization checks.
--
Jeff Sturm
jeff.sturm@appnet.com