Native compilation article

[Tom Tromey]

>>>>> "Martyn" == Martyn Honeyford <MARTYNH@uk.ibm.com> writes:

[ gcj -vs- case-insensitive file systems ]

Martyn> Is there any work underway to sort this out (or is it a
Martyn> nightmare to fix?)

I don't think anybody is working on it.  I wouldn't call it a
nightmare to fix.  It probably isn't trivial though.

Martyn> It would be great to be able to use gcj under cygwin.

It will happen eventually.  It all depends on interest from hackers.
I don't use Cygwin, so unless Red Hat gets a contract for it, it is
unlikely I'll be doing the work...

Martyn> As an aside - I hope no-one out there thinks I have anything
Martyn> against GCJ.  On the contrary, I think it is great and I hope
Martyn> the article will encourage people to check it out for
Martyn> themselves - the article *does* encourage people to ignore the
Martyn> limited benchmarks and try it for themselves on their own
Martyn> applications.

That's not the message I got from the article.  I thought the message
was pretty clearly "native compilers aren't worth it".

That said, I don't mind comparisons between gcj and other platforms.
In fact, I think they are important.  Of course, I'd prefer
comparisons against the fastest release off CVS that actually works
:-), but I realize that's probably unrealistic.

Tom