This is the mail archive of the java@gcc.gnu.org mailing list for the Java project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: sqrt interpreter Q

From: shudo at computer dot org
To: java at gcc dot gnu dot org
Date: Fri, 08 Feb 2002 18:55:11 +0900 (JST)
Subject: Re: sqrt interpreter Q
References: <3C62C85C.2665595A@pop.agri.ch><87adukr3hx.fsf@creche.redhat.com>

> Andreas> [titanium:~] andreast% gij my_sqrt
> Andreas> sqrt_int: -2147483648
> Andreas> sqrt_double: NaN
> Andreas> sqrt_long: -9223372034707292160

> I haven't looked into it more deeply than this.  It does suggest that
> the f2* and d2* instructions in the interpreter might need to be
> modified to handle the special cases.

I second you. Java spec needs `(int) NaN' translated to `0'.
The gij does not treat the special case, I guess.

The f2i, f2l, d2i and d2l instructions have to handle NaN specially.
Unfortunately, this means those instructions involve a conditional
branch.

  Kazuyuki Shudo	shudo@computer.org	http://www.shudo.net/

References:
- sqrt interpreter Q
  - From: Andreas Tobler
- Re: sqrt interpreter Q
  - From: Tom Tromey

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]