Re: Problems with float Math.min(float,float) etc
> "Eric" == Eric Blake <[EMAIL PROTECTED]> writes: Eric> * java/lang/Double.java (equals, compare): Fix 0.0 vs. -0.0 math. I don't think anything was wrong with the old logic in equals(). Tom ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
Re: build question
Brian, Thanks for your reply. I check out the CVS version, however it does not have a *configure* file in it, even though the INSTALL file does mention to run it first. I use the command cvs -d:pserver:[EMAIL PROTECTED]:/cvsroot/classpath co classpath to fetch the cvs copy. So, what should I do to proceed? Thanks, Larry Brian Jones writes: > > <[EMAIL PROTECTED]> writes: > > > Hello, All, > > > > I am tring to build classpath-0.02 on Linux. However, it seems > > the Japhar package is required by autoconf configure. I checked > > the install file, but Japhar is not listed as required. > > So I want to ask is Japhar package still necessary to install classpath, > > as I want to use it only with intel orp and sablevm? If not, what > > is the right command line for configure? If so, what is the role > > of Japhar in the building process? > > At this point, you would be better served to build classpath from > CVS. I'm expecting that we'll release a new version soon but I'm > waiting on something from the FSF. The current version does not > depend on Japhar or any other VM to build from CVS. > > Brian > -- > Brian Jones <[EMAIL PROTECTED]> > ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
Re: Problems with float Math.min(float,float) etc
Tom Tromey wrote: > > > "Eric" == Eric Blake <[EMAIL PROTECTED]> writes: > > Eric> * java/lang/Double.java (equals, compare): Fix 0.0 vs. -0.0 math. > > I don't think anything was wrong with the old logic in equals(). > Yes, you are right that there were no logic errors in the old version of equals(). But the version I committed avoids 2 calls to a native method, and accesses the field "value" directly rather than calling "doubleValue()". -- This signature intentionally left boring. Eric Blake [EMAIL PROTECTED] BYU student, free software programmer ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
Re: Problems with float Math.min(float,float) etc
> "Eric" == Eric Blake <[EMAIL PROTECTED]> writes: Eric> Yes, you are right that there were no logic errors in the old Eric> version of equals(). But the version I committed avoids 2 calls Eric> to a native method, and accesses the field "value" directly Eric> rather than calling "doubleValue()". It seems to me that a JIT could easily inline doubleToLongBits. Probably they don't though. gcj will (not yet, but it is a simple addition). I think I'll keep the original code in libgcj for now. Tom ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
Re: State of AWT?
> "John" == jewel <[EMAIL PROTECTED]> writes: John> I got stuck with the code that handled peers. I can't remember John> specifics now, but basically it covered the order in which John> addNotify and similar methods are called. There were issues with John> inheritance. Ok, from this and other comments, I'm not going to worry too much about possibly introducing bugs. John> I think what is needed is quick review of how the peer system is John> meant to work, and from there things should flow easily. This isn't exactly what you're talking about, but part of libgcj plan was to simplify the peers by removing redundant methods. We were also planning to do event masking on the peer side (according to the code in libgcj, but it also seems like a potentially good idea). For the time being I'm going to ignore this difference; I've added a couple stub methods where I needed them. I'll be checking in the first merge patch soon. It doesn't cover everything, but it does do a lot. It merges various classes, and adds a lot of the 1.2 stuff that was in libgcj but not Classpath (this code is by no means complete, but it's a start). My general plan is to merge what we've got, and then start trying the existing Classpath peers with libgcj. Tom ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
Re: build question
<[EMAIL PROTECTED]> writes: > Brian, > > Thanks for your reply. I check out the CVS version, however it > does not have a *configure* file in it, even though the INSTALL > file does mention to run it first. > I use the command > cvs -d:pserver:[EMAIL PROTECTED]:/cvsroot/classpath co classpath > to fetch the cvs copy. > > So, what should I do to proceed? >From CVS, the directions in the HACKING file have more information. You should be able to use jikes or some versions of gcj to build Classpath. Brian -- Brian Jones <[EMAIL PROTECTED]> ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
202_jess bug?
I'm sorry that this does not pertain to Classpath per se, but it seems to be the kind of thing you guys would have thought about at some point... What is Class.getConstructor supposed to do when the matching constructor is not public? When I run the SPECjvm98 202_jess benchmark, with my slightly hacked Classpath library, it dies with the message "*** FATAL ***: Missing intrinsic function class". When I looked into it, the benchmark calls Class.getConstructor, which is supposed to return only a public constructor. The class it is trying to probe has only one constructor, which javap prints like this: spec.benchmarks._202_jess.jess._return(); To me, the lack of access specifiers here means that this constructor is not public. Thus, the correct behaviour for getConstructor is (I think) to throw a NoSuchMethodException, which it does. The benchmark clearly does not expect an exception to be thrown, and produces the error message. So what going on here? Should I just be happily returning the constructor even if it's not public? If so, what is the difference between getConstructor and getDeclaredConstructor? Thanks for your help. -- Patrick Doyle [EMAIL PROTECTED] ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath