From: Godmar Back (gback@cs.utah.edu)
Date: Tue Dec 08 1998 - 01:01:32 EST
Also, I noticed you tried to trade space for time by defining
arrays of masks; but would this really work? If it does work,
then maybe that's just because the jit-compiler is not able to
efficiently compile (1 << x) statements. Otherwise, I'd expect
loading the result from a table in memory to be slower.
- Godmar
>
> >
> > Yeah - sleep is busted can you fix?
> >
>
> Done.
>
> You also broke BitSet by replacing Edouard's implementation with
> a homebrewn one.
>
> This seems like a step back.
> Edouard's implementation corresponded to Sun's serial form (does yours?)
> and it was documented. Moreover, it was complete according to 1.2.
>
> I think it's bad that so little of Kaffe's sources is documented; if
> people add comments, I think they should be kept.
>
> - Godmar
>
> BitSetTest ... error compiling
> BitSetTest.java:11: function length() not found in class java.util.BitSet
> System.out.println(name + " = " + b + "\tsize " + b.size() + "\tlength " + b.length());
> ^
> BitSetTest.java:91: function andNot(java.util.BitSet) not found in class java.util.BitSet
> d.andNot(b);
> ^
> 2 errors
>
This archive was generated by hypermail 2b29 : Sat Sep 23 2000 - 19:57:07 EDT