Re: Multiple file compilation patch

[Mark Wielaard <mark at klomp dot org>]

Hi,

On Sun, Mar 18, 2001 at 06:00:45PM +0100, Mark Wielaard wrote:
> I ran jacks with it and it gave to following results:
> 
> gkjc:   Total   1630    Passed  1118    Skipped 29      Failed  483
> 
> That doesn't look good (it is even worse then runnen kjc under kaffe).
> But it shows that kjc compiled to a native binary with gcj (on powerpc)
> does work some of the time. I don't know how much of the failures are the
> result of a broken binary and how much are caused by faults in the libgcj
> runtime.
> 
> Compiling another java compiler with gcj and running jacks on the result
> seems a nice way to see how good gcj is at the moment :)

I found a bug in HashMap (libgcj/2357) as a workaround for now I just always
add 1 to the given initialCapacity. And gkjc (KopiSusu kjc compiled with gcj
now gives a lot better result:

gkjc:   Total   1630    Passed  1413    Skipped 7   Failed  210

Still not perfect (we need to lose another 100 failures) but much improved.
kjc now has less failures with gcj then under kaffe :) This is on powerpc.

Cheers,

Mark
-- 
Stuff to read:
    <http://www.toad.com/gnu/whatswrong.html>
  What's Wrong with Copy Protection, by John Gilmore