Re: Bug in GHC 5.04.2 for Mac OS X

2003-01-31 Thread Wolfgang Thaller
GPR13 is indeed considered nonvolatile, so it looks like the JVM is correct and the function made by createAdjustor is wrong. Thanks, fixed. (both in the HEAD and in the stable branch) I'm beginning to think that the PowerPC has way too many registers (and Intel still has too few - the good

Re: Bug in GHC 5.04.2 for Mac OS X

2003-01-30 Thread Ashley Yakeley
At 2003-01-30 20:27, I wrote: GPR13 is indeed considered nonvolatile, so it looks like the JVM is correct and the function made by createAdjustor is wrong. See enclosed example. -- Ashley Yakeley, Seattle WA GPR13Bug.tar.gz Description: Binary data