Bugzilla – Bug 249
[X86] Missing cast from ULong -> Double, cast FP -> bool and support for -9223372036854775808
Last modified: 2004-02-22 21:33:59
You need to log in before you can comment on or make changes to this bug.
The X86 backend doesn't have the following two casts implemented: From Type::ULongTy -> floating point From floating point -> bool Also, the asm writer produces invalid output when the constant "1ULL << 63" is emitted. -Chris
This patch adds support for 'cast ulong to double': http://mail.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20040216/012056.html This patch adds support for 'cast double to bool': http://mail.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20040216/012057.html This patch works around the gas bug that prevents -9223372036854775808 from working correctly: http://mail.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20040216/012059.html The SingleSource/Regression/C/casts.c test has been extended to test for these, and an X86 specific testcase test/Regression/CodeGen/X86/2004-02-22-Casts.llx has been added. -Chris