[LLVMbugs] [Bug 2300] Wrong indices in GPE with pointer arithmetic and uintptr_t

[bugzilla-daemon at cs.uiuc.edu]

http://llvm.org/bugs/show_bug.cgi?id=2300


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gohman at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #1 from Dan Gohman <gohman at apple.com>  2008-05-08 17:27:59 ---
This looks like the same bug as in PR2295. This testcase also
passes with LLVM trunk, while it fails with LLVM 2.2.

Some background for debugging these kinds of bugs:
The llvm-gcc front-end behaves differently at -O2 than at -O0,
and this is often why compiling with -O0 and optimizing the
result with opt -std-compile-opts produces different results.
To get past this, compile with -O2 (or whatever) and add
-mllvm -disable-llvm-optzns. This will run the front-end as if
it were an -O2 compile, but the LLVM optimizations won't run,
so you can get a bitcode file that you can use with
opt -std-compile-opts with a better chance of reproducing the
problem.

There's also -fdebug-pass-arguments, to print out the
command-line arguments for the passes llvm-gcc runs.


-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.