llvm-svn: 3255

* Boldify stuff that changes in the help output.

llvm-svn: 3254

llvm-svn: 3253

llvm-svn: 3251

llvm-svn: 3249

in choosing how to create int-set instructions.

llvm-svn: 3248

llvm-svn: 3247

after array offset correctly.  The type was not being updated for
array offsets!

llvm-svn: 3246

when folding or not folding GEPs.

llvm-svn: 3245

Fix 2 bugs in FoldGetElemChain so index vector is not modified
when no GEPs are folded in, and so a hasLeadingZero is computed
only for the last folded GEP, not the one after that if any.

llvm-svn: 3244

llvm-svn: 3243

llvm-svn: 3240

llvm-svn: 3239

llvm-svn: 3238

llvm-svn: 3237

llvm-svn: 3236

llvm-svn: 3234

Don't decompose if there are 2 indices with 0 as first index.
Compute Changed flag correctly in runOnBasicBlock().

llvm-svn: 3233

llvm-svn: 3232

llvm-svn: 3231

uses the modifier "%ld".  Now lli passes off "%lld" to the underlying runtime
library in this case.

llvm-svn: 3230

llvm-svn: 3227

llvm-svn: 3226

Implement feature: Cast's can now be converted to bitwise AND expressions.

llvm-svn: 3225

llvm-svn: 3224

llvm-svn: 3223

   - Reenable gep (gep x) -> x
   - Make instcombine do dead instruction elimination where it's really
     easy.  Now visitors don't have to ensure they aren't not processing
     dead instructions.

llvm-svn: 3222

If instcombine is making dead instructions, ALL that should be used is -die

llvm-svn: 3221

llvm-svn: 3220

llvm-svn: 3219

llvm-svn: 3218

llvm-svn: 3216

llvm-svn: 3215

defintiions dominate their uses

llvm-svn: 3214

llvm-svn: 3213

llvm-svn: 3212

Dominance calculation goes to VMCore library to be used by Verifier.

llvm-svn: 3210

llvm-svn: 3209

llvm-svn: 3208

llvm-svn: 3207