I checked this patch on my own build on RHEL 6. Regressions were OK.
Details
Details
Diff Detail
Diff Detail
- Repository
- rL LLVM
Event Timeline
| utils/PerfectShuffle/PerfectShuffle.cpp | ||
|---|---|---|
| 425 | This is horrible. We shouldnt leave dead code, and this should really be moved into a DEBUG() statement. However, that is beyond the scope of this particular change, and that isnt your fault. | |
| utils/TableGen/CodeGenDAGPatterns.cpp | ||
| 2011 | Is this used in the rest of the scope? | |
| utils/TableGen/CodeGenDAGPatterns.cpp | ||
|---|---|---|
| 2011 | No, it used only in loop. | |
| utils/TableGen/CodeGenDAGPatterns.cpp | ||
|---|---|---|
| 2011 | (Which reminds that patches uploaded with full context are easier to review) For this particular change, I rather read: unsigned FirstChild = isCommIntrinsic ? 1 : 0; // First operand is intrinsic id.
for (unsigned Child = FirstChild, Last = getNumChildren()-1; Child != Last; ++Child)
if (OnlyOnRHSOfCommutative(getChild(Child))) {
Reason="Immediate value must be on the RHS of commutative operators!"; | |
This is horrible. We shouldnt leave dead code, and this should really be moved into a DEBUG() statement. However, that is beyond the scope of this particular change, and that isnt your fault.