No analysis for this case yet.

[r26711]

[r26709]

[r26708]

[r26691]

[r26690]

  (with different arguments and register constraints for the same node now)
- Lots of cleanups/changes in the arm backend. We can represent all
  "shifter operands" now (but the code selection isn't optimal yet)
- More fixes all over the place - arm backend handles 164.gzip now

[r26673]

Restricted nodes are now correctly inserted before not restricted Nodes in reverse perfect elimination order

[r26671]

[r26670]

[r26663]

[r26662]

[r26661]

Bitset contain's information if node has register constraints or not

[r26660]

[r26659]

[r26658]

[r26657]

[r26656]

[r26652]

  disabled some broken transformers. (Queens, Quicksort and several other stuff from the
  testsuite work, but still lots of stuff broken)

[r26649]

[r26648]

[r26647]

[r26646]

[r26645]

[r26643]

[r26642]

[r26641]

[r26640]

[r26639]

[r26638]

[r26637]

[r26636]

[r26635]

[r26634]

[r26633]

[r26632]

[r26631]

[r26630]

[r26629]

[r26628]

[r26627]

[r26626]