fix a bug in constraint enforcement; improve phi handling and phi0 construction (still not finished)

[r26335]

benewalloc: cleanups; dump intermediate results; rework copy tracking logic; started working on SSA construction

[r26334]

- make phi nodes use backend_info instead of phi_handler hashmaps

[r26332]

[r26331]

[r26330]

[r26329]

- rename bespilloptions.* to bespill.* and bespill.* to bespillutil.*

[r26327]

[r26326]

[r26325]

[r26321]

  (We have a separate get_input, and get_output callback for now).
  This should make the code faster for now and is a first step towards
  changing the interface to query register constraints on the mode_T node
  itself instead of the Proj nodes.
- Handle middleend node constraints and stuff in benode.c instead of in each
  backend
- Remove irn_class_branch we already had is_cfop in the middleend
- Fix a bunch of bugs/problems in the process

[r26320]

[r26318]

[r26317]

[r26315]

[r26314]

[r26313]

[r26311]

[r26310]

cleanup backend: make pre_spill_prepare_constraint independent of chordal allocator structures (and rework it a bit in the process)

[r26304]

[r26303]

[r26302]

[r26301]

[r26299]

[r26298]

[r26297]

[r26280]

[r26260]

[r26241]

[r26239]

- SymConst's are now ALWAYS placed in the start block

[r26236]

[r26224]

[r26223]

[r26222]

[r26221]

rework python generators: we can make use of the classes/inheritance language features to nicely define our nodes

[r26220]

[r26190]

Bad nodes just like on Block nodes.
This simplifies get_irn_irg() and implements the idea the the irg can
be retrieved from "the block" of a node.

[r26188]

[r26181]

[r26180]

[r26179]