Fixed a non-critical bug in the register allocator

*** empty log message ***

  Slight cleanup on the Alpha.
  Added a bunch of instructions to the x86 instruction set.
  The module ra-rewrite-with-renaming has been improved.
  These should have no effect on SML/NJ.
  CVS tag=leunga-20000515-alpha-x86-ra

merging changes from devel branch; see HISTORY

fixed important anchor-related bug in ml-build

merged my path anchor changes into main trunk
more config/srcarchiveurl! YOU NEED NEW BINFILES more config/srcarchiveurl!
Read HISTORY and src/system/README!

CM.sources implemented; CM.makedepend dropped

    Various bug fixes and new features for C--, Moby and MLRISC optimizations.
    See smlnj/HISTORY for details.  CVS tag: leunga-20000510-moby-c--ssa

  Fermin has found a few assembly problems with constant expressions
  generated in LabelExp.  Mostly, the problems involve extra parentheses,
  which choke on dumb assemblers.  This is his fix.

Version 110.28

merged changes from devel branch (added noweb tool)

fixes for bugs 1498 and 1510

merged pending changes into main trunk; see HISTORY; new binfiles!!

fixed install.sh (bug 1563)

added .fun as a file name suffix known to CM

1. Alpha

   PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
   in C--.

2. RA

   Added a flag "ra-dump-size" to print out the size of the flowgraph
   and the interference graph.

update ml-yacc documentation and examples for 110.*

   This update synchronizes my repository with Yale's.  Most of these
changes are related to C--, Moby, and my optimizations.  It should have
little impact on SML/NJ.

   CVS tag leunga-20000420-ssa-c---stuff

changed install.sh script to use/handle shorter archive names

Version 110.27

   Assembly fix for x86 and other miscellaneous changes to MLRISC

Oops, forgot about this one.
What kind of file format is that ?

    Forgot these files in the last update

    New peephole code, improvements to x86 instruction selection, and
    fixes for SSA and the machine description tool.

    CVS tag=leunga-20000406-peephole-x86-SSA

merging last bits before 110.27 code freeze

      Changes to assembly output and fixes a bug in the x86 assembler
      CVS tag=leunga-20000404-x86-asm

    Fixes related to C--, Moby and my optimizations.
    CVS tag=leunga-20000404-C--Moby

merging branch blume_devel_v110p26p2_1 (elimination of corenv)

   1. Renamed the constructor CALL in MLTREE by popular demand.
   2. Added a bunch of files from my repository.  These are currently
      used by other non-SMLNJ backends.

   This update contains a new module for extracting aliasing information.
   CVS tag: leunga-20000331-aliasing

fixes for c-libraries.c and bug 1556

changes to modmap generator and CM tool parameter mechanism (see HISTORY)

merging back development branch blume_devel_v110p26p1_3...
This involves changes to CM and the removal of CMStaticEnv from
the compiler.  See the HISTORY file for more information.

   Fixed a slight problem with gc types annotations.

   To turn on gc types, do this:

       CM.autoload "sml/compiler.cm";
       Compiler.Control.MLRISC.getFlag "mlrisc-gc-types" := true;

   This update contains major changes to the code generator and various
back ends.  Please see the entry leunga-20000327-mlriscGen_hppa_alpha_x86
in the file sml/HISTORY for details.

1. X86 fixes/changes

   a.  The old code generated for SETcc was completely wrong.
       The Intel optimization guide is VERY misleading.

2. ALPHA fixes/changes

   a.  Added the instructions LDBU, LDWU, STB, STW as per Fermin's suggestion.
   b.  Added a new mode byteWordLoadStores to the functor parameter to Alpha()
   c.  Added reassociation code for address computation.

*** empty log message ***

merging back blume_devel_v110p26p1_1

improved version of Dave's fixpt script added

More assembly output problems involving the indexed addressing mode
on the x86 have been found and corrected. Thanks to Fermin Reig for the
fix.

The interface and implementation of the register allocator have been changed
slightly to accommodate the possibility to skip the register allocation
phases completely and go directly to memory allocation.  This is needed
for C-- use.

This fix only affects the x86 assembly output.

BOOT and HISTORY file for blume_main_v110p26p1_0

BOOT and HISTORY updated

HISTORY file creation