compiler.cfg.linear-scan.*: dont use a hashtable here, it fixes register

allocation so it's always in the same order as machine-registers

basis/compiler/cfg/linear-scan/allocation/allocation-tests.factor

@@ -4,7 +4,7 @@ compiler.cfg.linear-scan.live-intervals cpu.architecture
 cpu.x86.assembler.operands heaps kernel namespaces system tools.test ;
 IN: compiler.cfg.linear-scan.allocation.tests
 
-: unassigned-interval ( -- live-interval )
+: interval-[30,46] ( -- live-interval )
     T{ live-interval-state
        { vreg 49 }
        { start 30 } { end 46 }
@@ -18,17 +18,31 @@ IN: compiler.cfg.linear-scan.allocation.tests
        { reg-class int-regs }
     } clone ;
 
+: interval-[30,60] ( -- live-interval )
+    T{ live-interval-state
+       { vreg 25 }
+       { start 30 } { end 60 }
+       { reg-class int-regs }
+       { reg RAX }
+    } ;
+
 cpu x86.64? [
     ! assign-registers
-    { R8 } [
-        { { int-regs V{ } } { float-regs V{ } } } active-intervals set
-        unassigned-interval dup machine-registers assign-register reg>>
+    { RAX } [
+        f machine-registers init-allocator
+        interval-[30,46] dup machine-registers assign-register reg>>
     ] unit-test
 
     ! register-status
-    { { R8 1/0. } } [
-        { { int-regs V{ } } { float-regs V{ } } } active-intervals set
-        unassigned-interval machine-registers register-status
+    { { RAX 1/0. } } [
+        f machine-registers init-allocator
+        interval-[30,46] machine-registers register-status
+    ] unit-test
+
+    { { RCX 1/0. } } [
+        f machine-registers init-allocator
+        interval-[30,60] add-active
+        interval-[30,46] machine-registers register-status
     ] unit-test
 ] when
 
@@ -83,7 +97,7 @@ cpu x86.64? [
     <min-heap> unhandled-min-heap set
     f f <basic-block> <cfg> cfg set
     40 progress set
-    T{ sync-point { n 40 } } unassigned-interval spill-at-sync-point
+    T{ sync-point { n 40 } } interval-[30,46] spill-at-sync-point
 ] unit-test
 
 ! spill-at-sync-point?

basis/compiler/cfg/linear-scan/allocation/allocation.factor

@@ -1,6 +1,6 @@
 ! Copyright (C) 2008, 2010 Slava Pestov.
 ! See http://factorcode.org/license.txt for BSD license.
-USING: accessors assocs combinators combinators.short-circuit
+USING: accessors arrays assocs combinators combinators.short-circuit
 compiler.cfg.linear-scan.allocation.spilling
 compiler.cfg.linear-scan.allocation.state
 compiler.cfg.linear-scan.live-intervals compiler.utilities fry
@@ -18,10 +18,13 @@ IN: compiler.cfg.linear-scan.allocation
         _ add-use-position
     ] each ;
 
+: free-positions ( registers reg-class -- avail-registers )
+    of [ 1/0. 2array ] map ;
+
 : register-status ( new registers -- free-pos )
-    over reg-class>> free-positions
-    [ inactive-positions ] [ active-positions ] [ nip ] 2tri
-    >alist alist-max ;
+    over reg-class>> free-positions [
+        [ inactive-positions ] [ active-positions ] 2bi
+    ] keep alist-max ;
 
 : no-free-registers? ( result -- ? )
     second 0 = ; inline

basis/compiler/cfg/linear-scan/allocation/state/state.factor

@@ -3,8 +3,7 @@
 USING: accessors arrays assocs combinators compiler.cfg
 compiler.cfg.instructions
 compiler.cfg.linear-scan.live-intervals compiler.cfg.registers
-cpu.architecture fry heaps kernel linked-assocs math
-math.order namespaces sequences ;
+cpu.architecture fry heaps kernel math math.order namespaces sequences ;
 FROM: assocs => change-at ;
 IN: compiler.cfg.linear-scan.allocation.state
 
@@ -140,9 +139,6 @@ SYMBOL: spill-slots
     H{ } clone spill-slots set
     -1 progress set ;
 
-: free-positions ( registers reg-class -- assoc )
-    of [ 1/0. ] H{ } <linked-assoc> map>assoc ;
-
 : add-use-position ( n reg assoc -- )
     [ [ min ] when* ] change-at ;