From 7cc1505b153bf0892a231685a1ebbe824d8e58e3 Mon Sep 17 00:00:00 2001
From: 4am <a2_4am@sent.com>
Date: Fri, 17 Dec 2021 18:29:35 -0500
Subject: [PATCH] shave some cycles

---
 src/fx/fx.dgr.fizzle.a              | 10 +++---
 src/fx/fx.dhgr.2bit.fizzle.a        |  8 ++---
 src/fx/fx.dhgr.2pass.lr.a           |  6 +---
 src/fx/fx.dhgr.bar.dissolve.a       |  6 +---
 src/fx/fx.dhgr.diagonal.a           |  6 +---
 src/fx/fx.dhgr.fizzle.a             |  9 +++--
 src/fx/fx.dhgr.precomputed.1bit.a   |  8 ++---
 src/fx/fx.dhgr.precomputed.2bit.a   |  8 ++---
 src/fx/fx.dhgr.r.by.pixel.a         |  8 ++---
 src/fx/fx.dhgr.redlines.a           |  7 ++--
 src/fx/fx.gr.fizzle.a               | 10 +++---
 src/fx/fx.hgr.1bit.fizzle.a         |  8 ++---
 src/fx/fx.hgr.2bit.fizzle.a         |  9 +++--
 src/fx/fx.hgr.checkerboard.fizzle.a |  9 +++--
 src/fx/fx.hgr.fizzle.a              | 10 +++---
 src/fx/fx.hgr.palette.fizzle.a      |  9 +++--
 src/fx/fx.hgr.precomputed.1bit.a    |  8 ++---
 src/fx/fx.hgr.precomputed.2bit.a    |  8 ++---
 src/fx/fx.hgr.precomputed.3bit.a    |  8 ++---
 src/fx/fx.shr.2pass.lr.a            |  8 ++---
 src/fx/fx.shr.common.a              | 52 +++++++++++++++++++++++++++++
 src/fx/fx.shr.diagonal.a            | 10 ++----
 src/fx/fx.shr.fizzle.a              |  7 ++--
 src/fx/fx.shr.lr.a                  |  9 ++---
 src/fx/fx.shr.lr2.a                 | 10 ++----
 src/fx/fx.shr.precomputed.a         | 49 +++++----------------------
 src/fx/fx.shr.radial.a              | 48 +++++---------------------
 src/fx/fx.shr.ud.a                  | 10 ++----
 src/fx/macros.a                     | 22 ++++++++++++
 29 files changed, 161 insertions(+), 219 deletions(-)

diff --git a/src/fx/fx.dgr.fizzle.a b/src/fx/fx.dgr.fizzle.a
index 4be43781e..03a774b4b 100644
--- a/src/fx/fx.dgr.fizzle.a
+++ b/src/fx/fx.dgr.fizzle.a
@@ -5,11 +5,11 @@
 !to "build/DGR.FIZZLE",plain
 *=$6000
 
-         ldx   #(end-start+1)        ; copy LFSR code to zero page
--        ldy   start-2, x
-         sty   $FE, x
-         dex
-         bne   -
+         !source "src/fx/macros.a"
+
+         +OVERCOPY_TO_0 start, end
+         ;X=0
+         ;Y=0
          jmp   loop
 
 start
diff --git a/src/fx/fx.dhgr.2bit.fizzle.a b/src/fx/fx.dhgr.2bit.fizzle.a
index aec88a694..0f653a439 100644
--- a/src/fx/fx.dhgr.2bit.fizzle.a
+++ b/src/fx/fx.dhgr.2bit.fizzle.a
@@ -8,11 +8,9 @@
 addrs     = $6100                    ; [256 bytes, page-aligned, duplicated in auxmem]
 copymasks = $6200                    ; [256 bytes, page-aligned, duplicated in auxmem]
 
-         ldx   #(end-start)          ; copy LFSR code to zero page
--        lda   start-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         !source "src/fx/macros.a"
+
+         +COPY_TO_0 start, end
          ;X=0
          ;A=3
 --       ldy   #$40                  ; create identical copymask tables in main and aux memory
diff --git a/src/fx/fx.dhgr.2pass.lr.a b/src/fx/fx.dhgr.2pass.lr.a
index 9a32d7f3e..fe75e766d 100644
--- a/src/fx/fx.dhgr.2pass.lr.a
+++ b/src/fx/fx.dhgr.2pass.lr.a
@@ -10,11 +10,7 @@ hgr1hi   =     $0300                 ; [$C0 bytes, main memory only]
 
          !source "src/fx/macros.a"
 
-         ldx   #(end-start)          ; copy code to zero page
--        lda   start-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         +COPY_TO_0 start, end
          ;X=0
          +BUILD_HGR_LOOKUP_TABLES_X_IS_ALREADY_0 hgrlo, hgr1hi
          jmp   loop
diff --git a/src/fx/fx.dhgr.bar.dissolve.a b/src/fx/fx.dhgr.bar.dissolve.a
index e1d43d503..93a0763db 100644
--- a/src/fx/fx.dhgr.bar.dissolve.a
+++ b/src/fx/fx.dhgr.bar.dissolve.a
@@ -10,11 +10,7 @@ hgr1hi   =     $0300                 ; [$C0 bytes, main memory only]
 
          !source "src/fx/macros.a"
 
-         ldx   #(end-start)          ; copy code to zero page
--        lda   start-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         +COPY_TO_0 start, end
          ;X=0
          +BUILD_HGR_LOOKUP_TABLES_X_IS_ALREADY_0 hgrlo, hgr1hi
          jmp   loop
diff --git a/src/fx/fx.dhgr.diagonal.a b/src/fx/fx.dhgr.diagonal.a
index 1a763b05b..c5805c135 100644
--- a/src/fx/fx.dhgr.diagonal.a
+++ b/src/fx/fx.dhgr.diagonal.a
@@ -11,11 +11,7 @@ copymasks=     $02C0                 ; [$08 bytes, different values in main and
 
          !source "src/fx/macros.a"
 
-         ldx   #(end-start)          ; copy code to zero page
--        lda   start-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         +COPY_TO_0 start, end
          ;X=0
          +BUILD_HGR_LOOKUP_TABLES_X_IS_ALREADY_0 hgrlo, hgr1hi
 
diff --git a/src/fx/fx.dhgr.fizzle.a b/src/fx/fx.dhgr.fizzle.a
index f4f0887b0..14fe3348a 100644
--- a/src/fx/fx.dhgr.fizzle.a
+++ b/src/fx/fx.dhgr.fizzle.a
@@ -7,6 +7,8 @@
 
 addrs=$BF                            ; [$40 bytes]
 
+         !source "src/fx/macros.a"
+
          ldx   #$1F                  ; build address lookup table
 -        txa
          eor   #$20
@@ -15,11 +17,8 @@ addrs=$BF                            ; [$40 bytes]
          sta   addrs+$20, x
          dex
          bpl   -
-         ldx   #(end-start+1)        ; copy LFSR code to zero page
--        ldy   start-2, x
-         sty   $FE, x
-         dex
-         bne   -
+
+         +OVERCOPY_TO_0 start, end
          ;X=0
          ;Y=0
          jmp   copyaux
diff --git a/src/fx/fx.dhgr.precomputed.1bit.a b/src/fx/fx.dhgr.precomputed.1bit.a
index 4e0b89e17..19bd86da5 100644
--- a/src/fx/fx.dhgr.precomputed.1bit.a
+++ b/src/fx/fx.dhgr.precomputed.1bit.a
@@ -55,11 +55,7 @@
          +BUILD_HGR_MIRROR_LOOKUP_TABLES hgrlomirror, hgr1himirror
          +BUILD_MIRROR_COLS mirror_cols
          +COPY_SELF_TO_AUXMEM
-         ldx   #(end-start)          ; copy InputLoop code to zero page
--        lda   start-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         +COPY_TO_0 start, end
          jmp   InputLoop
 start
 !pseudopc 0 {
@@ -67,7 +63,7 @@ Exit1Bit rts
 InputLoop
          ldy   #0
 input=*+1
-         ldx   .coords, y            ; first value: HGR row (only 0..95 will be in input array)
+         ldx   .coords               ; first value: HGR row (only 0..95 will be in input array)
          bmi   Exit1Bit              ; if > 127 then we're done
          +ROW_X_TO_BASE_ADDRESSES
          +ROW_X_TO_MIRROR_ADDRESSES
diff --git a/src/fx/fx.dhgr.precomputed.2bit.a b/src/fx/fx.dhgr.precomputed.2bit.a
index ca45f02c2..c233e5bc5 100644
--- a/src/fx/fx.dhgr.precomputed.2bit.a
+++ b/src/fx/fx.dhgr.precomputed.2bit.a
@@ -57,11 +57,7 @@
          +BUILD_HGR_LOOKUP_TABLES_X_IS_ALREADY_0 hgrlo, hgr1hi
          +BUILD_MIRROR_COLS mirror_cols
          +COPY_SELF_TO_AUXMEM
-         ldx   #(end-start)          ; copy InputLoop code to zero page
--        lda   start-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         +COPY_TO_0 start, end
          jmp   InputLoop
 
 start
@@ -70,7 +66,7 @@ Exit2Bit rts
 InputLoop
          ldy   #0
 input=*+1
-         ldx   .coords, y            ; first value: HGR row + 1
+         ldx   .coords               ; first value: HGR row + 1
          beq   Exit2Bit              ; if 0 then we're done
          +ROW_X_TO_2BIT_BASE_ADDRESSES
 
diff --git a/src/fx/fx.dhgr.r.by.pixel.a b/src/fx/fx.dhgr.r.by.pixel.a
index 6649c0279..c1c345284 100644
--- a/src/fx/fx.dhgr.r.by.pixel.a
+++ b/src/fx/fx.dhgr.r.by.pixel.a
@@ -11,12 +11,8 @@ copymasks=     $02C1                 ; [$07 bytes, different values in main and
 
          !source "src/fx/macros.a"
 
-         ldx   #(end-start)          ; copy code to zero page
--        lda   start-1, x
-         sta   $FF, x
-         dex
-         bne   -
-
+         +COPY_TO_0 start, end
+         ;X=0
          +BUILD_HGR_LOOKUP_TABLES_X_IS_ALREADY_0 hgrlo, hgr1hi
 
          ldy   #7                    ; copy copymask arrays into place in main and auxmem
diff --git a/src/fx/fx.dhgr.redlines.a b/src/fx/fx.dhgr.redlines.a
index bd345b9d2..4f1810aca 100644
--- a/src/fx/fx.dhgr.redlines.a
+++ b/src/fx/fx.dhgr.redlines.a
@@ -14,11 +14,8 @@ hgr1hi   =     $0300                 ; [$C0 bytes, main memory only]
 
          !source "src/fx/macros.a"
 
-         ldx   #(end-start)          ; copy code to zero page
--        lda   start-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         +COPY_TO_0 start, end
+         ;X=0
          +BUILD_HGR_LOOKUP_TABLES_X_IS_ALREADY_0 hgrlo, hgr1hi
          jmp   loop
 
diff --git a/src/fx/fx.gr.fizzle.a b/src/fx/fx.gr.fizzle.a
index b184275fc..e016691f1 100644
--- a/src/fx/fx.gr.fizzle.a
+++ b/src/fx/fx.gr.fizzle.a
@@ -5,11 +5,11 @@
 !to "build/GR.FIZZLE",plain
 *=$6000
 
-         ldx   #(end-start+1)        ; copy LFSR code to zero page
--        ldy   start-2, x
-         sty   $FE, x
-         dex
-         bne   -
+         !source "src/fx/macros.a"
+
+         +OVERCOPY_TO_0 start, end
+         ;X=0
+         ;Y=0
          jmp   loop
 
 start
diff --git a/src/fx/fx.hgr.1bit.fizzle.a b/src/fx/fx.hgr.1bit.fizzle.a
index 183d28b80..941e5997d 100755
--- a/src/fx/fx.hgr.1bit.fizzle.a
+++ b/src/fx/fx.hgr.1bit.fizzle.a
@@ -8,11 +8,9 @@
 addrs     = $6100                    ; [256 bytes, page-aligned]
 copymasks = $6200                    ; [256 bytes, page-aligned]
 
-         ldx   #(end-start)          ; copy LFSR code to zero page
--        lda   start-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         !source "src/fx/macros.a"
+
+         +COPY_TO_0 start, end
          ;X=0
          ;A=1
 --       ldy   #$20                  ; create copymask table
diff --git a/src/fx/fx.hgr.2bit.fizzle.a b/src/fx/fx.hgr.2bit.fizzle.a
index c64821a34..82c685c98 100644
--- a/src/fx/fx.hgr.2bit.fizzle.a
+++ b/src/fx/fx.hgr.2bit.fizzle.a
@@ -8,11 +8,10 @@
 addrs     = $6100                    ; [256 bytes, page-aligned]
 copymasks = $80                      ; [128 bytes, should not cross page boundary]
 
-         ldx   #(end-start)          ; copy LFSR code to zero page
--        lda   start-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         !source "src/fx/macros.a"
+
+         +COPY_TO_0 start, end
+         ;X=0
 --       ldy   #$20                  ; create copymask table
          ora   #%10000000
 -        sta   copymasks, x
diff --git a/src/fx/fx.hgr.checkerboard.fizzle.a b/src/fx/fx.hgr.checkerboard.fizzle.a
index 985847934..6143249b3 100644
--- a/src/fx/fx.hgr.checkerboard.fizzle.a
+++ b/src/fx/fx.hgr.checkerboard.fizzle.a
@@ -12,6 +12,8 @@ hgrlo     = $6400                    ; [$C0 bytes, should not cross page boundar
 hgrhi     = $6500                    ; [$C0 bytes, should not cross page boundary]
 shapemask = $8000                    ; [$2000 bytes, page-aligned]
 
+         !source "src/fx/macros.a"
+
          ldx   #0                    ; build an HGR base address lookup table,
 -        txa                         ; except the first address is |shapemask| instead of $2000
          and   #$F8
@@ -70,11 +72,8 @@ shapemask = $8000                    ; [$2000 bytes, page-aligned]
          inx
          cpx   #$C0
          bne   --
-         ldx   #(end-start)          ; copy LFSR code to zero page
--        lda   start-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         +COPY_TO_0 start, end
+         ;X=0
 --       ldy   #$20                  ; create copymask lookup table
          ora   #%10000000
 -        sta   copymasks, x
diff --git a/src/fx/fx.hgr.fizzle.a b/src/fx/fx.hgr.fizzle.a
index 8db424b38..2dc528b4d 100644
--- a/src/fx/fx.hgr.fizzle.a
+++ b/src/fx/fx.hgr.fizzle.a
@@ -5,11 +5,11 @@
 !to "build/FX.INDEXED/FIZZLE",plain
 *=$6000
 
-         ldx   #(end-start+1)        ; copy LFSR code to zero page
--        ldy   start-2, x
-         sty   $FE, x
-         dex
-         bne   -
+         !source "src/fx/macros.a"
+
+         +OVERCOPY_TO_0 start, end
+         ;X=0
+         ;Y=0
          jmp   loop
 
 start
diff --git a/src/fx/fx.hgr.palette.fizzle.a b/src/fx/fx.hgr.palette.fizzle.a
index d946f0f5a..6f97fdd54 100644
--- a/src/fx/fx.hgr.palette.fizzle.a
+++ b/src/fx/fx.hgr.palette.fizzle.a
@@ -7,12 +7,11 @@
 
          !source "src/fx/macros.a"   ; no code in here
 
-         ldx   #(end-start+1)        ; copy LFSR code to zero page
--        ldy   start-2, x
-         sty   $FE, x
-         dex
-         bne   -
+         +OVERCOPY_TO_0 start, end
+         ;X=0
+         ;Y=0
          bit   <exit
+         ;V=1
          jmp   loop
 
 start
diff --git a/src/fx/fx.hgr.precomputed.1bit.a b/src/fx/fx.hgr.precomputed.1bit.a
index da037ebd2..1a79ac857 100644
--- a/src/fx/fx.hgr.precomputed.1bit.a
+++ b/src/fx/fx.hgr.precomputed.1bit.a
@@ -267,11 +267,7 @@ aslmod   cmp   #$1A
          +BUILD_HGR_MIRROR_LOOKUP_TABLES hgrlomirror, hgr1himirror
          +BUILD_MIRROR_COLS mirror_cols
          +BUILD_SPARSE_BITMASKS copymasks, mirror_copymasks
-         ldx   #(end-start)          ; copy InputLoop code to zero page
--        lda   start-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         +COPY_TO_0 start, end
          jmp   InputLoop
 start
 !pseudopc 0 {
@@ -279,7 +275,7 @@ Exit1Bit rts
 InputLoop
          ldy   #0
 input=*+1
-         ldx   .coords, y            ; first value: HGR row (only 0..95 will be in input array)
+         ldx   .coords               ; first value: HGR row (only 0..95 will be in input array)
          bmi   Exit1Bit              ; if > 127 then we're done
          +ROW_X_TO_BASE_ADDRESSES
          +ROW_X_TO_MIRROR_ADDRESSES
diff --git a/src/fx/fx.hgr.precomputed.2bit.a b/src/fx/fx.hgr.precomputed.2bit.a
index 89fa89127..e097002f0 100644
--- a/src/fx/fx.hgr.precomputed.2bit.a
+++ b/src/fx/fx.hgr.precomputed.2bit.a
@@ -265,11 +265,7 @@ zerotbl  !byte $f0, $f2, $ca, $d2, $d8, $e0, $e2, $e6, $ea, $ee
          +BUILD_HGR_LOOKUP_TABLES hgrlo, hgr1hi
          +BUILD_MIRROR_COLS mirror_cols
          +BUILD_SPARSE_BITMASKS_2BIT copymasks, mirror_copymasks
-         ldx   #(end-start-1)        ; copy InputLoop code to zero page
--        lda   start, x
-         sta   $0, x
-         dex
-         bpl   -
+         +COPY_TO_0 start, end
          jmp   InputLoop
 start
 !pseudopc 0 {
@@ -277,7 +273,7 @@ Exit2Bit rts
 InputLoop
          ldy   #0
 input=*+1
-         ldx   .coords, y            ; first value: HGR row + 1
+         ldx   .coords               ; first value: HGR row + 1
          beq   Exit2Bit              ; if 0 then we're done
          +ROW_X_TO_2BIT_BASE_ADDRESSES
 
diff --git a/src/fx/fx.hgr.precomputed.3bit.a b/src/fx/fx.hgr.precomputed.3bit.a
index 9082695c4..bb42a9cb4 100644
--- a/src/fx/fx.hgr.precomputed.3bit.a
+++ b/src/fx/fx.hgr.precomputed.3bit.a
@@ -228,11 +228,7 @@ ripplezp
          +BUILD_3BIT_HGR_LOOKUP_TABLES
          +BUILD_EXTRA_COLS
          +BUILD_SPARSE_BITMASKS_3BIT
-         ldx   #(end-start-1)        ; copy InputLoop code to zero page
--        lda   start, x
-         sta   $0, x
-         dex
-         bpl   -
+         +COPY_TO_0 start, end
          jmp   InputLoop
 start
 !pseudopc 0 {
@@ -240,7 +236,7 @@ Exit3Bit rts
 InputLoop
          ldy   #0
 input=*+1
-         lda   .coords, y
+         lda   .coords
          bmi   Exit3Bit              ; if high bit is 1 then we're done
          cmp   #$40
          php
diff --git a/src/fx/fx.shr.2pass.lr.a b/src/fx/fx.shr.2pass.lr.a
index ae9ca1d88..6edefeae7 100644
--- a/src/fx/fx.shr.2pass.lr.a
+++ b/src/fx/fx.shr.2pass.lr.a
@@ -5,17 +5,13 @@
 !to "build/FX.INDEXED/SHR.TWOPASS.LR",plain
 *=$A000
 
+         !source "src/fx/macros.a"
          !source "src/fx/fx.shr.common.a"
 
 shrlo = $301                         ; $C8 bytes
 shrhi = $38                          ; $C8 bytes
 
-         ldx   #(end-start-1)        ; copy loop to zero page
--        lda   start, x
-         sta   $00, x
-         dex
-         bpl   -
-
+         +COPY_TO_0 start, end
          +BUILD_SHR_LOOKUP_TABLES shrlo, shrhi
          ;X=0
          +COPY_SCB_AND_PALETTES
diff --git a/src/fx/fx.shr.common.a b/src/fx/fx.shr.common.a
index b42d06750..b3ef30c39 100644
--- a/src/fx/fx.shr.common.a
+++ b/src/fx/fx.shr.common.a
@@ -64,3 +64,55 @@
          inx
          bne   -
 }
+
+!macro BUILD_SHR_MIRROR_ROWS_AND_COLS .mirror_rows, .mirror_cols {
+; out:   X=0
+;        Z=1
+         ldx   #$C6
+         ldy   #$00
+-        tya
+         sta   .mirror_cols-$27, x
+         sta   .mirror_rows-1, x
+         iny
+         dex
+         bne   -
+}
+
+!macro SHR_COPY_TO_STAGE_2 {
+; in:    X=0 /!\
+; out:   X=0
+;        Z=1
+;        WRITEMAINMEM active
+         +WRITE_MAIN
+-        lda   $A000, x
+         sta   $9D00, x
+         lda   $A100, x
+         sta   $9E00, x
+         inx
+         bne   -
+}
+
+!macro SHR_RESTORE_FROM_STAGE_2 {
+; out:   X=0
+;        Z=1
+         ldx   #$00
+-        lda   $9D00, x              ; copy main code back to $A000
+         sta   $A000, x              ; so it can be called again if necessary
+         lda   $9E00, x
+         sta   $A100, x
+         inx
+         bne   -
+}
+
+!macro LOAD_SHR_COORDINATES_AT .coords, .filename, .scratch {
+; out:   WRITEMAINMEM active
+;        LC RAM2 active and read/write
+         ldx   .filename             ; LOAD_FILE_AT macro destroys pathname
+-        lda   .filename, x          ; so we need to make a copy
+         sta   .scratch, x
+         dex
+         bpl   -
+
+         +READ_RAM2_WRITE_RAM2
+         +LOAD_FILE_AT .scratch, .coords
+}
diff --git a/src/fx/fx.shr.diagonal.a b/src/fx/fx.shr.diagonal.a
index 47d619dc6..c96ce1599 100644
--- a/src/fx/fx.shr.diagonal.a
+++ b/src/fx/fx.shr.diagonal.a
@@ -5,17 +5,13 @@
 !to "build/FX.INDEXED/SHR.DIAGONAL",plain
 *=$A000
 
-         !source "src/fx/fx.shr.common.a"
-
 shrlo = $301                         ; $C8 bytes
 shrhi = $201                         ; $C8 bytes
 
-         ldx   #(end-start-1)        ; copy loop to zero page
--        lda   start, x
-         sta   $00, x
-         dex
-         bpl   -
+         !source "src/fx/macros.a"
+         !source "src/fx/fx.shr.common.a"
 
+         +COPY_TO_0 start, end
          +BUILD_SHR_LOOKUP_TABLES shrlo, shrhi
          ;X=0
          +COPY_SCB_AND_PALETTES
diff --git a/src/fx/fx.shr.fizzle.a b/src/fx/fx.shr.fizzle.a
index 66f8c26f4..6e2a266c2 100644
--- a/src/fx/fx.shr.fizzle.a
+++ b/src/fx/fx.shr.fizzle.a
@@ -5,6 +5,7 @@
 !to "build/FX.INDEXED/SHR.FIZZLE",plain
 *=$A000
 
+         !source "src/fx/macros.a"
          !source "src/fx/fx.shr.common.a"
 
 addrs=$7F                            ; [128 bytes]
@@ -17,11 +18,7 @@ addrs=$7F                            ; [128 bytes]
          dex
          bne   -
 
-         ldx   #(end-start+1)        ; copy LFSR code to zero page
--        ldy   start-2, x
-         sty   $FE, x
-         dex
-         bne   -
+         +OVERCOPY_TO_0 start, end
          ;X=0
          ;Y=0
 
diff --git a/src/fx/fx.shr.lr.a b/src/fx/fx.shr.lr.a
index cfdd2d77c..795663fad 100644
--- a/src/fx/fx.shr.lr.a
+++ b/src/fx/fx.shr.lr.a
@@ -5,21 +5,16 @@
 !to "build/FX.INDEXED/SHR.LR",plain
 *=$A000
 
+         !source "src/fx/macros.a"
          !source "src/fx/fx.shr.common.a"
 
 shrlo = $301                         ; $C8 bytes
 shrhi = $37                          ; $C8 bytes
 
          +BUILD_SHR_LOOKUP_TABLES shrlo, shrhi
-
-         ldx   #(end-start+1)        ; copy loop to zero page
--        ldy   start-2, x
-         sty   $FE, x
-         dex
-         bne   -
+         +OVERCOPY_TO_0 start, end
          ;X=0
          ;Y=0
-
          +COPY_SCB_AND_PALETTES
          ;WRITEAUXMEM active
 
diff --git a/src/fx/fx.shr.lr2.a b/src/fx/fx.shr.lr2.a
index 59e1f7871..20628fa52 100644
--- a/src/fx/fx.shr.lr2.a
+++ b/src/fx/fx.shr.lr2.a
@@ -5,17 +5,13 @@
 !to "build/FX.INDEXED/SHR.LR2",plain
 *=$A000
 
-         !source "src/fx/fx.shr.common.a"
-
 shrlo = $301                         ; $C8 bytes
 shrhi = $38                          ; $C8 bytes
 
-         ldx   #(end-start-1)        ; copy loop to zero page
--        lda   start, x
-         sta   $00, x
-         dex
-         bpl   -
+         !source "src/fx/macros.a"
+         !source "src/fx/fx.shr.common.a"
 
+         +COPY_TO_0 start, end
          +BUILD_SHR_LOOKUP_TABLES shrlo, shrhi
          ;X=0
          +COPY_SCB_AND_PALETTES
diff --git a/src/fx/fx.shr.precomputed.a b/src/fx/fx.shr.precomputed.a
index 396026504..10bbf6dfa 100644
--- a/src/fx/fx.shr.precomputed.a
+++ b/src/fx/fx.shr.precomputed.a
@@ -8,63 +8,32 @@ mirror_cols = $1E29       ; $A0 bytes but clobbers $28 bytes before
 mirror_rows = $1F01       ; $C7 bytes
 coords   = $9F00          ; $1F41 bytes
 
-         !source "src/constants.a"
-         !source "src/macros.a"
+         !source "src/fx/macros.a"
          !source "src/fx/fx.shr.common.a"
 
 start
-         ldx   #$C6
-         ldy   #$00
--        tya
-         sta   mirror_cols-$27, x
-         sta   mirror_rows-1, x
-         iny
-         dex
-         bne   -
-
+         +BUILD_SHR_MIRROR_ROWS_AND_COLS mirror_rows, mirror_cols
+         ;X=0
          +BUILD_SHR_LOOKUP_TABLES shrlo, shrhi
          ;X=0
          +COPY_SCB_AND_PALETTES
-         ;WRITEAUXMEM active
-         +WRITE_MAIN
          ;X=0
--        lda   start, x
-         sta   $9D00, x
-         lda   start+$100, x
-         sta   $9E00, x
-         inx
-         bne   -
+         ;WRITEAUXMEM active
+         +SHR_COPY_TO_STAGE_2
          jmp   stage2
 
 !pseudopc *-$300 {
 stage2
-         ldx   CoordinatesFile           ; LOAD_FILE_AT macro destroys pathname
--        lda   CoordinatesFile, x        ; so we need to make a copy
-         sta   CoordinatesFileCopy, x
-         dex
-         bpl   -
+         +LOAD_SHR_COORDINATES_AT coords, CoordinatesFile, CoordinatesFileCopy
+         ;WRITEMAINMEM active
 
-         +READ_RAM2_WRITE_RAM2
-         +LOAD_FILE_AT CoordinatesFileCopy, coords
-         ;WRITEMAINMEM active after LOAD_FILE_AT macro
-
-         ldx   #(endzp-startzp)      ; copy loop code to zero page
--        lda   startzp-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         +COPY_TO_0 startzp, endzp
 
          +WRITE_AUX
          jsr   InputLoop
          +WRITE_MAIN
 
-         ldx   #$00
--        lda   $9D00, x              ; copy main code back to $A000
-         sta   start, x              ; so it can be called again if necessary
-         lda   $9E00, x
-         sta   start+$100, x
-         inx
-         bne   -
+         +SHR_RESTORE_FROM_STAGE_2
          +READ_RAM1_WRITE_RAM1
          rts
 
diff --git a/src/fx/fx.shr.radial.a b/src/fx/fx.shr.radial.a
index f98b9470e..bea96c1f1 100644
--- a/src/fx/fx.shr.radial.a
+++ b/src/fx/fx.shr.radial.a
@@ -13,65 +13,35 @@ mirror_rows    = $1F01    ; $C7 bytes
 coords         = $9F00    ; $1F41 bytes
 last_coords = coords+$1F3E
 
-         !source "src/constants.a"
-         !source "src/macros.a"
+         !source "src/fx/macros.a"
          !source "src/fx/fx.shr.common.a"
 
 start
-         ldx   #$C6
-         ldy   #$00
--        tya
-         sta   mirror_cols-$27, x
-         sta   mirror_rows-1, x
-         iny
-         dex
-         bne   -
-
+         +BUILD_SHR_MIRROR_ROWS_AND_COLS mirror_rows, mirror_cols
+         ;X=0
          +BUILD_SHR_LOOKUP_TABLES shrlo, shrhi
          ;X=0
          +COPY_SCB_AND_PALETTES
-         ;WRITEAUXMEM active
-         +WRITE_MAIN
          ;X=0
--        lda   start, x
-         sta   $9D00, x
-         lda   start+$100, x
-         sta   $9E00, x
-         inx
-         bne   -
+         ;WRITEAUXMEM active
+         +SHR_COPY_TO_STAGE_2
          jmp   stage2
 
 !pseudopc *-$300 {
 stage2
-         ldx   CoordinatesFile           ; LOAD_FILE_AT macro destroys pathname
--        lda   CoordinatesFile, x        ; so we need to make a copy
-         sta   CoordinatesFileCopy, x
-         dex
-         bpl   -
+         +LOAD_SHR_COORDINATES_AT coords, CoordinatesFile, CoordinatesFileCopy
+         ;WRITEMAINMEM active
 
-         +READ_RAM2_WRITE_RAM2
-         +LOAD_FILE_AT CoordinatesFileCopy, coords
-         ;WRITEMAINMEM active after LOAD_FILE_AT macro
          lda   #$80
          sta   coords-2
 
-         ldx   #(endzp-startzp)      ; copy loop code to zero page
--        lda   startzp-1, x
-         sta   $FF, x
-         dex
-         bne   -
+         +COPY_TO_0 startzp, endzp
 
          +WRITE_AUX
          jsr   LoopBL
          +WRITE_MAIN
 
-         ldx   #$00
--        lda   $9D00, x              ; copy main code back to $A000
-         sta   start, x              ; so it can be called again if necessary
-         lda   $9E00, x
-         sta   start+$100, x
-         inx
-         bne   -
+         +SHR_RESTORE_FROM_STAGE_2
          +READ_RAM1_WRITE_RAM1
          rts
 
diff --git a/src/fx/fx.shr.ud.a b/src/fx/fx.shr.ud.a
index 910eae88b..fe467c3ff 100644
--- a/src/fx/fx.shr.ud.a
+++ b/src/fx/fx.shr.ud.a
@@ -5,19 +5,15 @@
 !to "build/FX.INDEXED/SHR.UD",plain
 *=$A000
 
-         !source "src/fx/fx.shr.common.a"
-
 shrlo = $301                         ; $C8 bytes
 shrhi = $38                          ; $C8 bytes
 reverseshrlo = $BD01                 ; $C8 bytes
 reverseshrhi = $BE01                 ; $C8 bytes
 
-         ldx   #(end-start-1)        ; copy loop to zero page
--        lda   start, x
-         sta   $00, x
-         dex
-         bpl   -
+         !source "src/fx/macros.a"
+         !source "src/fx/fx.shr.common.a"
 
+         +COPY_TO_0 start, end
          +BUILD_SHR_REVERSE_LOOKUP_TABLES reverseshrlo, reverseshrhi
          +BUILD_SHR_LOOKUP_TABLES shrlo, shrhi
          ;X=0
diff --git a/src/fx/macros.a b/src/fx/macros.a
index b3a76d63e..c60890be9 100644
--- a/src/fx/macros.a
+++ b/src/fx/macros.a
@@ -161,5 +161,27 @@
 +
 }
 
+!macro COPY_TO_0 .start, .end {
+; out:   X=0
+;        Z=1
+         ldx   #(.end-.start)
+-        lda   .start-1, x
+         sta   $FF, x
+         dex
+         bne   -
+}
+
+!macro OVERCOPY_TO_0 .start, .end {
+; over-copy region to $00
+; clobbers $FF
+; out:   X=0
+;        Y=last byte before start (e.g. 0 if the last instruction is JMP $0000)
+         ldx   #(.end-.start+1)
+-        ldy   .start-2, x
+         sty   $FE, x
+         dex
+         bne   -
+}
+
 _FX_MACROS_=*
 }