; E04-S04: control-flow emitter fixture in x128.
;
; The fixture generates the same routine twice with flow.s: first in counting
; mode (a sizing pass that writes no bytes), then in emitting mode. Both
; passes must produce identical label offsets and total lengths. The emitted
; routine is then executed and must compute:
;
;   a = 0; while a < 5: a = a + 1            -> a = 5
;   if a = 5 then path = 10 else path = 20   -> path = 10
;   for i = 1 to 3: sum = sum + i            -> sum = 16, i = 4
;   sum = add(sum)                           -> sum = 17 (call + return)
;   return sum                               -> A = 17
;
; A separate routine takes a long conditional branch (distance >= 128) over
; 125 bytes of filler and must reach its target. The fixture also checks the
; five-byte branch encoding: short form has the distance byte plus three NOPs,
; long form has +3 and a live JMP.
;
; Mailbox data:
;   +0     program result (17)
;   +1     a (5)
;   +2     if/else path marker (10)
;   +3     i after the for loop (4)
;   +4     long-branch result (5A)
;   +5     long-branch wrong-path marker (0)
;   +6     label/length mismatches between the sizing and emission passes (0)
;   +7     branch-form bits: 1 short form verified, 2 long form verified (3)
; Failure bits: 1 sizing, 2 program result, 4 if/else, 8 for loop,
; 16 call/return, 32 long branch, 64 branch bytes, 128 emitter error.

.setcpu "6502"
.include "../../src/common/c128.inc"
.include "../../src/common/mailbox.inc"
.include "../../src/common/codegen.inc"
.include "../../src/common/flow.inc"
.include "../../src/common/frames.inc"

.import cg_init, cg_len, cg_err, cg_code, cg_here
.import cg_lda_imm, cg_ldx_imm, cg_lda_abs, cg_sta_abs
.import cg_adc_abs, cg_cmp_abs, cg_jsr_abs
.import cg_clc, cg_sec, cg_rts
.import fl_init, fl_label, fl_jmp, fl_branch, fl_finish, fl_get_label
.import fl_branch_true, fl_call_byte, fl_prologue, fl_ret_byte
.import fl_err, fl_enter, fl_leave, fl_discard
.import frame_init, frame_enter, frame_leave, frame_discard, frame_push_word

; Label ids (reset by every fl_init).
L_WHILE         = 0
L_ENDWHILE      = 1
L_ELSE          = 2
L_ENDIF         = 3
L_FOR           = 4
L_ENDFOR        = 5
L_ERR           = 6
L_LONG          = 7
LABEL_IDS       = 7            ; ids 0..6 belong to the main routine

.macro EMIT_LDA addr
        lda #<(addr)
        ldx #>(addr)
        jsr cg_lda_abs
.endmacro

.macro EMIT_STA addr
        lda #<(addr)
        ldx #>(addr)
        jsr cg_sta_abs
.endmacro

.macro EMIT_ADC addr
        lda #<(addr)
        ldx #>(addr)
        jsr cg_adc_abs
.endmacro

.macro EMIT_CMP addr
        lda #<(addr)
        ldx #>(addr)
        jsr cg_cmp_abs
.endmacro

.macro EMIT_LDA_IMM value
        lda #value
        jsr cg_lda_imm
.endmacro

.macro EMIT_LDX_IMM value
        lda #value
        jsr cg_ldx_imm
.endmacro

.macro EMIT_JSR addr
        lda #<(addr)
        ldx #>(addr)
        jsr cg_jsr_abs
.endmacro

; Branch to a global error label even when it is out of relative range.
.macro ON_CARRY target
        bcc :+
        jmp target
:
.endmacro

.segment "LOADADDR"
        .word $1C01

.segment "EXEHDR"
        .word basic_end
        .word 10
        .byte $9E
        .byte "7181"
        .byte $00
basic_end:
        .word $0000

.segment "MAILBOX"
        .res MB_SIZE

.segment "STARTUP"
.export test_entry
test_entry:
        php
        sei
        ldx #MB_SIZE - 1
        lda #$00
@clr:   sta MB_BASE,x
        dex
        bpl @clr
        lda #MB_MAGIC0_VALUE
        sta MB_MAGIC0
        lda #MB_MAGIC1_VALUE
        sta MB_MAGIC1
        lda #MB_PROD_PROGRAM
        sta MB_PRODUCER
        lda #MB_ST_RUNNING
        sta MB_STATUS
        lda #$00
        sta fail_mask
        sta mismatch
        sta form_mask
        sta MB_CODE
        sta MB_CODE+1
        jsr init_vars

        ; Runtime addresses for the emitted prologue/call/epilogue.
        lda #<frame_enter
        sta fl_enter
        lda #>frame_enter
        sta fl_enter+1
        lda #<frame_leave
        sta fl_leave
        lda #>frame_leave
        sta fl_leave+1
        lda #<frame_discard
        sta fl_discard
        lda #>frame_discard
        sta fl_discard+1

        lda MMU_CR
        sta saved_cr
        lda #CR_RAM0_KERNAL
        sta MMU_CR

        ; ------------------------------------------- main routine: sizing pass
        lda #FL_MODE_COUNT
        jsr fl_init
        jsr gen_main
        ON_CARRY emitter_error
        jsr fl_finish
        ON_CARRY emitter_error
        lda cg_len
        sta main_size
        jsr save_labels

        ; ------------------------------------------ main routine: emit pass 2
        jsr frame_init
        lda #FL_MODE_EMIT
        jsr fl_init
        jsr gen_main
        ON_CARRY emitter_error
        jsr fl_finish
        ON_CARRY emitter_error
        lda cg_len
        cmp main_size
        beq @len_ok
        inc mismatch
        lda #$01
        jsr fail
@len_ok:
        jsr compare_labels

        ; short branch: distance byte plus three NOPs
        ldy pos_while
        lda cg_code,y
        cmp #128
        bcs @short_bad
        iny
        lda cg_code,y
        cmp #$EA
        bne @short_bad
        iny
        lda cg_code,y
        cmp #$EA
        bne @short_bad
        iny
        lda cg_code,y
        cmp #$EA
        bne @short_bad
        lda #$01
        sta form_mask
        jmp @run_main
@short_bad:
        lda #$40
        jsr fail

        ; ------------------------------------------------ execute the routine
@run_main:
        jsr frame_init
        lda #$00
        jsr run_code
        sta prog_result

        lda prog_result
        cmp #17
        beq @check_a
        lda #$02
        jsr fail
@check_a:
        lda v_a
        cmp #5
        beq @check_if
        lda #$04
        jsr fail
@check_if:
        lda v_ifpath
        cmp #10
        beq @check_for
        lda #$04
        jsr fail
@check_for:
        lda v_i
        cmp #4
        beq @check_call
        lda #$08
        jsr fail
@check_call:
        lda v_sum
        cmp #17
        beq @long_begin
        lda #$10
        jsr fail

        ; ------------------------------- long branch routine: sizing + emit
@long_begin:
        lda #FL_MODE_COUNT
        jsr fl_init
        jsr gen_long
        ON_CARRY emitter_error
        jsr fl_finish
        ON_CARRY emitter_error
        lda cg_len
        sta long_size

        lda #FL_MODE_EMIT
        jsr fl_init
        jsr gen_long
        ON_CARRY emitter_error
        jsr fl_finish
        ON_CARRY emitter_error
        lda cg_len
        cmp long_size
        beq @long_len_ok
        inc mismatch
        lda #$01
        jsr fail
@long_len_ok:
        ; long form: +3 then JMP, and the target is at least 128 bytes away
        ldy pos_long
        lda cg_code,y
        cmp #$03
        bne @long_bad
        iny
        lda cg_code,y
        cmp #$4C
        bne @long_bad
        ldy pos_long
        dey
        lda cg_code,y
        cmp #$F0                       ; BNE was inverted to BEQ
        bne @long_bad
        lda #L_LONG
        jsr fl_get_label
        bcs @long_bad
        txa
        sec
        sbc pos_long
        sec
        sbc #$01
        cmp #128
        bcc @long_bad
        lda form_mask
        ora #$02
        sta form_mask
        jmp @run_long
@long_bad:
        lda #$40
        jsr fail

@run_long:
        lda #$00
        jsr run_code
        lda v_long
        cmp #$5A
        beq @check_bad
        lda #$20
        jsr fail
@check_bad:
        lda v_bad
        beq report
        lda #$20
        jsr fail

        jmp report

; ---------------------------------------------------------- report
report:
        lda saved_cr
        sta MMU_CR
        lda prog_result
        sta MB_DATA+0
        lda v_a
        sta MB_DATA+1
        lda v_ifpath
        sta MB_DATA+2
        lda v_i
        sta MB_DATA+3
        lda v_long
        sta MB_DATA+4
        lda v_bad
        sta MB_DATA+5
        lda mismatch
        sta MB_DATA+6
        lda form_mask
        sta MB_DATA+7
        lda fail_mask
        sta MB_CODE
        lda #$00
        sta MB_CODE+1
        lda #MB_ST_FAIL
        ldx fail_mask
        bne @status
        lda #MB_ST_PASS
@status:
        sta MB_STATUS
        ldy #$00
@print: lda message,y
        beq @end
        sty print_y
        jsr CHROUT
        ldy print_y
        iny
        bne @print
@end:   plp
        rts

emitter_error:
        lda #$80
        jsr fail
        jmp report

; ------------------------------------------------------------- generators
; if/else/while/for, one call, and a value return.
gen_main:
        lda #$00                       ; prologue: no locals
        ldx #$00
        ldy #L_ERR
        jsr fl_prologue
        ON_CARRY gen_out

        EMIT_LDA_IMM $00               ; a = 0
        EMIT_STA v_a
        EMIT_LDA_IMM $00               ; sum = 0
        EMIT_STA v_sum

        lda #L_WHILE                   ; while a < 5
        jsr fl_label
        EMIT_LDA v_a
        EMIT_CMP v_five
        jsr cg_here                    ; the exit branch's distance byte
        clc
        adc #$01
        sta pos_while
        lda #$B0                       ; BCS: a >= 5 leaves the loop
        ldx #L_ENDWHILE
        jsr fl_branch
        ON_CARRY gen_out
        EMIT_LDA v_a                   ; a = a + 1
        jsr cg_clc
        EMIT_ADC v_one
        EMIT_STA v_a
        lda #L_WHILE                   ; back edge: absolute jump
        jsr fl_jmp
        ON_CARRY gen_out
        lda #L_ENDWHILE
        jsr fl_label

        EMIT_LDA v_a                   ; if a = 5
        EMIT_CMP v_five
        lda #$D0                       ; BNE: else
        ldx #L_ELSE
        jsr fl_branch
        ON_CARRY gen_out
        EMIT_LDA_IMM 10
        EMIT_STA v_ifpath
        EMIT_STA v_sum
        lda #L_ENDIF
        jsr fl_jmp
        ON_CARRY gen_out
        lda #L_ELSE
        jsr fl_label
        EMIT_LDA_IMM 20
        EMIT_STA v_ifpath
        EMIT_STA v_sum
        lda #L_ENDIF
        jsr fl_label

        EMIT_LDA_IMM 1                 ; for i = 1 to 3
        EMIT_STA v_i
        lda #L_FOR
        jsr fl_label
        EMIT_LDA v_i
        EMIT_CMP v_four
        lda #$B0                       ; BCS: i >= 4 leaves the loop
        ldx #L_ENDFOR
        jsr fl_branch
        ON_CARRY gen_out
        EMIT_LDA v_sum                 ; sum = sum + i
        jsr cg_clc
        EMIT_ADC v_i
        EMIT_STA v_sum
        EMIT_LDA v_i                   ; i = i + 1
        jsr cg_clc
        EMIT_ADC v_one
        EMIT_STA v_i
        lda #L_FOR
        jsr fl_jmp
        ON_CARRY gen_out
        lda #L_ENDFOR
        jsr fl_label

        EMIT_LDA v_sum                 ; push(sum) for add()
        EMIT_LDX_IMM 0
        EMIT_JSR frame_push_word
        ON_CARRY gen_out
        lda #$B0                       ; push refused a frame
        ldx #L_ERR
        jsr fl_branch
        ON_CARRY gen_out
        lda #<add                      ; sum = add(sum)
        ldx #>add
        ldy #2
        jsr fl_call_byte
        ON_CARRY gen_out
        EMIT_STA v_sum

        EMIT_LDA v_sum                 ; return sum
        jsr fl_ret_byte
        ON_CARRY gen_out

        lda #L_ERR                     ; frame or argument-push failure
        jsr fl_label
        lda #$00
        jsr cg_sec
        jmp cg_rts

; A long taken branch over 125 bytes of filler.
gen_long:
        EMIT_LDA_IMM $00
        EMIT_STA v_bad
        EMIT_LDA_IMM $01
        jsr cg_here                    ; remember the branch's distance byte
        clc
        adc #$01
        sta pos_long
        lda #L_LONG
        jsr fl_branch_true
        ON_CARRY gen_out
        EMIT_LDA_IMM $AA               ; wrong path if the branch fails
        EMIT_STA v_bad
        lda #25                        ; X is clobbered by the emit macros
        sta fill_count
@fill:  EMIT_LDA_IMM $11
        EMIT_STA v_filler
        dec fill_count
        bne @fill
        lda #L_LONG
        jsr fl_label
        EMIT_LDA_IMM $5A
        EMIT_STA v_long
        jmp cg_rts

; Shared early exit for both generators: C=1 from the failing call.
gen_out:
        rts

; ---------------------------------------------------------------- helpers
; A = code-buffer offset. Run the snippet and return its A.
run_code:
        clc
        adc #<cg_code
        sta @target+1
        lda #>cg_code
        adc #$00
        sta @target+2
@target:
        jsr $FFFF
        rts

save_labels:
        lda #$00
        sta label_id
@loop:  lda label_id
        cmp #LABEL_IDS
        beq @done
        jsr fl_get_label
        bcs @missing
        ldy label_id
        txa
        sta saved_labels,y
        inc label_id
        jmp @loop
@missing:
        inc mismatch
        lda #$01
        jsr fail
        inc label_id
        jmp @loop
@done:  rts

compare_labels:
        lda #$00
        sta label_id
@loop:  lda label_id
        cmp #LABEL_IDS
        beq @done
        jsr fl_get_label
        bcs @bad
        ldy label_id
        txa
        cmp saved_labels,y
        beq @next
@bad:   inc mismatch
        lda #$01
        jsr fail
@next:  inc label_id
        jmp @loop
@done:  rts

init_vars:
        lda #$01
        sta v_one
        lda #$03
        sta v_three
        lda #$04
        sta v_four
        lda #$05
        sta v_five
        lda #$00
        sta v_a
        sta v_sum
        sta v_i
        sta v_ifpath
        sta v_bad
        sta v_long
        sta v_filler
        sta prog_result
        rts

; add(value as word) as byte: returns value + 1.
add:
        lda #$00
        ldx #$00
        jsr frame_enter
        bcs @fault
        ldy #2
        lda (FRAME_FP),y
        clc
        adc #$01
        sta add_result
        jsr frame_leave
        lda add_result
        clc
        rts
@fault: lda #$00
        sec
        rts

.segment "CODE"
fail:
        ora fail_mask
        sta fail_mask
        rts

.segment "RODATA"
message:        .byte "NOVA128 CONTROL TEST DONE", $0D, $00

.segment "BSS"
fail_mask:      .res 1
print_y:        .res 1
saved_cr:       .res 1
label_id:       .res 1
mismatch:       .res 1
form_mask:      .res 1
main_size:      .res 1
long_size:      .res 1
pos_while:      .res 1
pos_long:       .res 1
prog_result:    .res 1
v_one:          .res 1
v_three:        .res 1
v_four:         .res 1
v_five:         .res 1
v_a:            .res 1
v_sum:          .res 1
v_i:            .res 1
v_ifpath:       .res 1
v_bad:          .res 1
v_long:         .res 1
v_filler:       .res 1
add_result:     .res 1
saved_labels:   .res LABEL_IDS
fill_count:     .res 1