paintchars Fonts02 Does this look right?

I was trying to find what I’ve messed up such that each character of a string text all appears plotted at the same (wrong) place, and I came across this.

; set up terminating address of string from R7, else use -1 if R2 bit 7 clear

        MOV     wrchbase,R1                     ; wrchbase (R11) -> points to string
        TST     R2,#paint_length
        ADDNE   R7,wrchbase,R7
        MOVEQ   R7,#&FFFFFFFF                   ; must use unsigned comparison, and ignore V set
        STR     R7,paintendaddress

; convert coordinates to millipoints (input units depend on R2 bits 4,5)

        Push    "ycoord,xcoord"                 ; these are pulled MUCH later!

        TST     R2,#paint_osunits               ; 1 ==> convert to 1/72000th"
        STREQ   R3,xco72                        ; coords in 1/72000th inch
        STREQ   R4,yco72

The problem is that xcoord is R8 (OK) and ycoord is R7, which is now the pointer to the end of the string or all ones. This value gets carefully preserved on the stack for later use… R3 and R4 actually contain the x and y coordinates I want, in OS units, and the conversion is done after the push…

A disassembly, annotated:

fc2677b8:       e1a0b001        mov     fp, r1                                                                         
fc2677bc:       e3120080        tst     r2, #128        ; 0x80                                                         
fc2677c0:       108b7007        addne   r7, fp, r7                                                                     
fc2677c4:       03e07000        mvneq   r7, #0                                                                         
fc2677c8:       e58c74c4        str     r7, [ip, #1220] ; 0x4c4 paintendaddress                                        
fc2677cc:       e92d0180        push    {r7, r8}                        ycoord,xcoord                                  
fc2677d0:       e3120010        tst     r2, #16                 paintosunits                                           
fc2677d4:       058c33ac        streq   r3, [ip, #940]  ; 0x3ac                                                        
fc2677d8:       058c43b0        streq   r4, [ip, #944]  ; 0x3b0                                                        
fc2677dc:       0b00022d        bleq    fc268098 calcxcoord                                                            

It doesn’t seem to cause any problems, but I can’t see what’s going on.

The code is at 0xfc2677b8 in 5.28, (19-Oct-20).

Another interpretation is that it’s pushing whichever registers xcoord,ycoord happen to be in order to avoid them being corrupted by the subsequent millipoint calculation?

I think it’s probably redundant debug information; the only thing that happens to it (unless its input to a function) is that it may be used for debug output.