[decomp] shadow-cpu-h defines shadow-edge twice (and differently)

[xTVaser]

Kinda interesting...I presume in the original code, the second definition won out, but our decompiler/compiler doesn't handle this very well

lw v1, type(s7)           ;; [118] (set! v1-11 type) [] -> [v1: <the etype type> ]
    lwu t9, 16(v1)            ;; [119] (set! t9-21 (l.wu (+ v1-11 16)))
                              ;; [v1: <the etype type> ] -> [t9: (function symbol type int type) ]
    daddiu a0, s7, shadow-edge;; [120] (set! a0-21 'shadow-edge) [] -> [a0: symbol ]
    lw a1, structure(s7)      ;; [121] (set! a1-21 structure) [] -> [a1: <the etype structure> ]
    ld a2, L94(fp)            ;; [122] (set! a2-20 (l.d L94)) [] -> [a2: uint ]
    jalr ra, t9               ;; [123] (call! a0-21 a1-21 a2-20)
...
lw v1, type(s7)           ;; [140] (set! v1-13 type) [] -> [v1: <the etype type> ]
    lwu t9, 16(v1)            ;; [141] (set! t9-25 (l.wu (+ v1-13 16)))
                              ;; [v1: <the etype type> ] -> [t9: (function symbol type int type) ]
    daddiu a0, s7, shadow-edge;; [142] (set! a0-25 'shadow-edge) [] -> [a0: symbol ]
    lw a1, structure(s7)      ;; [143] (set! a1-25 structure) [] -> [a1: <the etype structure> ]
    ld a2, L97(fp)            ;; [144] (set! a2-24 (l.d L97)) [] -> [a2: uint ]
    jalr ra, t9               ;; [145] (call! a0-25 a1-25 a2-24)
                              ;; [a0: symbol a1: <the etype structure> a2: uint t9: (function symbol type int type) ] -> [v0: type ]

[water111]

I worked around this in #1246
It's not ideal, but I don't think it's worth the effort to fix because it happened only once.