From: Paul Khuong <pvk@pv...>  20091117 13:05:59

On 20091117, at 7:38 AM, alexander a donkov wrote: > sorry to bother you again about this topic. > i took your word that it would not be hard to change the complex+/real operation for my needs. > My question is besides the floattran.lisp file where else should i look to make that change? > I did change the four lines in floattran.lisp with explicit coerce 0, but after compilation it still adds the real and complex as if promoting the real number to a complex with zero im part. in src/compiler/x8664/float.lisp, look for (defrealcomplexop + t nil addps +/realcomplexsinglefloat +/complexrealsinglefloat 3 addpd +/realcomplexdoublefloat +/complexrealdoublefloat 4) (defrealcomplexop  nil nil subps /realcomplexsinglefloat /complexrealsinglefloat 3 subpd /realcomplexdoublefloat /complexrealdoublefloat 4) and replace that with (defrealcomplexop + t nil addss +/realcomplexsinglefloat +/complexrealsinglefloat 3 addsd +/realcomplexdoublefloat +/complexrealdoublefloat 4) (defrealcomplexop  nil nil subss /realcomplexsinglefloat /complexrealsinglefloat 3 subsd /realcomplexdoublefloat /complexrealdoublefloat 4) Save, and start a fresh build. Paul Khuong 