Cpp check ругается, можно ли так писать? Опции

[ViKo]

Имею массив из float и указатель на него.

Код

  pSigRe = SigRe;
  for (uint32_t i = FFT_N; i--; ) {
    *pSigRe++ = log2f(*pSigRe) * 15.0515;
  }

При проверке cpp check выдает ошибку.
Кратко: Expression '*pSigRe++=log2f(*pSigRe)*15.0515' depends on order of evaluation of side effects
Сообщение: Expression '*pSigRe++=log2f(*pSigRe)*15.0515' depends on order of evaluation of side effects
Точно, так нельзя писать?

[mcheb]

Решил проверить в GCC

Код

    uint16_t i;
    float SigRe[16],*pSigRe;
    pSigRe = SigRe;
    for(i=0; i<16; i++) SigRe[i] = 2.*(i+1);
    for(i=0; i<16; i++) *pSigRe++ = log2f(*pSigRe) * 15.0515;
    for(i=0; i<16; i++) printf("i = %u SigRe[%u] = %f12.5\n",i,i,SigRe[i]);

    for(i=0; i<16; i++) SigRe[i] = 2.*(i+1);
    for(i=0; i<16; i++) SigRe[i] = log2f(SigRe[i]) * 15.0515;
    for(i=0; i<16; i++) printf("i = %u SigRe[%u] = %f12.5\n",i,i,SigRe[i]);

Получил

Код

i = 0 SigRe[0] = 30.10300112.5
i = 1 SigRe[1] = 38.90756212.5
i = 2 SigRe[2] = 45.15449912.5
i = 3 SigRe[3] = 50.00000012.5
i = 4 SigRe[4] = 53.95906412.5
i = 5 SigRe[5] = 57.30640412.5
i = 6 SigRe[6] = 60.20600112.5
i = 7 SigRe[7] = 62.76362612.5
i = 8 SigRe[8] = 65.05149812.5
i = 9 SigRe[9] = 67.12113212.5
i = 10 SigRe[10] = 69.01056712.5
i = 11 SigRe[11] = 70.74866512.5
i = 12 SigRe[12] = 72.35790312.5
i = 13 SigRe[13] = 73.85606412.5
i = 14 SigRe[14] = 75.25750012.5
i = 15 SigRe[15] = -1913.32995612.5
i = 0 SigRe[0] = 15.05150012.5
i = 1 SigRe[1] = 30.10300112.5
i = 2 SigRe[2] = 38.90756212.5
i = 3 SigRe[3] = 45.15449912.5
i = 4 SigRe[4] = 50.00000012.5
i = 5 SigRe[5] = 53.95906412.5
i = 6 SigRe[6] = 57.30640412.5
i = 7 SigRe[7] = 60.20600112.5
i = 8 SigRe[8] = 62.76362612.5
i = 9 SigRe[9] = 65.05149812.5
i = 10 SigRe[10] = 67.12113212.5
i = 11 SigRe[11] = 69.01056712.5
i = 12 SigRe[12] = 70.74866512.5
i = 13 SigRe[13] = 72.35790312.5
i = 14 SigRe[14] = 73.85606412.5
i = 15 SigRe[15] = 75.25750012.5
Hello world!

Ругается на
Prj\Hello\main.c:12:32: warning: operation on 'pSigRe' may be undefined [-Wsequence-point]
for(i=0; i<16; i++) *pSigRe++ = log2f(*pSigRe) * 15.0515;
Выходит, что нельзя так.