Potentially faulty line of code in nb_celp.c #37
Description
At the line 765 in nb_celpc.c it is the line of code "int qe = scal_quant32(ol_gain, ol_gain_table, 32);". The problems with it is that I am not sure if qe is supposed to fit in 32 bits (I know it should because scal_quant32 is used only there), so an explicit data type spx_word32_t would be helpful to know what results the function returns me :), but the other problem is that on some platforms int has less than 32 bits, so it will generate bit true problems. In this case I should declare qe to be a large enough data type; thing I shouldn't do if the data type was more explicit (spx_word32_t).
I have not searched for a specific set of parameters to make scal_quant32 return a value which will be wraped on a system with an int less than 32 bits, but the name alone of that function and its single use case tells me there exists such values for its parameters (to cause a bit true problem between native and platform).