sizeof - C++ fread: unsigned char and short mixture -
i have sample code here:
unsigned char *m_fbytes; m_fbytes = (unsigned char*)malloc(m_ibytelen1framedecoded*sizeof(short)); int err; err = fread(m_fbytes, sizeof(short), 960, fin); curr_read = err; for(int i=0;i<curr_read;i++) { opus_int32 s; s=m_fbytes[2*i+1]<<8|m_fbytes[2*i]; s=((s&0xffff)^0x8000)-0x8000; m_in[i]=s; } int ilen = encode(m_enc, m_in, m_ibytelen1framedecoded, m_data, m_max_payload_bytes); i don't understand line:
err = fread(m_fbytes, sizeof(short), 960, fin); the authors of code read bytes (aka unsigned char), pass sizeof(short).
why use "sizeof(short)"?
the code contains couple of tacit assumptions make unnecessarily fragile. way read bytes used
s=m_fbytes[2*i+1]<<8|m_fbytes[2*i]; s=((s&0xffff)^0x8000)-0x8000; the assumptions are
char_bit == 8; that's pretty safe assumption nowadays, unless 1 deals exotic hardware, it's not guaranteed.sizeof(short) == 2; that's true, not guaranteed either.
the way read bytes used, each adjacent pair of them make 1 unit in final result (and using 2 instead of sizeof(short) more robust, if sizeof(short) > 2, use part of read data [if sizeof(short) == 1, char_bit must @ least 16, in case stuff break in different way]).
the size parameter fread used specify how many bytes make 1 unit, no partial units read, , buffer contains err complete units.
Comments
Post a Comment