Yes, short int is 16-bit. Also, yes it does produce correct results for 0..32767. Thanks. -----Original Message----- From: John (Eljay) Love-Jensen [mailto:eljay@xxxxxxxxx] Sent: Wednesday, April 14, 2010 1:52 PM To: Gederberg, Thomas K; GCC-help Subject: Re: Problem with cast from double to unsigned short int Hi Thomas, I presume that short on your platform is 16-bit. Also, I presume that this line... is = (short int) ((double)(ius)); ...yields correct results when ius is 0 through 32767. Is that correct? I think that when "ius > 32767" on such a platform, the result is undefined behavior. If I am correct: as such, -32768 is a reasonable value. So is -13000. But I may be mistaken, and hopefully someone can site ISO 14882 chapter/verse. > Is this a bug with gcc? No, I think it is a bug in the code. Sincerely, --Eljay
