Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!henry From: henry@utzoo.UUCP (Henry Spencer) Newsgroups: net.lang.c,net.micro.pc Subject: Re: Lattice versus CI86 Message-ID: <4856@utzoo.UUCP> Date: Wed, 2-Jan-85 13:07:40 EST Article-I.D.: utzoo.4856 Posted: Wed Jan 2 13:07:40 1985 Date-Received: Wed, 2-Jan-85 13:07:40 EST References: <233@gumby.UUCP> <3194@alice.UUCP> <6267@boring.UUCP>, <5022@rochester.UUCP> Organization: U of Toronto Zoology Lines: 15 > > - Doesn't implement ++ and -- for float/double > > While ++ and -- on float/double suggest they mean the same thing as > {+-}= 1.0, some compilers object to this, with some justification. > This is a non-portable construct. {+-}= 1.0 is clearer to read anyway. It's more than just suggested. K&R, section 7.2, says in so many words that "++x" is equivalent to "x+=1". There is nothing ill-defined or implementation-dependent about the semantics of the latter, even when applied to floating-point. "--x", "x++", and "x--" are defined with reference to "++x". A compiler which does not implement the ++ and -- operators on floating-point operands is wrong, period. -- Henry Spencer @ U of Toronto Zoology {allegra,ihnp4,linus,decvax}!utzoo!henry