Re: C arithmetic, was Software proofs, was Are there different

From	Keith Thompson <Keith.S.Thompson+u@gmail.com>
Newsgroups	comp.compilers
Subject	Re: C arithmetic, was Software proofs, was Are there different
Date	2023-02-05 16:14 -0800
Organization	None to speak of
Message-ID	<23-02-025@comp.compilers> (permalink)
References	<23-01-092@comp.compilers> <23-02-003@comp.compilers> <23-02-019@comp.compilers>

Show all headers | View raw

[...]
> [The history of formal description of programming languages is not
> encouraging. Back in the 1970s the ANSI PL/I standard was defined in
> terms of formal operations on trees, but even so it had bugs/errors.
> Also, per your comment on addition, in a lot of cases the practical
> definition turns into whatever the underlying machine's instruction
> does. Until recently the C standard was deliberately unclear about
> whether arithmetic was ones- or twos-complement. -John]

The C standard still doesn't mandate two's-complement.

The 1990 edition of the C standard was vague about integer
representations, though it did impose some requirements.  It required
a "pure binary" representation, and it required, for example,
a value that is within the range of both int and unsigned int to
have the same representation in both.  That excluded some of the
more exotic possibilites.  (It mentioned 2's complement a few times,
but only as an example.)

The 1999 standard explicitly required one of three representations:
sign and magnitude, two's complement, or one's complement.  It also
explicitly permitted padding bits (bits that don't contribute
to the value, and that can either be ignored or create trap
representations).

The 2011 standard corrected the spelling of "ones' complement",
but otherwise didn't change anything significant.  (The 2017 edition
was a minor update.)

The upcoming 2023 standard mandates two's complement.  And it
requires INT_MIN to be exactly -INT_MAX-1; previously INT_MIN could
be equal to -INT_MAX, and -INT_MAX-1 could be a trap representation.
It still permits padding bits and trap representations.

Note that twos's complement *representation* doesn't imply two's
complement *behavior* on overflow.  Signed integer overflow still
has undefined behavior.

--
Keith Thompson (The_Other_Keith) Keith.S.Thompson+u@gmail.com
Working, but not speaking, for XCOM Labs
void Void(void) { Void(); } /* The recursive call of the void */

Back to comp.compilers | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

Re: Are there different programming languages that are compiled to the same intermediate language? Martin Ward <mwardgkc@gmail.com> - 2023-01-31 14:04 +0000
  Re: Are there different programming languages that are compiled to the same intermediate language? arnold@freefriends.org (Aharon Robbins) - 2023-02-01 08:07 +0000
    Re: Software proofs, was Are there different programming languages that are compiled to the same intermediate language? Spiros Bousbouras <spibou@gmail.com> - 2023-02-05 15:14 +0000
      Re: C arithmetic, was Software proofs, was Are there different Keith Thompson <Keith.S.Thompson+u@gmail.com> - 2023-02-05 16:14 -0800
        Re: C arithmetic, was Software proofs, was Are there different gah4 <gah4@u.washington.edu> - 2023-02-06 13:26 -0800
          Re: C arithmetic, was Software proofs, was Are there different Hans-Peter Diettrich <DrDiettrich1@netscape.net> - 2023-02-07 14:31 +0100
            Re: C arithmetic, was Software proofs, was Are there different gah4 <gah4@u.washington.edu> - 2023-02-08 01:10 -0800
              Re: C arithmetic, was Software proofs, was Are there different gah4 <gah4@u.washington.edu> - 2023-02-09 00:26 -0800
                Re: C arithmetic, was Software proofs, was Are there different gah4 <gah4@u.washington.edu> - 2023-02-11 00:01 -0800
                Re: C arithmetic, was Software proofs, was Are there different drb@ihatespam.msu.edu (Dennis Boone) - 2023-02-12 04:37 +0000
            Re: C arithmetic, was Software proofs, was Are there different anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2023-02-08 10:19 +0000
              Re: C arithmetic, was Software proofs, was Are there different Hans-Peter Diettrich <DrDiettrich1@netscape.net> - 2023-02-10 19:11 +0100
                Re: C arithmetic, was Software proofs, was Are there different gah4 <gah4@u.washington.edu> - 2023-02-10 23:47 -0800
                Re: C arithmetic, was Software proofs, was Are there different anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2023-02-11 22:34 +0000
                Re: old floating point, C arithmetic, was Software proofs, was Are there different gah4 <gah4@u.washington.edu> - 2023-02-11 22:48 -0800
            Re: C arithmetic, was Software proofs, was Are there different Hans-Peter Diettrich <DrDiettrich1@netscape.net> - 2023-02-08 15:24 +0100
              Re: C arithmetic, was Software proofs, was Are there different gah4 <gah4@u.washington.edu> - 2023-02-09 00:37 -0800

csiph-web