Path: csiph.com!eternal-september.org!feeder.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: Keith Thompson <kst-u@mib.org>
Newsgroups: comp.lang.c
Subject: Re: va_start without the first argument?
Date: Tue, 14 Mar 2017 11:34:42 -0700
Organization: None to speak of
Lines: 42
Message-ID: <ln4lyvk8ql.fsf@kst-u.example.com>
References: <b0f99f2b-09a3-4510-89dd-c3511fba3321@googlegroups.com> <oa4csk$jd4$1@dont-email.me> <oa4f3e$r2v$1@dont-email.me> <7fda2e02-9a6a-411f-9d91-1ad9030ec778@googlegroups.com> <oa4n2f$nq6$1@dont-email.me> <a6aade9b-b090-4e38-a131-11f15f25261c@googlegroups.com> <lnd1dlf364.fsf@kst-u.example.com> <zBBxA.61776$Bp6.33909@fx36.iad> <oa71j7$rrb$1@dont-email.me> <7dc37a16-7135-4e50-accf-591c923908c4@googlegroups.com> <oa8p90$n6q$1@dont-email.me> <lnh92vkf22.fsf@kst-u.example.com> <UcVxA.62280$Bp6.51791@fx36.iad>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Injection-Info: mx02.eternal-september.org; posting-host="68145e82dff5cec9d4077a7833a68c0b"; logging-data="29072"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX19Tz73EF3RxGMO0DF+jghSJ"
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.1 (gnu/linux)
Cancel-Lock: sha1:jxLONEn04QebKv+g5C7IZl525Xs= sha1:2fmjrJg7FPfpQuuG8SMMeoSRwgY=
Xref: csiph.com comp.lang.c:106067

scott@slp53.sl.home (Scott Lurndal) writes:
> Keith Thompson <kst-u@mib.org> writes:
>>David Brown <david.brown@hesbynett.no> writes:
>>[...]
>>> #ifdef __GNUC__
>>> #define IS_COMPILE_TIME_CONSTANT(x) __builtin_constant_p(x)
>>
>>I'd parenthesize this:
>>
>>#define IS_COMPILE_TIME_CONSTANT(x) (__builtin_constant_p(x))
>>
>>I'm not sure whether it can make a difference, but it's easier to add
>>the parentheses than to prove that they're not needed.
>
> I think if you are adding additional parenthesis, I'd be tempted to write it:
>
>    #define IS_COMPILE_TIME_CONSTANT(x) __builtin_constant_p((x))

Why?  In both David's version and mine, the reference to x in
the macro definition is already parenthesized.  The fact that the
parentheses are for a function call (or macro invocation?) rather
than grouping doesn't matter as far as operator precedence is
concerned.  Similarly, if a macro definition uses an argument as
an array index, it's already surrounded by [] so it doesn't need
additional ().

In general, for a macro whose arguments are expressions and
that expands to an expression, each reference to each argument
should be parenthesized, and the entire definition should also
be parenthesized.  Thus:

    #define ADD(x, y) ((x) + (y))

But:

    #define INDEX(a, i) ((a)[i]) // no need for additional parentheses on i

-- 
Keith Thompson (The_Other_Keith) kst-u@mib.org  <http://www.ghoti.net/~kst>
Working, but not speaking, for JetHead Development, Inc.
"We must do something.  This is something.  Therefore, we must do this."
    -- Antony Jay and Jonathan Lynn, "Yes Minister"