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: GCC vs Clang
Date: Fri, 12 Aug 2016 11:29:38 -0700
Organization: None to speak of
Lines: 36
Message-ID: <lnpopdq1rh.fsf@kst-u.example.com>
References: <dvsj5aFakf4U4@mid.individual.net> <yZamz.1131820$GD.1074203@fx33.am4> <add98d01-a1f0-4576-b3b1-8b830e5d7026@googlegroups.com> <noev9h$gg9$1@dont-email.me> <lnoa50twco.fsf@kst-u.example.com> <nofm46$1cb$1@dont-email.me> <lny444sbl5.fsf@kst-u.example.com> <nofv94$3hv$1@dont-email.me> <4p0oqbh4ail533kg8acdg2u1p3n2d47n6f@4ax.com> <nohgst$bh3$1@dont-email.me> <u9ppqb5m4fs5596s3vffoeihor08qcl192@4ax.com> <7ed4b3db-62d1-4613-ab0a-84c7f921334c@googlegroups.com> <bd0rqblo10ujb41utg6i9srb8j4sgfdibg@4ax.com> <nok8f1$quv$1@dont-email.me> <a2c288b5-6b9a-410d-acef-637f6ea2ec6d@googlegroups.com> <nokheo$oaa$1@dont-email.me> <2c83dfac-49cf-4c4b-9782-fe4e3c0be7ff@googlegroups.com> <lntweqoth2.fsf@kst-u.example.com> <395b1cf7-3bd2-4464-937c-708b410710dd@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Injection-Info: mx02.eternal-september.org; posting-host="945944de09706c9b4e29b53c9d2efdc2"; logging-data="15949"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX18Af7bR0PI/TVDRoIO/dlPz"
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.1 (gnu/linux)
Cancel-Lock: sha1:ZLup/erUBZ+pyBHUfJmyj4Gmc3Q= sha1:UF2Pxk0OTlYnmi1GgrvMNbSXY2g=
Xref: csiph.com comp.lang.c:92203

supercat@casperkitty.com writes:
> On Friday, August 12, 2016 at 10:33:24 AM UTC-5, Keith Thompson wrote:
>> That's possible on Linux too (well, at least for 32 and 64 bits).
>> On my system, gcc has "-m32", "-mx32", and "-m64" options.  16-bit int
>> is not supported as far as I know.
>
> Does it use macros to rename standard-library functions like "sscanf"
> to indicate whether they expect "%d" to overwrite 4 or 8 bytes, does
> it bundle standard-library functions into executables, or does it have
> some other way of handling such issues?

It uses separate shared libraries.

    $ ldd hello-m*
    hello-m32:
            linux-gate.so.1 =>  (0xf779e000)
            libc.so.6 => /lib/i386-linux-gnu/libc.so.6 (0xf75be000)
            /lib/ld-linux.so.2 (0xf77a1000)
    hello-m64:
            linux-vdso.so.1 =>  (0x00007ffea198d000)
            libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f0b2c8ad000)
            /lib64/ld-linux-x86-64.so.2 (0x00007f0b2cca2000)
    hello-mx32:
            linux-vdso.so.1 =>  (0xfff9a000)
            libc.so.6 => /libx32/libc.so.6 (0xf7406000)
            /libx32/ld-linux-x32.so.2 (0xf77d0000)
    $

Most of the header files are the same (I think the compiler predefines
some macros that some of the gcc-specific headers test).

-- 
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"