public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
From: Cheng Rk <crquan@ymail.com>
To: Dennis <crquan@fedoraproject.org>,
	 "Justin P. Mattock" <justinmattock@gmail.com>
Cc: gcc@gcc.gnu.org
Subject: Re: on how to compile gcc-4.6 correctly?
Date: Mon, 06 Sep 2010 09:54:00 -0000	[thread overview]
Message-ID: <501025.9515.qm@web77013.mail.sg1.yahoo.com> (raw)
In-Reply-To: <4C84B1E3.7020502@gmail.com>

--- On Mon, 6/9/10, Justin P. Mattock <justinmattock@gmail.com> wrote:
> From: Justin P. Mattock <justinmattock@gmail.com>
> Subject: Re: on how to compile gcc-4.6 correctly?
> To: "Dennis" <crquan@fedoraproject.org>
> Cc: gcc@gcc.gnu.org, crquan@ymail.com
> Date: Monday, 6 September, 2010, 9:18 AM
> On 09/05/2010 08:17 PM, Dennis
[...]
> hm... an obvious question would be if you have your
> toolchain correct?

Yes, I think the binutils/gmp/mpfr/mpc all dependencies are correct, otherwise why I can compile all gcc-4.5.x version correctly?
On the same system, gcc-4.5.x (4.5.0 and 4.5.1 and 4.5-20100902/... multiple snapshots) all can compile correctly, I install them into separate directories, and have verified those compilers can compile many source packages correctly;

> (but you probably do..(I used this to build 
> 4.6.0..:http://cross-lfs.org/view/svn/x86_64-64/))

but this still didn't explain what's the "ggc_alloc_cleared_lang_type" macro? all through gcc source it's only used by gcc/c-decl.c file, but it has no definition under gcc source header files, nor under /usr/include;

> 
> in your case I see something with lang... maybe your CFLAGS
> are set 
> wrong to the wrong machine?! over here my CFLAGS look like
> this:
> CFLAGS="-m64 -mtune=core2 -march=core2 -O2 -pipe
> -fomit-frame-pointer 
> -fstack-protector -w" CXXFLAGS="${CFLAGS}"
> MAKEOPTS="{-j3}"

I didn't use any special CFLAGS, when I try to compile gcc-4.5.x or this gcc-4.6 snapshot, I just use "../gcc-4.6-20100905/configure --with-arch-64=athlon64 --with-arch-32=i386", I tried to get a multilib compiler that could generate default athlon64 optimized code for 64bit and i386 code for 32bit;

The same "--with-arch-64=athlon64 --with-arch-32=i386" works well during compiling gcc-4.5.x but why not work on gcc-4.6 ???

> 
> keep in mind this is for a intel iMac... your system could
> be diff..
> 
> Justin P. Mattock
> 

Thanks a lot~


  reply	other threads:[~2010-09-06  9:54 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-09-06  3:17 Dennis
2010-09-06  9:18 ` Justin P. Mattock
2010-09-06  9:54   ` Cheng Rk [this message]
2010-09-06 10:32     ` Justin Mattock
2010-09-06 14:52 ` Kai Ruottu
2010-09-13 14:19   ` Dennis, CHENG Renquan
2010-09-13 20:04     ` Kai Ruottu
2010-09-13 20:15     ` Laurynas Biveinis
2010-09-13 20:31       ` Dennis, CHENG Renquan
2010-09-13 21:37         ` Laurynas Biveinis
2010-09-13 21:42           ` Dennis, CHENG Renquan
2010-09-14 14:01             ` Dennis, CHENG Renquan
2010-09-14 14:23               ` Kai Ruottu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=501025.9515.qm@web77013.mail.sg1.yahoo.com \
    --to=crquan@ymail.com \
    --cc=crquan@fedoraproject.org \
    --cc=gcc@gcc.gnu.org \
    --cc=justinmattock@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).