public inbox for gcc-help@gcc.gnu.org
 help / color / mirror / Atom feed
From: ludo@gnu.org (Ludovic Courtès)
To: gcc-help@gcc.gnu.org
Subject: Re: Passing `-B', `-isystem' and `-Wl,dynamic-linker' flags
Date: Mon, 14 Apr 2008 18:03:00 -0000	[thread overview]
Message-ID: <87d4os63vh.fsf@gnu.org> (raw)
In-Reply-To: <8763ultvy5.fsf@gnu.org>

Hi,

ludo@gnu.org (Ludovic Courtès) writes:

> Thus, when building GCC, `-B' and `-isystem' flags must be passed so
> that it finds the libc headers, as well as a `-Wl,-dynamic-linker' flag
> to pass the path to the dynamic linker.  However, none of the
> environment variables supposedly honored by `configure' seem to suffice
> to propagate these flags through all the GCC build tree and all the
> build stages (currently, NixOS GCC builds circumvent this problems by
> hacking around Makefiles).
>
> What's the recommended way to achieve this?

FWIW, I solved it by passing various environment variables, either at
`configure' time, at `make' time, or both.  Namely:

  1. I passed `CFLAGS_FOR_BUILD' and `LDFLAGS_FOR_BUILD' with the
     relevant `-B', `-isystem' and `-Wl,-dynamic-linker' flags to the
     top-level `configure'.

  2. I then ran `make' with

       make BOOT_LDFLAGS="-B/path/to/libc/lib \
                          -Wl,-dynamic-linker=/path/to/libc/lib/ld-linux.so.2" \
            C_INCLUDE_PATH="$PWD/prev-gcc/include:$PWD/prev-gcc/include-fixed:/path/to/libc/include" \
            CPPFLAGS="-DSSIZE_MAX=LONG_MAX"

     Not all parts of the build require these flags, but it doesn't hurt
     to pass them from the beginning.  The `CPPFLAGS' setting works
     around the fact that GCC ended up using its own <limits.h> (under
     `prev-gcc/include-fixed'), which doesn't define `SSIZE_MAX'.

This is somewhat kludgey and inconvenient, maybe redundant, but I
couldn't find any better way.

One complication is that some environment variables are captured and
substituted by `config.status', while others are only read by makefiles
(and sometimes then captured by a sub-`configure'), making step (2)
above unavoidable.

Thanks,
Ludovic.

      reply	other threads:[~2008-04-14 12:03 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-04-14  8:25 Ludovic Courtès
2008-04-14 18:03 ` Ludovic Courtès [this message]

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=87d4os63vh.fsf@gnu.org \
    --to=ludo@gnu.org \
    --cc=gcc-help@gcc.gnu.org \
    /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).