Buildroot Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Korsgaard <jacmet@uclibc.org>
To: buildroot@busybox.net
Subject: [Buildroot] coreutils build failing
Date: Thu, 10 Sep 2009 09:28:40 +0200	[thread overview]
Message-ID: <87ocpjw95j.fsf@macbook.be.48ers.dk> (raw)
In-Reply-To: <4AA87826.5020506@rodgers.sdcoxmail.com> (Stephen Rodgers's message of "Wed\, 09 Sep 2009 20\:53\:10 -0700")

>>>>> "Stephen" == Stephen Rodgers <hwstar@rodgers.sdcoxmail.com> writes:

Hi,

 Stephen> As you requested, I tarred up the config.log from the failed
 Stephen> coreutils-7.4 build. You should be able to wget
 Stephen> config.log.tar.gz using the following url:

 Stephen> http://test.qrvc.com/config.log.tar.gz

Thanks. The bug is in the coreutils configure script:

grep -i -A1 -B1 '^[^|].*dest.*exists' config.log

configure:28670: result: yes
configure:28672: checking whether rename is broken when the destination exists
configure:28684: result: no
--
gl_cv_func_re_compile_pattern_working=no
gl_cv_func_rename_dest_exists_bug=no
gl_cv_func_rename_trailing_dest_slash_bug=yes
--
#define RENAME_TRAILING_SLASH_BUG 1
#define RENAME_DEST_EXISTS_BUG 1
#define rename rpl_rename_dest_slash

It correctly detects that you don't need the windows-only rename
workaround, but still it ends up defining RENAME_DEST_EXISTS_BUG.

The logic (m4/rename.m4) goes:

  if test $gl_cv_func_rename_trailing_slash_bug = yes ||
     test $gl_cv_func_rename_dest_exists_bug = yes; then
    AC_LIBOBJ([rename])
    AC_DEFINE([rename], [rpl_rename],
      [Define to rpl_rename if the replacement function should be used.])
    if test $gl_cv_func_rename_trailing_slash_bug; then
      AC_DEFINE([RENAME_TRAILING_SLASH_BUG], [1],
        [Define if rename does not work for source file names with a trailing
         slash, like the one from SunOS 4.1.1_U1.])
    fi
    if test $gl_cv_func_rename_dest_exists_bug; then
      AC_DEFINE([RENAME_DEST_EXISTS_BUG], [1],
        [Define if rename does not work when the destination file exists,
         as on Windows.])
    fi
    gl_PREREQ_RENAME
  fi

Notice that the 2nd test doesn't do 'test $var = yes', so if any of the
2 variables are true both defines will get set. Just above we see:

 dnl When crosscompiling, assume rename is broken.
 gl_cv_func_rename_trailing_slash_bug=yes)

So the quick fix is to tell configure that we don't need the trailing
slash bugfix (which also just adds to the size and isn't needed on
uclibc).

I've done that in git (dfca1b2e0e) - Please give it a try.

I've just checked, and all of this is removed in the recently released
coreutils 7.5, so we might consider upgrading to that.

-- 
Bye, Peter Korsgaard

  reply	other threads:[~2009-09-10  7:28 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-09-06 17:05 [Buildroot] coreutils build failing Stephen Rodgers
2009-09-07  5:35 ` Peter Korsgaard
2009-09-07 16:08   ` Stephen Rodgers
2009-09-07 18:01     ` Peter Korsgaard
2009-09-07 18:36       ` Stephen Rodgers
2009-09-07 21:25         ` Peter Korsgaard
     [not found]           ` <4AA58BB1.50906@rodgers.sdcoxmail.com>
2009-09-09 14:29             ` Peter Korsgaard
2009-09-10  3:53               ` Stephen Rodgers
2009-09-10  7:28                 ` Peter Korsgaard [this message]
2009-09-11  9:11                   ` Stephen Rodgers
2009-09-11 11:21                     ` Peter Korsgaard
2009-09-11 14:16                       ` Stephen Rodgers
2009-09-11 18:44                         ` Peter Korsgaard
2009-09-12  2:16                           ` Stephen Rodgers
  -- strict thread matches above, loose matches on Subject: below --
2009-09-07 22:49 Stephen Rodgers

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=87ocpjw95j.fsf@macbook.be.48ers.dk \
    --to=jacmet@uclibc.org \
    --cc=buildroot@busybox.net \
    /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