public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
From: Ralf Wildenhues <Ralf.Wildenhues@gmx.de>
To: binutils@sourceware.org, gdb-patches@sourceware.org
Cc: Paolo Bonzini <bonzini@gnu.org>
Subject: Re: toplevel configure hits sed program length limit on HP-UX
Date: Tue, 08 Sep 2009 17:57:00 -0000	[thread overview]
Message-ID: <20090908175659.GB24114@gmx.de> (raw)
In-Reply-To: <4AA41B9B.6040304@gnu.org>

Hello binutils, gdb-patches,

I've pushed this patch over to src as well, pasted below for convenience
<http://gcc.gnu.org/ml/gcc-patches/2009-09/msg00427.html>

* Paolo Bonzini wrote on Sun, Sep 06, 2009 at 10:29:15PM CEST:
> On 09/06/2009 08:01 PM, Ralf Wildenhues wrote:
> >OK to apply?  Tested by ensuring old and new toplevel Makefile are
> >identical.
> 
> Ok after the slush finishes.

Thanks, done.

Cheers,
Ralf

Fix toplevel config.status for HP-UX sed.

ChangeLog:
2009-09-08  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>

	* configure.ac: Do not use $extrasub for replacing @if/@endif
	parts in Makefile; instead, use additional arguments to
	AC_CONFIG_COMMANDS to do the replacement manually, with several
	sed invocations, to avoid HP-UX sed command limits.
	* configure: Regenerate.

diff --git a/configure.ac b/configure.ac
index 592a8e3..a85c3bd 100644
--- a/configure.ac
+++ b/configure.ac
@@ -2410,12 +2410,13 @@ INSTALL_GDB_TK=`echo ${GDB_TK} | sed s/-all-/-install-/g`
 # configuration, so that the top-level Makefile reconfigures them,
 # like we used to do when configure itself was recursive.
 
-# Loop over modules.  $extrasub must be used with care, limiting as
-# much as possible the usage of range addresses.  That's because autoconf
-# splits the sed script to overcome limits in the number of commands,
-# and relying on carefully-timed sed passes may turn out to be very hard
-# to maintain later.  In this particular case, you just have to be careful
-# not to nest @if/@endif pairs, because configure will not warn you at all.
+# Loop over modules.  We used to use the "$extrasub" feature from Autoconf
+# but now we're fixing up the Makefile ourselves with the additional
+# commands passed to AC_CONFIG_FILES.  Use separate variables
+# extrasub-{build,host,target} not because there is any reason to split
+# the substitutions up that way, but only to remain below the limit of
+# 99 commands in a script, for HP-UX sed.
+# Do not nest @if/@endif pairs, because configure will not warn you at all.
 
 AC_ARG_ENABLE([bootstrap],
 [  --enable-bootstrap      enable bootstrapping @<:@yes if native build@:>@],,
@@ -2502,18 +2503,20 @@ fi
 AC_MSG_RESULT($BUILD_CONFIG)
 AC_SUBST(BUILD_CONFIG)
 
+extrasub_build=
 for module in ${build_configdirs} ; do
   if test -z "${no_recursion}" \
      && test -f ${build_subdir}/${module}/Makefile; then
     echo 1>&2 "*** removing ${build_subdir}/${module}/Makefile to force reconfigure"
     rm -f ${build_subdir}/${module}/Makefile
   fi
-  extrasub="$extrasub
+  extrasub_build="$extrasub_build
 /^@if build-$module\$/d
 /^@endif build-$module\$/d
 /^@if build-$module-$bootstrap_suffix\$/d
 /^@endif build-$module-$bootstrap_suffix\$/d"
 done
+extrasub_host=
 for module in ${configdirs} ; do
   if test -z "${no_recursion}"; then
     for file in stage*-${module}/Makefile prev-${module}/Makefile ${module}/Makefile; do
@@ -2523,12 +2526,13 @@ for module in ${configdirs} ; do
       fi
     done
   fi
-  extrasub="$extrasub
+  extrasub_host="$extrasub_host
 /^@if $module\$/d
 /^@endif $module\$/d
 /^@if $module-$bootstrap_suffix\$/d
 /^@endif $module-$bootstrap_suffix\$/d"
 done
+extrasub_target=
 for module in ${target_configdirs} ; do
   if test -z "${no_recursion}" \
      && test -f ${target_subdir}/${module}/Makefile; then
@@ -2542,14 +2546,15 @@ for module in ${target_configdirs} ; do
     *) target_bootstrap_suffix=no-bootstrap ;;
   esac
 
-  extrasub="$extrasub
+  extrasub_target="$extrasub_target
 /^@if target-$module\$/d
 /^@endif target-$module\$/d
 /^@if target-$module-$target_bootstrap_suffix\$/d
 /^@endif target-$module-$target_bootstrap_suffix\$/d"
 done
 
-extrasub="$extrasub
+# Do the final fixup along with target modules.
+extrasub_target="$extrasub_target
 /^@if /,/^@endif /d"
 
 # Create the serialization dependencies.  This uses a temporary file.
@@ -3167,4 +3172,12 @@ case "$target" in
 esac
 AC_SUBST(compare_exclusions)
 
-AC_OUTPUT(Makefile)
+AC_CONFIG_FILES([Makefile],
+  [sed "$extrasub_build" Makefile |
+   sed "$extrasub_host" |
+   sed "$extrasub_target" > mf$$
+   mv -f mf$$ Makefile],
+  [extrasub_build="$extrasub_build"
+   extrasub_host="$extrasub_host"
+   extrasub_target="$extrasub_target"])
+AC_OUTPUT

           reply	other threads:[~2009-09-08 17:57 UTC|newest]

Thread overview: expand[flat|nested]  mbox.gz  Atom feed
 [parent not found: <4AA41B9B.6040304@gnu.org>]

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=20090908175659.GB24114@gmx.de \
    --to=ralf.wildenhues@gmx.de \
    --cc=binutils@sourceware.org \
    --cc=bonzini@gnu.org \
    --cc=gdb-patches@sourceware.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).