To comment on the following update, log in, then open the issue:
http://www.openoffice.org/issues/show_bug.cgi?id=112803
Issue #|112803
Summary|config.guess mis-guesses x86_64-apple-darwin10.4.0 (co
|nfigure fails)
Component|tools
Version|DEV300m83
Platform|Macintosh
URL|
OS/Version|Mac OS X
Status|NEW
Status whiteboard|
Keywords|
Resolution|
Issue type|DEFECT
Priority|P3
Subcomponent|configure
Assigned to|sb
Reported by|sb
------- Additional comments from s...@openoffice.org Tue Jun 29 21:26:11 +0000
2010 -------
After upgrading to Mac OS X 10.6.4 and Xcode 3.2.3 (don't know which is the
culprit, but I would
assume Xcode), a simple configure with no additional switches no longer works:
[...]
configure: creating ./config.status
config.status: creating set_soenv
config.status: creating Makefile
Setting up the environment for building OpenOffice.org
Use of uninitialized value $OUTPATH in concatenation (.) or string at
./set_soenv line 774.
Setting platform independent values... Setting Mac OS X/Darwin specific
values... done
Use of uninitialized value $outfile in concatenation (.) or string at
./set_soenv line 920.
Setting the environment variables... done
Setting the aliases... done
Use of uninitialized value $outfile in concatenation (.) or string at
./set_soenv line 1514.
Use of uninitialized value $outfile in concatenation (.) or string at
./set_soenv line 1514.
Cannot open : No such file or directory
The problem appears to be as follows: configure calls set_soenv, which in turn
calls config.guess, which
guesses CC_FOR_BUILD to be in-PATH gcc = /usr/bin/gcc-4.2, and then calls
$CC_FOR_BUILD -E on
#ifdef __LP64__
IS_64BIT_ARCH
#endif
to determine whether the architecture is actually x86_64 rather than i386 as
reported by uname -p.
Now, on my updated machine, __LP64__ is now defined for plain /usr/bin/gcc-4.2
(while it still is not
defined for /usr/bin/gcc-4.0), so config.guess reports
x86_64-apple-darwin10.4.0 (instead of i386-
apple-darwin10.4.0), which in turn setso_env does not know how to handle.
A plain configure on Mac OS X actually sets CC=/usr/bin/gcc-4.0 (and puts that
into
MacOSXX86Env.Set.sh), but does not export it when calling
set_soenv/config.guess. So, a quick fix is
---8<---
diff --git a/configure.in b/configure.in
--- a/configure.in
+++ b/configure.in
@@ -1452,7 +1452,7 @@
fi
if test "$_os" = "Darwin" -a "$GCCVER" -ge "040100" ; then
if test -z "$save_CC" -a -x "$GCC_HOME/bin/gcc-4.0" ; then
- CC=$GCC_HOME/bin/gcc-4.0
+ export CC=$GCC_HOME/bin/gcc-4.0
GCCVER2=`"$CC" -dumpversion | $AWK -F. '{ print \$1*10000+\$2*100+\$3
}'`
if test "$GCCVER2" -ge "040000" -a "$GCCVER2" -lt "040100" ; then
GCCVER=$GCCVER2
---8<---
but I wonder whether it would not be better to generally export CC into
set_soenv/config.guess, so that
it does not potentially mis-guess its CC_FOR_BUILD on any platform.
---------------------------------------------------------------------
Please do not reply to this automatically generated notification from
Issue Tracker. Please log onto the website and enter your comments.
http://qa.openoffice.org/issue_handling/project_issues.html#notification
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@tools.openoffice.org
For additional commands, e-mail: issues-h...@tools.openoffice.org
---------------------------------------------------------------------
To unsubscribe, e-mail: allbugs-unsubscr...@openoffice.org
For additional commands, e-mail: allbugs-h...@openoffice.org
