bug-gnulib
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

an autoconf expert challenge


From: Bruno Haible
Subject: an autoconf expert challenge
Date: Fri, 20 Oct 2006 15:52:56 +0200
User-agent: KMail/1.9.1

Since gl_FUNC_ALLOCA may invoke AC_LIBOBJ, it must not be AC_REQUIREd -
otherwise it won't work if you use it in two different gnulib libraries from
the same configure.ac file.

Does anyone know how to write an automatic check for this rule: that
autoconf macros that directly or indirectly invoke AC_LIBOBJ are not
AC_REQUIREd?


2006-10-19  Bruno Haible  <address@hidden>

        * m4/allocsa.m4 (gl_ALLOCSA): Invoke gl_FUNC_ALLOCA, don't AC_REQUIRE
        it.

*** gnulib-20061019/m4/allocsa.m4       2006-10-13 00:10:35.000000000 +0200
--- gnulib-20061019-modified/m4/allocsa.m4      2006-10-20 02:12:31.000000000 
+0200
***************
*** 1,4 ****
! # allocsa.m4 serial 4
  dnl Copyright (C) 2003-2004, 2006 Free Software Foundation, Inc.
  dnl This file is free software; the Free Software Foundation
  dnl gives unlimited permission to copy and/or distribute it,
--- 1,4 ----
! # allocsa.m4 serial 5
  dnl Copyright (C) 2003-2004, 2006 Free Software Foundation, Inc.
  dnl This file is free software; the Free Software Foundation
  dnl gives unlimited permission to copy and/or distribute it,
***************
*** 8,14 ****
  [
    dnl Use the autoconf tests for alloca(), but not the AC_SUBSTed variables
    dnl @ALLOCA@ and @address@hidden
!   AC_REQUIRE([gl_FUNC_ALLOCA])
    AC_REQUIRE([gl_EEMALLOC])
    AC_REQUIRE([AC_TYPE_LONG_LONG_INT])
    AC_REQUIRE([gt_TYPE_LONGDOUBLE])
--- 8,14 ----
  [
    dnl Use the autoconf tests for alloca(), but not the AC_SUBSTed variables
    dnl @ALLOCA@ and @address@hidden
!   gl_FUNC_ALLOCA
    AC_REQUIRE([gl_EEMALLOC])
    AC_REQUIRE([AC_TYPE_LONG_LONG_INT])
    AC_REQUIRE([gt_TYPE_LONGDOUBLE])




reply via email to

[Prev in Thread] Current Thread [Next in Thread]