bug-gnulib
[Top][All Lists]
Advanced

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

Re: gettext vs. gnulib m4 macros


From: Karl Berry
Subject: Re: gettext vs. gnulib m4 macros
Date: Tue, 24 Feb 2015 22:33:37 GMT

    I agree with adding them to srclist.txt

Ok, thanks.  Sounds good, I will.  So now my question is, are all of the
*.m4 files in gettext-runtime/m4 really maintained in gettext?  I.e.,
they should all be added to srclist.txt?

$ ls gettext-0.19.4/gettext-runtime/m4/
ChangeLog         gettext.m4  intlmacosx.m4    nls.m4           uintmax_t.m4
Makefile.am       glibc2.m4   intmax.m4        po.m4            visibility.m4
Makefile.in       glibc21.m4  inttypes-pri.m4  printf-posix.m4  wchar_t.m4
README            iconv.m4    inttypes_h.m4    progtest.m4      wint_t.m4
codeset.m4        intdiv0.m4  lcmessage.m4     size_max.m4      xsize.m4
extern-inline.m4  intl.m4     lock.m4          stdint_h.m4
fcntl-o.m4        intldir.m4  longlong.m4      threadlib.m4

It seems bizarre (especially given the existence of
gettext-runtime/gnulib-m4), though I can believe that Bruno wrote them
all originally and things just went along from there.  (Not arguing for
that the situation should change, just want to be sure I understand.)

The only one with nontrivial differences between gettext and gnulib is
intl.m4 (diff below).  Please confirm that the gettext version should go
into gnulib ...

Thanks,
Karl

--- gettext/m4/intl.m4  2014-12-09 20:05:46.000000000 -0800
+++ gnulib/m4/intl.m4   2015-01-01 00:22:56.924372633 -0800
@@ -1,3 +1,3 @@
-# intl.m4 serial 28 (gettext-0.19)
-dnl Copyright (C) 1995-2014 Free Software Foundation, Inc.
+# intl.m4 serial 27 (gettext-0.18.3)
+dnl Copyright (C) 1995-2015 Free Software Foundation, Inc.
 dnl This file is free software; the Free Software Foundation
@@ -44,4 +44,2 @@
   AC_REQUIRE([gt_INTL_MACOSX])dnl
-  AC_REQUIRE([gl_EXTERN_INLINE])dnl
-  AC_REQUIRE([gt_GL_ATTRIBUTE])dnl
 
@@ -244,3 +242,4 @@
   dnl because plural.y uses bison specific features. It requires at least
-  dnl bison-2.7 for %define api.pure.
+  dnl bison-1.26 because earlier versions generate a plural.c that doesn't
+  dnl compile.
   dnl bison is only needed for the maintainer (who touches plural.y). But in
@@ -261,3 +260,3 @@
       '') ac_prog_version="v. ?.??, bad"; ac_verc_fail=yes;;
-      2.[7-9]* | [3-9].*)
+      1.2[6-9]* | 1.[3-9][0-9]* | [2-9].*)
 changequote([,])dnl
@@ -272,27 +271 @@
 ])
-
-dnl Copies _GL_UNUSED and _GL_ATTRIBUTE_PURE definitions from
-dnl gnulib-common.m4 as a fallback, if the project isn't using Gnulib.
-AC_DEFUN([gt_GL_ATTRIBUTE], [
-  m4_ifndef([gl_[]COMMON],
-    AH_VERBATIM([gt_gl_attribute],
-[/* Define as a marker that can be attached to declarations that might not
-    be used.  This helps to reduce warnings, such as from
-    GCC -Wunused-parameter.  */
-#ifndef _GL_UNUSED
-# if __GNUC__ >= 3 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 7)
-#  define _GL_UNUSED __attribute__ ((__unused__))
-# else
-#  define _GL_UNUSED
-# endif
-#endif
-
-/* The __pure__ attribute was added in gcc 2.96.  */
-#ifndef _GL_ATTRIBUTE_PURE
-# if __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 96)
-#  define _GL_ATTRIBUTE_PURE __attribute__ ((__pure__))
-# else
-#  define _GL_ATTRIBUTE_PURE /* empty */
-# endif
-#endif
-]))])



reply via email to

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