1993-05-23 20:50:09 +00:00
|
|
|
dnl This is an autoconf script.
|
1993-05-23 21:04:16 +00:00
|
|
|
dnl To rebuild the `configure' script from this, execute the command
|
|
|
|
dnl autoconf
|
|
|
|
dnl in the directory containing this script. You must have autoconf
|
|
|
|
dnl version 1.4 or later.
|
|
|
|
dnl
|
|
|
|
dnl The following text appears in the resulting `configure' script,
|
|
|
|
dnl explaining how to rebuild it.
|
1993-05-23 20:39:16 +00:00
|
|
|
[#!/bin/sh
|
1992-12-06 22:12:37 +00:00
|
|
|
#### Configuration script for GNU Emacs
|
|
|
|
#### Copyright (C) 1992 Free Software Foundation, Inc.
|
|
|
|
|
1993-05-23 21:04:16 +00:00
|
|
|
### Don't edit this script!
|
|
|
|
### This script was automatically generated by the `autoconf' program
|
|
|
|
### from the file `./configure.in'.
|
|
|
|
### To rebuild it, execute the command
|
|
|
|
### autoconf
|
|
|
|
### in the this directory. You must have autoconf version 1.4 or later.
|
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
### This file is part of GNU Emacs.
|
|
|
|
|
|
|
|
### GNU Emacs is free software; you can redistribute it and/or modify
|
|
|
|
### it under the terms of the GNU General Public License as published by
|
|
|
|
### the Free Software Foundation; either version 1, or (at your option)
|
|
|
|
### any later version.
|
|
|
|
|
|
|
|
### GNU Emacs is distributed in the hope that it will be useful,
|
|
|
|
### but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
### MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
### GNU General Public License for more details.
|
|
|
|
|
|
|
|
### You should have received a copy of the GNU General Public License
|
|
|
|
### along with GNU Emacs; see the file COPYING. If not, write to
|
|
|
|
### the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
|
|
|
|
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
|
|
|
|
### Since Emacs has configuration requirements that autoconf can't
|
|
|
|
### meet, this file is an unholy marriage of custom-baked
|
1993-05-24 07:02:01 +00:00
|
|
|
### configuration code and autoconf macros.
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
###
|
|
|
|
### We use the m4 quoting characters [ ] (as established by the
|
|
|
|
### autoconf system) to include large sections of raw sewage - Oops, I
|
|
|
|
### mean, shell code - in the final configuration script.
|
1992-12-06 22:12:37 +00:00
|
|
|
###
|
|
|
|
### Usage: configure config_name
|
|
|
|
###
|
|
|
|
### If configure succeeds, it leaves its status in config.status.
|
|
|
|
### If configure fails after disturbing the status quo,
|
|
|
|
### config.status is removed.
|
|
|
|
|
|
|
|
|
1993-03-11 06:58:58 +00:00
|
|
|
### Remove any more than one leading "." element from the path name.
|
|
|
|
### If we don't remove them, then another "./" will be prepended to
|
|
|
|
### the file name each time we use config.status, and the program name
|
|
|
|
### will get larger and larger. This wouldn't be a problem, except
|
|
|
|
### that since progname gets recorded in all the Makefiles this script
|
|
|
|
### produces, move-if-change thinks they're different when they're
|
|
|
|
### not.
|
1992-12-06 22:12:37 +00:00
|
|
|
###
|
|
|
|
### It would be nice if we could put the ./ in a \( \) group and then
|
|
|
|
### apply the * operator to that, so we remove as many leading ./././'s
|
|
|
|
### as are present, but some seds (like Ultrix's sed) don't allow you to
|
|
|
|
### apply * to a \( \) group. Bleah.
|
1993-03-11 06:58:58 +00:00
|
|
|
progname="`echo $0 | sed 's:^\./\./:\./:'`"
|
1992-04-01 10:45:51 +00:00
|
|
|
|
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
#### Usage messages.
|
|
|
|
|
1993-05-23 04:42:38 +00:00
|
|
|
short_usage="Usage: ${progname} CONFIGURATION [-OPTION[=VALUE] ...]
|
1992-11-20 17:08:32 +00:00
|
|
|
|
1992-04-01 10:45:51 +00:00
|
|
|
Set compilation and installation parameters for GNU Emacs, and report.
|
1992-11-20 17:08:32 +00:00
|
|
|
CONFIGURATION specifies the machine and operating system to build for.
|
1993-05-23 04:42:38 +00:00
|
|
|
--with-x Support the X Window System.
|
1993-05-25 16:47:29 +00:00
|
|
|
--with-x=no Don't support X.
|
|
|
|
--x-includes=DIR Search for X header files in DIR.
|
|
|
|
--x-libraries=DIR Search for X libraries in DIR.
|
1993-05-23 04:42:38 +00:00
|
|
|
--with-gcc Use GCC to compile Emacs.
|
|
|
|
--with-gcc=no Don't use GCC to compile Emacs.
|
|
|
|
--run-in-place Use libraries and data files directly out of the
|
|
|
|
source tree.
|
|
|
|
--srcdir=DIR Look for source in DIR.
|
|
|
|
--prefix=DIR Install files below dir.
|
1992-12-06 22:12:37 +00:00
|
|
|
|
1992-04-01 10:45:51 +00:00
|
|
|
If successful, ${progname} leaves its status in config.status. If
|
1992-12-06 22:12:37 +00:00
|
|
|
unsuccessful after disturbing the status quo, it removes config.status."
|
|
|
|
|
|
|
|
|
|
|
|
#### Option processing.
|
1992-04-01 10:45:51 +00:00
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
### Record all the arguments, so we can save them in config.status.
|
1992-10-17 22:07:23 +00:00
|
|
|
arguments="$@"
|
1992-05-10 18:15:10 +00:00
|
|
|
|
1993-05-03 02:07:41 +00:00
|
|
|
### These values are used to comment and uncomment different values
|
|
|
|
### for the path variables in the Makefile, to choose the installed
|
|
|
|
### configuration or the run-in-place configuration.
|
|
|
|
rip_paths='#disabled# '
|
|
|
|
inst_paths=''
|
|
|
|
|
1993-05-23 04:42:38 +00:00
|
|
|
### Establish some default values.
|
|
|
|
prefix='/usr/local'
|
1993-05-24 06:08:13 +00:00
|
|
|
exec_prefix='${prefix}'
|
1993-05-23 04:42:38 +00:00
|
|
|
|
1993-05-29 20:07:50 +00:00
|
|
|
### Don't use shift -- that destroys the argument list, which autoconf needs
|
|
|
|
### to produce config.status. It turns out that "set - ${arguments}" doesn't
|
|
|
|
### work portably.
|
|
|
|
index=0
|
|
|
|
while [ $index -lt $# ]; do
|
|
|
|
index=`expr $index + 1`
|
|
|
|
arg=`eval echo '$'$index`
|
1992-04-01 10:45:51 +00:00
|
|
|
case "${arg}" in
|
1992-12-06 22:12:37 +00:00
|
|
|
|
|
|
|
## Anything starting with a hyphen we assume is an option.
|
|
|
|
-* )
|
|
|
|
|
|
|
|
## Separate the switch name from the value it's being given.
|
1992-04-01 10:45:51 +00:00
|
|
|
case "${arg}" in
|
1992-09-14 22:08:02 +00:00
|
|
|
-*=*)
|
1992-10-17 22:07:23 +00:00
|
|
|
opt=`echo ${arg} | sed 's:^-*\([^=]*\)=.*$:\1:'`
|
1992-09-14 22:08:02 +00:00
|
|
|
val=`echo ${arg} | sed 's:^-*[^=]*=\(.*\)$:\1:'`
|
1992-04-01 10:45:51 +00:00
|
|
|
valomitted=no
|
|
|
|
;;
|
1992-09-14 22:08:02 +00:00
|
|
|
-*)
|
1992-12-06 22:12:37 +00:00
|
|
|
## If FOO is a boolean argument, --FOO is equivalent to
|
|
|
|
## --FOO=yes. Otherwise, the value comes from the next
|
|
|
|
## argument - see below.
|
1992-10-17 22:07:23 +00:00
|
|
|
opt=`echo ${arg} | sed 's:^-*\(.*\)$:\1:'`
|
1992-04-01 10:45:51 +00:00
|
|
|
val="yes"
|
|
|
|
valomitted=yes
|
|
|
|
;;
|
|
|
|
esac
|
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
## Change `-' in the option name to `_'.
|
1993-03-25 01:14:34 +00:00
|
|
|
optname="${opt}"
|
1992-10-17 22:07:23 +00:00
|
|
|
opt="`echo ${opt} | tr - _`"
|
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
## Process the option.
|
|
|
|
case "${opt}" in
|
1992-04-01 10:45:51 +00:00
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
## Has the user specified which window systems they want to support?
|
|
|
|
"with_x" | "with_x11" | "with_x10" )
|
|
|
|
## Make sure the value given was either "yes" or "no".
|
1992-04-01 10:45:51 +00:00
|
|
|
case "${val}" in
|
|
|
|
y | ye | yes ) val=yes ;;
|
|
|
|
n | no ) val=no ;;
|
|
|
|
* )
|
1993-03-25 01:14:34 +00:00
|
|
|
(echo "${progname}: the \`--${optname}' option is supposed to have a boolean value.
|
1992-12-06 22:12:37 +00:00
|
|
|
Set it to either \`yes' or \`no'."
|
|
|
|
echo "${short_usage}") >&2
|
1992-04-01 10:45:51 +00:00
|
|
|
exit 1
|
|
|
|
;;
|
|
|
|
esac
|
1992-12-06 22:12:37 +00:00
|
|
|
eval "${opt}=\"${val}\""
|
1992-04-01 10:45:51 +00:00
|
|
|
;;
|
1992-12-06 22:12:37 +00:00
|
|
|
|
1993-02-14 14:25:18 +00:00
|
|
|
## Has the user specified whether or not they want GCC?
|
1993-05-25 16:47:29 +00:00
|
|
|
"with_gcc" | "with_gnu_cc" )
|
1993-02-14 14:25:18 +00:00
|
|
|
## Make sure the value given was either "yes" or "no".
|
|
|
|
case "${val}" in
|
|
|
|
y | ye | yes ) val=yes ;;
|
|
|
|
n | no ) val=no ;;
|
|
|
|
* )
|
1993-03-25 01:14:34 +00:00
|
|
|
(echo "${progname}: the \`--${optname}' option is supposed to have a boolean value.
|
1993-02-14 14:25:18 +00:00
|
|
|
Set it to either \`yes' or \`no'."
|
|
|
|
echo "${short_usage}") >&2
|
|
|
|
exit 1
|
|
|
|
;;
|
|
|
|
esac
|
|
|
|
eval "${opt}=\"${val}\""
|
|
|
|
;;
|
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
## Has the user specified a source directory?
|
|
|
|
"srcdir" )
|
|
|
|
## If the value was omitted, get it from the next argument.
|
1992-04-01 10:45:51 +00:00
|
|
|
if [ "${valomitted}" = "yes" ]; then
|
1992-12-06 22:12:37 +00:00
|
|
|
## Get the next argument from the argument list, if there is one.
|
1993-05-29 20:07:50 +00:00
|
|
|
if [ $index = $# ]; then
|
1993-03-25 01:14:34 +00:00
|
|
|
(echo "${progname}: You must give a value for the \`--${optname}' option, as in
|
|
|
|
\`--${optname}=FOO'."
|
1992-12-06 22:12:37 +00:00
|
|
|
echo "${short_usage}") >&2
|
1992-04-24 08:11:54 +00:00
|
|
|
exit 1
|
|
|
|
fi
|
1993-05-29 20:07:50 +00:00
|
|
|
index=`expr $index + 1`
|
|
|
|
val=`eval echo '$'$index`
|
1992-04-01 10:45:51 +00:00
|
|
|
fi
|
1992-12-06 22:12:37 +00:00
|
|
|
srcdir="${val}"
|
|
|
|
;;
|
|
|
|
|
1993-03-23 05:31:35 +00:00
|
|
|
## Has the user tried to tell us where the X files are?
|
|
|
|
## I think these are dopey, but no less than three alpha
|
|
|
|
## testers, at large sites, have said they have their X files
|
|
|
|
## installed in odd places.
|
1993-03-25 01:14:34 +00:00
|
|
|
"x_includes" )
|
1993-03-23 05:31:35 +00:00
|
|
|
## If the value was omitted, get it from the next argument.
|
|
|
|
if [ "${valomitted}" = "yes" ]; then
|
|
|
|
## Get the next argument from the argument list, if there is one.
|
1993-05-29 20:07:50 +00:00
|
|
|
if [ $index = $# ]; then
|
1993-03-25 01:14:34 +00:00
|
|
|
(echo "${progname}: You must give a value for the \`--${optname}' option, as in
|
|
|
|
\`--${optname}=FOO'."
|
1993-03-23 05:31:35 +00:00
|
|
|
echo "${short_usage}") >&2
|
|
|
|
exit 1
|
|
|
|
fi
|
1993-05-29 20:07:50 +00:00
|
|
|
index=`expr $index + 1`
|
|
|
|
val=`eval echo '$'$index`
|
1993-03-23 05:31:35 +00:00
|
|
|
fi
|
|
|
|
x_includes="${val}"
|
1993-03-23 05:53:36 +00:00
|
|
|
C_SWITCH_X_SITE="-I${x_includes}"
|
1993-03-23 05:31:35 +00:00
|
|
|
;;
|
1993-03-25 01:14:34 +00:00
|
|
|
"x_libraries" )
|
1993-03-23 05:31:35 +00:00
|
|
|
## If the value was omitted, get it from the next argument.
|
|
|
|
if [ "${valomitted}" = "yes" ]; then
|
|
|
|
## Get the next argument from the argument list, if there is one.
|
1993-05-29 20:07:50 +00:00
|
|
|
if [ $index = $# ]; then
|
1993-03-25 01:14:34 +00:00
|
|
|
(echo "${progname}: You must give a value for the \`--${optname}' option, as in
|
|
|
|
\`--${optname}=FOO'."
|
1993-03-23 05:31:35 +00:00
|
|
|
echo "${short_usage}") >&2
|
|
|
|
exit 1
|
|
|
|
fi
|
1993-05-29 20:07:50 +00:00
|
|
|
index=`expr $index + 1`
|
|
|
|
val=`eval echo '$'$index`
|
1993-03-23 05:31:35 +00:00
|
|
|
fi
|
|
|
|
x_libraries="${val}"
|
1993-03-23 05:53:36 +00:00
|
|
|
LD_SWITCH_X_SITE="-L${x_libraries}"
|
1993-03-23 05:31:35 +00:00
|
|
|
;;
|
|
|
|
|
1993-05-03 02:07:41 +00:00
|
|
|
## Should this use the "development configuration"?
|
|
|
|
"run_in_place" )
|
|
|
|
rip_paths=''
|
|
|
|
inst_paths='#disabled# '
|
|
|
|
;;
|
|
|
|
|
1993-06-09 11:59:12 +00:00
|
|
|
## Has the user specified an installation prefix?
|
1993-05-23 04:42:38 +00:00
|
|
|
"prefix" )
|
|
|
|
## If the value was omitted, get it from the next argument.
|
|
|
|
if [ "${valomitted}" = "yes" ]; then
|
|
|
|
## Get the next argument from the argument list, if there is one.
|
1993-05-29 20:07:50 +00:00
|
|
|
if [ $index = $# ]; then
|
1993-05-23 04:42:38 +00:00
|
|
|
(echo "${progname}: You must give a value for the \`--${optname}' option, as in
|
|
|
|
\`--${optname}=FOO'."
|
|
|
|
echo "${short_usage}") >&2
|
|
|
|
exit 1
|
|
|
|
fi
|
1993-05-29 20:07:50 +00:00
|
|
|
index=`expr $index + 1`
|
|
|
|
val=`eval echo '$'$index`
|
1993-05-23 04:42:38 +00:00
|
|
|
fi
|
|
|
|
prefix="${val}"
|
|
|
|
;;
|
|
|
|
|
1993-06-09 11:59:12 +00:00
|
|
|
## Has the user specified an installation prefix?
|
1993-05-24 06:08:13 +00:00
|
|
|
"exec_prefix" )
|
|
|
|
## If the value was omitted, get it from the next argument.
|
|
|
|
if [ "${valomitted}" = "yes" ]; then
|
|
|
|
## Get the next argument from the argument list, if there is one.
|
1993-05-29 20:07:50 +00:00
|
|
|
if [ $index = $# ]; then
|
1993-05-24 06:08:13 +00:00
|
|
|
(echo "${progname}: You must give a value for the \`--${optname}' option, as in
|
|
|
|
\`--${optname}=FOO'."
|
|
|
|
echo "${short_usage}") >&2
|
|
|
|
exit 1
|
|
|
|
fi
|
1993-05-29 20:07:50 +00:00
|
|
|
index=`expr $index + 1`
|
|
|
|
val=`eval echo '$'$index`
|
1993-05-24 06:08:13 +00:00
|
|
|
fi
|
|
|
|
exec_prefix="${val}"
|
|
|
|
;;
|
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
## Has the user asked for some help?
|
|
|
|
"usage" | "help" )
|
1993-05-23 04:42:38 +00:00
|
|
|
echo "${short_usage}" | more
|
1992-12-06 22:12:37 +00:00
|
|
|
exit
|
|
|
|
;;
|
1992-04-01 10:45:51 +00:00
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
## We ignore all other options silently.
|
|
|
|
esac
|
1992-04-01 10:45:51 +00:00
|
|
|
;;
|
1992-12-06 22:12:37 +00:00
|
|
|
|
|
|
|
## Anything not starting with a hyphen we assume is a
|
|
|
|
## configuration name.
|
1992-04-01 10:45:51 +00:00
|
|
|
*)
|
1992-11-20 17:08:32 +00:00
|
|
|
configuration=${arg}
|
1992-04-01 10:45:51 +00:00
|
|
|
;;
|
1992-12-06 22:12:37 +00:00
|
|
|
|
1992-04-01 10:45:51 +00:00
|
|
|
esac
|
|
|
|
done
|
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
if [ "${configuration}" = "" ]; then
|
1993-05-26 21:34:10 +00:00
|
|
|
echo '- You did not tell me what kind of host system you want to configure.
|
|
|
|
- I will attempt to guess the kind of system this is.' 1>&2
|
|
|
|
guesssys=`echo ${progname} | sed 's/configure$/config.guess/'`
|
|
|
|
if configuration=`${guesssys}` ; then
|
|
|
|
echo "- Looks like this is a ${configuration}" 1>&2
|
|
|
|
else
|
|
|
|
echo '- Failed to guess the system type. You need to tell me.' 1>&2
|
1993-05-27 22:02:16 +00:00
|
|
|
echo "${short_usage}" >&2
|
1993-05-26 21:34:10 +00:00
|
|
|
exit 1
|
|
|
|
fi
|
1992-12-06 22:12:37 +00:00
|
|
|
fi
|
|
|
|
|
|
|
|
#### Decide where the source is.
|
|
|
|
case "${srcdir}" in
|
|
|
|
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
## If it's not specified, see if `.' or `..' might work.
|
1992-12-06 22:12:37 +00:00
|
|
|
"" )
|
1993-05-25 04:56:17 +00:00
|
|
|
confdir=`echo $0 | sed 's|//|/|' | sed 's|/[^/]*$||'`
|
|
|
|
if [ -f $confdir/src/lisp.h -a -f $confdir/lisp/version.el ]; then
|
1993-05-28 03:24:51 +00:00
|
|
|
srcdir="${confdir}"
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
else
|
1993-05-25 04:56:17 +00:00
|
|
|
if [ -f "./src/lisp.h" -a -f "./lisp/version.el" ]; then
|
1993-05-30 18:56:45 +00:00
|
|
|
srcdir='.'
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
else
|
1993-05-25 04:56:17 +00:00
|
|
|
if [ -f "../src/lisp.h" -a -f "../lisp/version.el" ]; then
|
1993-05-30 18:56:45 +00:00
|
|
|
srcdir='..'
|
1993-05-25 04:56:17 +00:00
|
|
|
else
|
|
|
|
(echo "\
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
${progname}: Neither the current directory nor its parent seem to
|
|
|
|
contain the Emacs sources. If you do not want to build Emacs in its
|
|
|
|
source tree, you should run \`${progname}' in the directory in which
|
|
|
|
you wish to build Emacs, using its \`--srcdir' option to say where the
|
|
|
|
sources may be found."
|
1993-05-25 04:56:17 +00:00
|
|
|
echo "${short_usage}") >&2
|
|
|
|
exit 1
|
|
|
|
fi
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
fi
|
|
|
|
fi
|
1992-12-06 22:12:37 +00:00
|
|
|
;;
|
|
|
|
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
## Otherwise, check if the directory they specified is okay.
|
1992-12-06 22:12:37 +00:00
|
|
|
* )
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
if [ ! -d "${srcdir}" -o ! -f "${srcdir}/src/lisp.h" -o ! -f "${srcdir}/lisp/version.el" ]; then
|
|
|
|
(echo "\
|
|
|
|
${progname}: The directory specified with the \`--srcdir' option,
|
|
|
|
\`${srcdir}', doesn't seem to contain the Emacs sources. You should
|
|
|
|
either run the \`${progname}' script at the top of the Emacs source
|
|
|
|
tree, or use the \`--srcdir' option to specify where the Emacs sources
|
|
|
|
are."
|
1992-12-06 22:12:37 +00:00
|
|
|
echo "${short_usage}") >&2
|
|
|
|
exit 1
|
|
|
|
fi
|
|
|
|
;;
|
1993-05-28 03:24:51 +00:00
|
|
|
esac
|
1992-12-06 22:12:37 +00:00
|
|
|
|
1993-05-28 03:24:51 +00:00
|
|
|
#### Make srcdir absolute, if it isn't already. It's important to
|
|
|
|
#### avoid running the path through pwd unnecessary, since pwd can
|
|
|
|
#### give you automounter prefixes, which can go away.
|
|
|
|
case "${srcdir}" in
|
|
|
|
/* ) ;;
|
1993-05-28 04:52:55 +00:00
|
|
|
. )
|
|
|
|
## We may be able to use the $PWD environment variable to make this
|
|
|
|
## absolute. But sometimes PWD is inaccurate.
|
|
|
|
if [ "${PWD}" != "" ] && [ "`(cd ${PWD} ; pwd)`" = "`pwd`" ] ; then
|
|
|
|
srcdir="$PWD"
|
|
|
|
else
|
|
|
|
srcdir="`(cd ${srcdir}; pwd)`"
|
1993-05-29 19:41:13 +00:00
|
|
|
fi
|
1993-05-28 04:52:55 +00:00
|
|
|
;;
|
1993-05-28 03:24:51 +00:00
|
|
|
* ) srcdir="`(cd ${srcdir}; pwd)`" ;;
|
1992-12-06 22:12:37 +00:00
|
|
|
esac
|
|
|
|
|
1993-05-30 18:24:08 +00:00
|
|
|
#### Make sure that the source directory doesn't already have a
|
|
|
|
#### configured system in it.
|
1993-05-30 21:40:18 +00:00
|
|
|
if [ `pwd` != `(cd ${srcdir} && pwd)` ] \
|
|
|
|
&& [ -f "${srcdir}/src/config.h" ] ; then
|
|
|
|
(echo "${progname}: the directory tree \`${srcdir}' is being used"
|
|
|
|
echo " as a build directory right now; it has been configured in its own"
|
|
|
|
echo " right. You can't use srcdir in this situation.") >&2
|
1993-05-30 18:24:08 +00:00
|
|
|
exit 1
|
|
|
|
fi
|
1993-05-28 03:24:51 +00:00
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
### Make the necessary directories, if they don't exist.
|
|
|
|
if [ ! -d ./src ]; then
|
|
|
|
mkdir ./src
|
|
|
|
fi
|
|
|
|
if [ ! -d ./lib-src ]; then
|
|
|
|
mkdir ./lib-src
|
|
|
|
fi
|
|
|
|
if [ ! -d ./cpp ]; then
|
|
|
|
mkdir ./cpp
|
|
|
|
fi
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
if [ ! -d ./oldXMenu ]; then
|
|
|
|
mkdir ./oldXMenu
|
|
|
|
fi
|
1993-05-15 22:57:27 +00:00
|
|
|
if [ ! -d ./etc ]; then
|
|
|
|
mkdir ./etc
|
|
|
|
fi
|
1992-12-06 22:12:37 +00:00
|
|
|
|
|
|
|
#### Given the configuration name, set machfile and opsysfile to the
|
|
|
|
#### names of the m/*.h and s/*.h files we should use.
|
|
|
|
|
|
|
|
### Canonicalize the configuration name.
|
1992-11-20 17:08:32 +00:00
|
|
|
echo "Checking the configuration name."
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
if configuration=`${srcdir}/config.sub "${configuration}"` ; then : ; else
|
1992-11-20 17:08:32 +00:00
|
|
|
exit $?
|
|
|
|
fi
|
1992-10-17 22:07:23 +00:00
|
|
|
|
1993-03-11 06:58:58 +00:00
|
|
|
### If you add support for a new configuration, add code to this
|
|
|
|
### switch statement to recognize your configuration name and select
|
|
|
|
### the appropriate operating system and machine description files.
|
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
### You would hope that you could choose an m/*.h file pretty much
|
|
|
|
### based on the machine portion of the configuration name, and an s-
|
|
|
|
### file based on the operating system portion. However, it turns out
|
|
|
|
### that each m/*.h file is pretty manufacturer-specific - for
|
|
|
|
### example, apollo.h, hp9000s300.h, mega68k, news.h, and tad68k are
|
|
|
|
### all 68000 machines; mips.h, pmax.h, and news-risc are all MIPS
|
|
|
|
### machines. So we basically have to have a special case for each
|
|
|
|
### configuration name.
|
1993-03-11 06:58:58 +00:00
|
|
|
###
|
1992-11-20 17:08:32 +00:00
|
|
|
### As far as handling version numbers on operating systems is
|
|
|
|
### concerned, make sure things will fail in a fixable way. If
|
|
|
|
### /etc/MACHINES doesn't say anything about version numbers, be
|
|
|
|
### prepared to handle anything reasonably. If version numbers
|
|
|
|
### matter, be sure /etc/MACHINES says something about it.
|
1993-03-11 06:58:58 +00:00
|
|
|
###
|
|
|
|
### Eric Raymond says we should accept strings like "sysvr4" to mean
|
|
|
|
### "System V Release 4"; he writes, "The old convention encouraged
|
|
|
|
### confusion between `system' and `release' levels'."
|
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
machine='' opsys='' unported='false'
|
|
|
|
case "${configuration}" in
|
|
|
|
|
|
|
|
## Alliant machines
|
1992-10-17 22:07:23 +00:00
|
|
|
## Strictly speaking, we need the version of the alliant operating
|
|
|
|
## system to choose the right machine file, but currently the
|
|
|
|
## configuration name doesn't tell us enough to choose the right
|
|
|
|
## one; we need to give alliants their own operating system name to
|
|
|
|
## do this right. When someone cares, they can help us.
|
|
|
|
fx80-alliant-* )
|
1992-11-20 17:08:32 +00:00
|
|
|
machine=alliant4 opsys=bsd4-2
|
1992-10-17 22:07:23 +00:00
|
|
|
;;
|
|
|
|
i860-alliant-* )
|
1992-11-20 17:08:32 +00:00
|
|
|
machine=alliant-2800 opsys=bsd4-3
|
1992-10-17 22:07:23 +00:00
|
|
|
;;
|
1992-04-01 10:45:51 +00:00
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
## Altos 3068
|
|
|
|
m68*-altos-sysv* )
|
|
|
|
machine=altos opsys=usg5-2
|
1992-10-17 22:07:23 +00:00
|
|
|
;;
|
1992-11-20 17:08:32 +00:00
|
|
|
|
|
|
|
## Amdahl UTS
|
|
|
|
580-amdahl-sysv* )
|
|
|
|
machine=amdahl opsys=usg5-2-2
|
1992-10-17 22:07:23 +00:00
|
|
|
;;
|
1992-04-01 10:45:51 +00:00
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
## Appallings - I mean, Apollos - running Domain
|
|
|
|
m68*-apollo* )
|
The GNU coding standards specify that CFLAGS should be left for
users to set.
* Makefile.in (CFLAGS): Let configure determine the default value
for this. Don't
have it default to DEFS.
(${SUBDIR}): Pass CFLAGS down to submakes, not DEFS.
(lib-src/Makefile, src/Makefile): Edit the default value for
CFLAGS into these files, not DEFS.
* configure.in (CFLAGS): Choose a default value for this - "-g"
normally, or "-g -O" if we're using GCC. Edit it into the
top-level Makefile.
* configure.in: When scanning the machine and system description
#include files, write their names to conftest.c properly.
* configure.in: In configuration name case for Apallos running
Domainios, set opsys, not opsysfile.
* configure.in: Use the autoconf AC_CONFIG_HEADER macro to produce
src/config.h, instead of AC_OUTPUT; the latter overwrites
src/config.h even when it hasn't changed, puts a makefile-style
comment at the top even though it's C code, and produces a
config.status script which doesn't do the job right.
* configure.in: Add AC_LN_S test, so we can tell whether or not we
can use a symbolic link to get the X Menu library into src.
* Makefile.in (LN_S): New variable.
(src/Makefile): Edit the value of LN_S into this makefile.
1993-05-09 22:52:15 +00:00
|
|
|
machine=apollo opsys=bsd4-2
|
1992-10-17 22:07:23 +00:00
|
|
|
;;
|
1992-11-20 17:08:32 +00:00
|
|
|
|
|
|
|
## AT&T 3b2, 3b5, 3b15, 3b20
|
|
|
|
we32k-att-sysv* )
|
|
|
|
machine=att3b opsys=usg5-2-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## AT&T 3b1 - The Mighty Unix PC!
|
|
|
|
m68*-att-sysv* )
|
|
|
|
machine=7300 opsys=usg5-2-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Bull sps7
|
|
|
|
m68*-bull-sysv* )
|
|
|
|
machine=sps7 opsys=usg5-2
|
1992-10-17 22:07:23 +00:00
|
|
|
;;
|
1992-11-20 17:08:32 +00:00
|
|
|
|
|
|
|
## CCI 5/32, 6/32 -- see "Tahoe".
|
|
|
|
|
|
|
|
## Celerity
|
|
|
|
## I don't know what configuration name to use for this; config.sub
|
|
|
|
## doesn't seem to know anything about it. Hey, Celerity users, get
|
|
|
|
## in touch with us!
|
|
|
|
celerity-celerity-bsd* )
|
|
|
|
machine=celerity opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Clipper
|
|
|
|
## What operating systems does this chip run that Emacs has been
|
|
|
|
## tested on?
|
|
|
|
clipper-* )
|
|
|
|
machine=clipper
|
|
|
|
## We'll use the catch-all code at the bottom to guess the
|
|
|
|
## operating system.
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Convex
|
|
|
|
*-convex-bsd* )
|
|
|
|
machine=convex opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Cubix QBx/386
|
|
|
|
i386-cubix-sysv* )
|
|
|
|
machine=intel386 opsys=usg5-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Cydra 5
|
|
|
|
cydra*-cydrome-sysv* )
|
|
|
|
machine=cydra5 opsys=usg5-3
|
|
|
|
;;
|
|
|
|
|
1993-06-10 05:46:10 +00:00
|
|
|
## Data General AViiON Machines
|
|
|
|
m88k-dg-dgux* )
|
|
|
|
machine=aviion opsys=dgux
|
|
|
|
;;
|
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
## DECstations
|
|
|
|
mips-dec-ultrix[0-3].* | mips-dec-ultrix4.0 | mips-dec-bsd4.2 )
|
|
|
|
machine=pmax opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
mips-dec-ultrix* | mips-dec-bsd* )
|
|
|
|
machine=pmax opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
mips-dec-osf* )
|
|
|
|
machine=pmax opsys=osf1
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Motorola Delta machines
|
|
|
|
m68*-motorola-sysv* )
|
|
|
|
machine=delta opsys=usg5-3
|
|
|
|
;;
|
|
|
|
m88k-motorola-sysv* | m88k-motorola-m88kbcs* )
|
|
|
|
machine=delta88k opsys=usg5-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Dual machines
|
|
|
|
m68*-dual-sysv* )
|
|
|
|
machine=dual opsys=usg5-2
|
|
|
|
;;
|
|
|
|
m68*-dual-uniplus* )
|
|
|
|
machine=dual opsys=unipl5-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Elxsi 6400
|
|
|
|
elxsi-elxsi-sysv* )
|
|
|
|
machine=elxsi opsys=usg5-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Encore machines
|
|
|
|
ns16k-encore-bsd* )
|
|
|
|
machine=ns16000 opsys=umax
|
|
|
|
;;
|
|
|
|
|
|
|
|
## The GEC 93 - apparently, this port isn't really finished yet.
|
|
|
|
|
|
|
|
## Gould Power Node and NP1
|
|
|
|
pn-gould-bsd4.2 )
|
|
|
|
machine=gould opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
pn-gould-bsd4.3 )
|
|
|
|
machine=gould opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
np1-gould-bsd* )
|
|
|
|
machine=gould-np1 opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Honeywell XPS100
|
|
|
|
xps*-honeywell-sysv* )
|
|
|
|
machine=xps100 opsys=usg5-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## HP 9000 series 200 or 300
|
|
|
|
m68*-hp-bsd* )
|
|
|
|
machine=hp9000s300 opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
## HP/UX 8 doesn't run on these machines, so use HP/UX 7.
|
1992-10-17 22:07:23 +00:00
|
|
|
m68*-hp-hpux* )
|
1992-11-20 17:08:32 +00:00
|
|
|
machine=hp9000s300 opsys=hpux
|
1992-10-17 22:07:23 +00:00
|
|
|
;;
|
1992-04-01 10:45:51 +00:00
|
|
|
|
1993-05-24 05:06:15 +00:00
|
|
|
## HP 9000 series 700 and 800, running HP/UX
|
1993-06-08 07:17:32 +00:00
|
|
|
hppa*-hp-hpux7* )
|
1993-06-08 07:21:54 +00:00
|
|
|
machine=hp9000s800 opsys=hpux
|
1993-05-24 05:06:15 +00:00
|
|
|
;;
|
1993-06-08 07:17:32 +00:00
|
|
|
hppa*-hp-hpux8* )
|
1993-06-08 07:21:54 +00:00
|
|
|
machine=hp9000s800 opsys=hpux8
|
1993-05-24 05:06:15 +00:00
|
|
|
;;
|
1993-06-08 07:17:32 +00:00
|
|
|
hppa*-hp-hpux9* )
|
1993-06-08 07:21:54 +00:00
|
|
|
machine=hp9000s800 opsys=hpux9
|
1993-06-08 07:17:32 +00:00
|
|
|
;;
|
|
|
|
|
|
|
|
## HP 9000 series 700 and 800, running HP/UX
|
1993-05-24 05:06:15 +00:00
|
|
|
hppa*-hp-hpux* )
|
|
|
|
## Cross-compilation? Nah!
|
1993-05-26 18:32:19 +00:00
|
|
|
case "`uname -r`" in
|
1993-06-08 07:17:32 +00:00
|
|
|
*.08.* ) machine=hp9000s800 opsys=hpux8 ;;
|
|
|
|
*.09.* ) machine=hp9000s800 opsys=hpux9 ;;
|
1993-05-26 18:32:19 +00:00
|
|
|
*) machine=hp9000s800 opsys=hpux ;;
|
1993-05-24 05:06:15 +00:00
|
|
|
esac
|
1992-11-20 17:08:32 +00:00
|
|
|
;;
|
|
|
|
|
|
|
|
## Orion machines
|
|
|
|
orion-orion-bsd* )
|
|
|
|
machine=orion opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
clipper-orion-bsd* )
|
|
|
|
machine=orion105 opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## IBM machines
|
|
|
|
i386-ibm-aix1.1 )
|
|
|
|
machine=ibmps2-aix opsys=usg5-2-2
|
|
|
|
;;
|
1993-05-31 04:24:14 +00:00
|
|
|
i386-ibm-aix1.[23] | i386-ibm-aix* )
|
1992-11-20 17:08:32 +00:00
|
|
|
machine=ibmps2-aix opsys=usg5-3
|
|
|
|
;;
|
1993-03-18 16:10:40 +00:00
|
|
|
rs6000-ibm-aix3.1 )
|
1992-11-20 17:08:32 +00:00
|
|
|
machine=ibmrs6000 opsys=aix3-1
|
|
|
|
;;
|
1993-03-18 16:10:40 +00:00
|
|
|
rs6000-ibm-aix3.2 | rs6000-ibm-aix* )
|
|
|
|
machine=ibmrs6000 opsys=aix3-2
|
|
|
|
;;
|
1993-05-29 06:10:43 +00:00
|
|
|
romp-ibm-bsd4-3 )
|
|
|
|
machine=ibmrt opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
romp-ibm-bsd4-2 )
|
|
|
|
machine=ibmrt opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
romp-ibm-aos4-3 )
|
|
|
|
machine=ibmrt opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
romp-ibm-aos4-2 )
|
1992-11-20 17:08:32 +00:00
|
|
|
machine=ibmrt opsys=bsd4-2
|
|
|
|
;;
|
1993-05-29 06:10:43 +00:00
|
|
|
romp-ibm-aos* )
|
|
|
|
machine=ibmrt opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
romp-ibm-bsd* )
|
|
|
|
machine=ibmrt opsys=bsd4-3
|
|
|
|
;;
|
1992-11-20 17:08:32 +00:00
|
|
|
romp-ibm-aix* )
|
|
|
|
machine=ibmrt-aix opsys=usg5-2-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Integrated Solutions `Optimum V'
|
|
|
|
m68*-isi-bsd4.2 )
|
|
|
|
machine=isi-ov opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
m68*-isi-bsd4.3 )
|
|
|
|
machine=isi-ov opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Intel 386 machines where we do care about the manufacturer
|
|
|
|
i[34]86-intsys-sysv* )
|
|
|
|
machine=is386 opsys=usg5-2-2
|
|
|
|
;;
|
|
|
|
## Intel 386 machines where we don't care about the manufacturer
|
1993-06-03 20:43:29 +00:00
|
|
|
i[34]86-*-* )
|
1992-11-20 17:08:32 +00:00
|
|
|
machine=intel386
|
|
|
|
case "${configuration}" in
|
|
|
|
*-isc1.* | *-isc2.[01]* ) opsys=386-ix ;;
|
1993-03-04 14:43:09 +00:00
|
|
|
*-isc2.2 ) opsys=isc2-2 ;;
|
|
|
|
*-isc* ) opsys=isc3-0 ;;
|
1993-05-23 02:53:53 +00:00
|
|
|
*-esix5* ) opsys=esix5r4 ;;
|
1992-11-20 17:08:32 +00:00
|
|
|
*-esix* ) opsys=esix ;;
|
|
|
|
*-xenix* ) opsys=xenix ;;
|
1993-05-20 00:56:54 +00:00
|
|
|
*-linux* ) opsys=linux ;;
|
1993-05-06 18:42:47 +00:00
|
|
|
*-sco3.2v4* ) opsys=sco4 ;;
|
1993-05-31 21:28:10 +00:00
|
|
|
*-bsd386* ) opsys=bsd386 ;;
|
1993-06-01 04:12:00 +00:00
|
|
|
*-386bsd ) opsys=386bsd ;;
|
1992-11-20 17:08:32 +00:00
|
|
|
## Otherwise, we'll fall through to the generic opsys code at the bottom.
|
|
|
|
esac
|
|
|
|
;;
|
|
|
|
|
1993-05-25 04:39:28 +00:00
|
|
|
## Intel 860
|
|
|
|
i860-*-sysvr4 )
|
|
|
|
machine=i860 opsys=usg5-4
|
|
|
|
;;
|
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
## Silicon Graphics machines
|
|
|
|
## Iris 2500 and Iris 2500 Turbo (aka the Iris 3030)
|
|
|
|
m68*-sgi-iris3.5 )
|
|
|
|
machine=irist opsys=iris3-5
|
|
|
|
;;
|
|
|
|
m68*-sgi-iris3.6 | m68*-sgi-iris*)
|
|
|
|
machine=irist opsys=iris3-6
|
|
|
|
;;
|
|
|
|
## Iris 4D
|
|
|
|
mips-sgi-irix3.* )
|
|
|
|
machine=iris4d opsys=irix3-3
|
|
|
|
;;
|
|
|
|
mips-sgi-irix4.* | mips-sgi-irix* )
|
|
|
|
machine=iris4d opsys=irix4-0
|
1992-10-17 22:07:23 +00:00
|
|
|
;;
|
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
## Masscomp machines
|
|
|
|
m68*-masscomp-rtu )
|
|
|
|
machine=masscomp opsys=rtu
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Megatest machines
|
|
|
|
m68*-megatest-bsd* )
|
|
|
|
machine=mega68 opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Workstations sold by MIPS
|
|
|
|
## This is not necessarily all workstations using the MIPS processor -
|
|
|
|
## Irises are produced by SGI, and DECstations by DEC.
|
|
|
|
|
|
|
|
## etc/MACHINES lists mips.h and mips4.h as possible machine files,
|
|
|
|
## and usg5-2-2 and bsd4-3 as possible OS files. The only guidance
|
|
|
|
## it gives for choosing between the alternatives seems to be "Use
|
|
|
|
## -machine=mips4 for RISCOS version 4; use -opsystem=bsd4-3 with
|
|
|
|
## the BSD world." I'll assume that these are instructions for
|
|
|
|
## handling two odd situations, and that every other situation
|
|
|
|
## should use mips.h and usg5-2-2, they being listed first.
|
1993-05-30 21:46:11 +00:00
|
|
|
mips-mips-usg* )
|
|
|
|
machine=mips4
|
|
|
|
## Fall through to the general code at the bottom to decide on the OS.
|
|
|
|
;;
|
1992-11-20 17:08:32 +00:00
|
|
|
mips-mips-riscos4* )
|
1993-05-30 21:40:18 +00:00
|
|
|
machine=mips4 opsys=bsd4-3
|
1992-11-20 17:08:32 +00:00
|
|
|
;;
|
|
|
|
mips-mips-bsd* )
|
|
|
|
machine=mips opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
mips-mips-* )
|
|
|
|
machine=mips opsys=usg5-2-2
|
|
|
|
;;
|
|
|
|
|
1993-05-03 02:07:41 +00:00
|
|
|
## NeXT
|
|
|
|
m68*-next-mach* | m68*-next-bsd* )
|
|
|
|
machine=next opsys=mach2
|
|
|
|
;;
|
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
## The complete machine from National Semiconductor
|
|
|
|
ns32k-ns-genix* )
|
|
|
|
machine=ns32000 opsys=usg5-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## NCR machines
|
1993-03-11 06:58:58 +00:00
|
|
|
m68*-ncr-sysv2* | m68*-ncr-sysvr2* )
|
1992-11-20 17:08:32 +00:00
|
|
|
machine=tower32 opsys=usg5-2-2
|
|
|
|
;;
|
1993-03-11 06:58:58 +00:00
|
|
|
m68*-ncr-sysv3* | m68*-ncr-sysvr3* )
|
1992-11-20 17:08:32 +00:00
|
|
|
machine=tower32v3 opsys=usg5-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Nixdorf Targon 31
|
|
|
|
m68*-nixdorf-sysv* )
|
|
|
|
machine=targon31 opsys=usg5-2-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Nu (TI or LMI)
|
|
|
|
m68*-nu-sysv* )
|
|
|
|
machine=nu opsys=usg5-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Plexus
|
|
|
|
m68*-plexus-sysv* )
|
|
|
|
machine=plexus opsys=usg5-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Prime EXL
|
|
|
|
i386-prime-sysv* )
|
|
|
|
machine=i386 opsys=usg5-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Pyramid machines
|
|
|
|
## I don't really have any idea what sort of processor the Pyramid has,
|
|
|
|
## so I'm assuming it is its own architecture.
|
|
|
|
pyramid-pyramid-bsd* )
|
|
|
|
machine=pyramid opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Sequent Balance
|
|
|
|
ns32k-sequent-bsd4.2 )
|
|
|
|
machine=sequent opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
ns32k-sequent-bsd4.3 )
|
|
|
|
machine=sequent opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
## Sequent Symmetry
|
|
|
|
i386-sequent-bsd* )
|
|
|
|
machine=symmetry opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## SONY machines
|
|
|
|
m68*-sony-bsd4.2 )
|
|
|
|
machine=news opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
m68*-sony-bsd4.3 )
|
|
|
|
machine=news opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
mips-sony-bsd* )
|
|
|
|
machine=news-risc opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Stride
|
|
|
|
m68*-stride-sysv* )
|
|
|
|
machine=stride opsys=usg5-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Suns
|
1993-05-15 22:57:27 +00:00
|
|
|
*-sun-sunos* | *-sun-bsd* | *-sun-solaris* )
|
1992-11-20 17:08:32 +00:00
|
|
|
case "${configuration}" in
|
|
|
|
m68*-sunos1* ) machine=sun1 ;;
|
|
|
|
m68*-sunos2* ) machine=sun2 ;;
|
|
|
|
m68* ) machine=sun3 ;;
|
|
|
|
i[34]86* ) machine=sun386 ;;
|
|
|
|
sparc* ) machine=sparc ;;
|
|
|
|
* ) unported=true ;;
|
|
|
|
esac
|
|
|
|
case "${configuration}" in
|
|
|
|
*-sunos4.0* ) opsys=sunos4-0 ;;
|
1993-05-26 17:43:22 +00:00
|
|
|
*-sunos4.1.3* ) opsys=sunos4-1-3 ;;
|
1992-11-20 17:08:32 +00:00
|
|
|
*-sunos4* | *-sunos ) opsys=sunos4-1 ;;
|
1993-05-15 22:57:27 +00:00
|
|
|
*-sunos5* | *-solaris* ) opsys=sol2 ;;
|
1992-11-20 17:08:32 +00:00
|
|
|
* ) opsys=bsd4-2 ;;
|
1992-10-17 22:07:23 +00:00
|
|
|
esac
|
1992-11-20 17:08:32 +00:00
|
|
|
;;
|
|
|
|
|
|
|
|
## Tadpole 68k
|
|
|
|
m68*-tadpole-sysv* )
|
|
|
|
machine=tad68k opsys=usg5-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Tahoe machines
|
|
|
|
tahoe-tahoe-bsd4.2 )
|
|
|
|
machine=tahoe opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
tahoe-tahoe-bsd4.3 )
|
|
|
|
machine=tahoe opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Tandem Integrity S2
|
|
|
|
mips-tandem-sysv* )
|
|
|
|
machine=tandem-s2 opsys=usg5-3
|
|
|
|
;;
|
|
|
|
|
1993-05-29 20:52:45 +00:00
|
|
|
## Tektronix XD88
|
|
|
|
m88k-tektronix-sysv3 )
|
|
|
|
machine=tekXD88 opsys=usg5-3
|
|
|
|
;;
|
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
## Tektronix 16000 box (6130?)
|
|
|
|
ns16k-tektronix-bsd* )
|
|
|
|
machine=ns16000 opsys=bsd4-2
|
|
|
|
;;
|
|
|
|
## Tektronix 4300
|
|
|
|
## src/m/tek4300.h hints that this is a m68k machine.
|
|
|
|
m68*-tektronix-bsd* )
|
|
|
|
machine=tex4300 opsys=bsd4-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Titan P2 or P3
|
|
|
|
## We seem to have lost the machine-description file titan.h!
|
|
|
|
titan-titan-sysv* )
|
|
|
|
machine=titan opsys=usg5-3
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Ustation E30 (SS5E)
|
|
|
|
m68*-unisys-uniplus* )
|
|
|
|
machine=ustation opsystem=unipl5-2
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Vaxen.
|
|
|
|
vax-dec-* )
|
|
|
|
machine=vax
|
|
|
|
case "${configuration}" in
|
1993-03-11 06:58:58 +00:00
|
|
|
*-bsd4.1 ) opsys=bsd4-1 ;;
|
|
|
|
*-bsd4.2 | *-ultrix[0-3].* | *-ultrix4.0 ) opsys=bsd4-2 ;;
|
|
|
|
*-bsd4.3 | *-ultrix* ) opsys=bsd4-3 ;;
|
1993-05-31 21:28:10 +00:00
|
|
|
*-bsd386 ) opsys=bsd386 ;;
|
1993-03-11 06:58:58 +00:00
|
|
|
*-sysv[01]* | *-sysvr[01]* ) opsys=usg5-0 ;;
|
|
|
|
*-sysv2* | *-sysvr2* ) opsys=usg5-2 ;;
|
|
|
|
*-vms* ) opsys=vms ;;
|
|
|
|
* ) unported=true
|
1992-10-17 22:07:23 +00:00
|
|
|
esac
|
|
|
|
;;
|
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
## Whitechapel MG1
|
|
|
|
ns16k-whitechapel-* )
|
|
|
|
machine=mg1
|
|
|
|
## We don't know what sort of OS runs on these; we'll let the
|
|
|
|
## operating system guessing code below try.
|
|
|
|
;;
|
|
|
|
|
|
|
|
## Wicat
|
|
|
|
m68*-wicat-sysv* )
|
|
|
|
machine=wicat opsys=usg5-2
|
|
|
|
;;
|
|
|
|
|
1992-10-17 22:07:23 +00:00
|
|
|
* )
|
1992-11-20 17:08:32 +00:00
|
|
|
unported=true
|
1992-10-17 22:07:23 +00:00
|
|
|
;;
|
|
|
|
esac
|
1992-04-01 10:45:51 +00:00
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
### If the code above didn't choose an operating system, just choose
|
|
|
|
### an operating system based on the configuration name. You really
|
|
|
|
### only want to use this when you have no idea what the right
|
|
|
|
### operating system is; if you know what operating systems a machine
|
|
|
|
### runs, it's cleaner to make it explicit in the case statement
|
|
|
|
### above.
|
1993-05-31 21:28:10 +00:00
|
|
|
if [ x"${opsys}" = x ]; then
|
1992-11-20 17:08:32 +00:00
|
|
|
case "${configuration}" in
|
|
|
|
*-bsd4.[01] ) opsys=bsd4-1 ;;
|
|
|
|
*-bsd4.2 ) opsys=bsd4-2 ;;
|
|
|
|
*-bsd4.3 ) opsys=bsd4-3 ;;
|
1993-03-11 06:58:58 +00:00
|
|
|
*-sysv0 | *-sysvr0 ) opsys=usg5-0 ;;
|
|
|
|
*-sysv2 | *-sysvr2 ) opsys=usg5-2 ;;
|
|
|
|
*-sysv2.2 | *-sysvr2.2 ) opsys=usg5-2-2 ;;
|
|
|
|
*-sysv3 | *-sysvr3 ) opsys=usg5-3 ;;
|
|
|
|
*-sysv4 | *-sysvr4 ) opsys=usg5-4 ;;
|
1993-05-29 20:31:06 +00:00
|
|
|
*-sysv4.2 | *-sysvr4.2 ) opsys=usg5-4-2 ;;
|
1992-11-20 17:08:32 +00:00
|
|
|
* )
|
|
|
|
unported=true
|
|
|
|
;;
|
|
|
|
esac
|
|
|
|
fi
|
|
|
|
|
|
|
|
if $unported ; then
|
|
|
|
(echo "${progname}: Emacs hasn't been ported to \`${configuration}' systems."
|
|
|
|
echo "${progname}: Check \`etc/MACHINES' for recognized configuration names."
|
|
|
|
) >&2
|
|
|
|
exit 1
|
|
|
|
fi
|
|
|
|
|
|
|
|
machfile="m/${machine}.h"
|
|
|
|
opsysfile="s/${opsys}.h"
|
|
|
|
|
1993-05-03 02:07:41 +00:00
|
|
|
]
|
|
|
|
AC_PREPARE(lisp)
|
The GNU coding standards specify that CFLAGS should be left for
users to set.
* Makefile.in (CFLAGS): Let configure determine the default value
for this. Don't
have it default to DEFS.
(${SUBDIR}): Pass CFLAGS down to submakes, not DEFS.
(lib-src/Makefile, src/Makefile): Edit the default value for
CFLAGS into these files, not DEFS.
* configure.in (CFLAGS): Choose a default value for this - "-g"
normally, or "-g -O" if we're using GCC. Edit it into the
top-level Makefile.
* configure.in: When scanning the machine and system description
#include files, write their names to conftest.c properly.
* configure.in: In configuration name case for Apallos running
Domainios, set opsys, not opsysfile.
* configure.in: Use the autoconf AC_CONFIG_HEADER macro to produce
src/config.h, instead of AC_OUTPUT; the latter overwrites
src/config.h even when it hasn't changed, puts a makefile-style
comment at the top even though it's C code, and produces a
config.status script which doesn't do the job right.
* configure.in: Add AC_LN_S test, so we can tell whether or not we
can use a symbolic link to get the X Menu library into src.
* Makefile.in (LN_S): New variable.
(src/Makefile): Edit the value of LN_S into this makefile.
1993-05-09 22:52:15 +00:00
|
|
|
AC_CONFIG_HEADER(src/config.h)
|
1993-05-03 02:07:41 +00:00
|
|
|
[
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
|
|
|
|
#### Choose a compiler.
|
|
|
|
case ${with_gcc} in
|
1993-04-10 08:30:27 +00:00
|
|
|
"yes" ) CC="gcc" GCC=1 ;;
|
|
|
|
"no" ) CC="cc" ;;
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
* )
|
The GNU coding standards specify that CFLAGS should be left for
users to set.
* Makefile.in (CFLAGS): Let configure determine the default value
for this. Don't
have it default to DEFS.
(${SUBDIR}): Pass CFLAGS down to submakes, not DEFS.
(lib-src/Makefile, src/Makefile): Edit the default value for
CFLAGS into these files, not DEFS.
* configure.in (CFLAGS): Choose a default value for this - "-g"
normally, or "-g -O" if we're using GCC. Edit it into the
top-level Makefile.
* configure.in: When scanning the machine and system description
#include files, write their names to conftest.c properly.
* configure.in: In configuration name case for Apallos running
Domainios, set opsys, not opsysfile.
* configure.in: Use the autoconf AC_CONFIG_HEADER macro to produce
src/config.h, instead of AC_OUTPUT; the latter overwrites
src/config.h even when it hasn't changed, puts a makefile-style
comment at the top even though it's C code, and produces a
config.status script which doesn't do the job right.
* configure.in: Add AC_LN_S test, so we can tell whether or not we
can use a symbolic link to get the X Menu library into src.
* Makefile.in (LN_S): New variable.
(src/Makefile): Edit the value of LN_S into this makefile.
1993-05-09 22:52:15 +00:00
|
|
|
] AC_PROG_CC [
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
esac
|
|
|
|
|
1993-05-18 22:51:56 +00:00
|
|
|
#### Some other nice autoconf tests. If you add a test here which
|
|
|
|
#### should make an entry in src/config.h, don't forget to add an
|
|
|
|
#### #undef clause to src/config.h.in for autoconf to modify.
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
]
|
1993-05-18 22:51:56 +00:00
|
|
|
dnl checks for programs
|
The GNU coding standards specify that CFLAGS should be left for
users to set.
* Makefile.in (CFLAGS): Let configure determine the default value
for this. Don't
have it default to DEFS.
(${SUBDIR}): Pass CFLAGS down to submakes, not DEFS.
(lib-src/Makefile, src/Makefile): Edit the default value for
CFLAGS into these files, not DEFS.
* configure.in (CFLAGS): Choose a default value for this - "-g"
normally, or "-g -O" if we're using GCC. Edit it into the
top-level Makefile.
* configure.in: When scanning the machine and system description
#include files, write their names to conftest.c properly.
* configure.in: In configuration name case for Apallos running
Domainios, set opsys, not opsysfile.
* configure.in: Use the autoconf AC_CONFIG_HEADER macro to produce
src/config.h, instead of AC_OUTPUT; the latter overwrites
src/config.h even when it hasn't changed, puts a makefile-style
comment at the top even though it's C code, and produces a
config.status script which doesn't do the job right.
* configure.in: Add AC_LN_S test, so we can tell whether or not we
can use a symbolic link to get the X Menu library into src.
* Makefile.in (LN_S): New variable.
(src/Makefile): Edit the value of LN_S into this makefile.
1993-05-09 22:52:15 +00:00
|
|
|
AC_LN_S
|
1993-05-03 02:07:41 +00:00
|
|
|
AC_PROG_CPP
|
1993-05-24 06:08:13 +00:00
|
|
|
AC_PROG_INSTALL
|
1993-06-09 04:06:57 +00:00
|
|
|
AC_PROG_YACC
|
1993-05-18 22:51:56 +00:00
|
|
|
|
|
|
|
dnl checks for UNIX variants that set `DEFS'
|
|
|
|
|
|
|
|
dnl checks for header files
|
1993-05-04 14:17:07 +00:00
|
|
|
AC_HAVE_HEADERS(sys/timeb.h sys/time.h)
|
1993-05-14 14:27:17 +00:00
|
|
|
AC_STDC_HEADERS
|
1993-05-15 20:48:34 +00:00
|
|
|
AC_TIME_WITH_SYS_TIME
|
1993-05-18 22:51:56 +00:00
|
|
|
|
1993-06-08 07:18:40 +00:00
|
|
|
dnl checks for library files
|
|
|
|
AC_HAVE_LIBRARY(-ldnet)
|
|
|
|
|
1993-05-18 22:51:56 +00:00
|
|
|
dnl checks for typedefs
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
AC_RETSIGTYPE
|
1993-05-18 22:51:56 +00:00
|
|
|
|
|
|
|
dnl checks for functions
|
1993-05-03 02:07:41 +00:00
|
|
|
AC_ALLOCA
|
1993-05-24 05:06:15 +00:00
|
|
|
AC_HAVE_FUNCS(gettimeofday gethostname dup2 rename)
|
1993-05-18 22:51:56 +00:00
|
|
|
|
|
|
|
dnl checks for structure members
|
1993-05-04 14:17:07 +00:00
|
|
|
AC_STRUCT_TM
|
1993-05-20 06:18:08 +00:00
|
|
|
AC_TIMEZONE
|
1993-05-18 22:51:56 +00:00
|
|
|
|
|
|
|
dnl checks for compiler characteristics
|
1993-05-03 02:07:41 +00:00
|
|
|
AC_CONST
|
1993-05-18 22:51:56 +00:00
|
|
|
|
|
|
|
dnl checks for operating system services
|
|
|
|
|
|
|
|
dnl other checks for UNIX variants
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
[
|
|
|
|
|
1992-04-01 10:45:51 +00:00
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
#### Choose a window system.
|
1992-04-19 08:53:55 +00:00
|
|
|
echo "Checking window system."
|
1993-05-03 02:07:41 +00:00
|
|
|
|
1992-11-20 17:08:32 +00:00
|
|
|
window_system=''
|
1992-10-17 22:07:23 +00:00
|
|
|
case "${with_x}" in
|
1992-09-14 22:08:02 +00:00
|
|
|
yes )
|
|
|
|
window_system=${window_system}x11
|
1992-10-17 22:07:23 +00:00
|
|
|
;;
|
|
|
|
no )
|
|
|
|
window_system=${window_system}none
|
1992-09-14 22:08:02 +00:00
|
|
|
esac
|
1992-10-17 22:07:23 +00:00
|
|
|
case "${with_x11}" in
|
1992-09-14 22:08:02 +00:00
|
|
|
yes )
|
|
|
|
window_system=${window_system}x11
|
1992-10-17 22:07:23 +00:00
|
|
|
;;
|
1992-09-14 22:08:02 +00:00
|
|
|
esac
|
|
|
|
case "${with_x10}" in
|
|
|
|
yes )
|
|
|
|
window_system=${window_system}x10
|
1992-10-17 22:07:23 +00:00
|
|
|
;;
|
1992-09-14 22:08:02 +00:00
|
|
|
esac
|
|
|
|
|
1992-04-01 10:45:51 +00:00
|
|
|
case "${window_system}" in
|
|
|
|
"none" | "x11" | "x10" ) ;;
|
|
|
|
"" )
|
1993-06-03 20:43:29 +00:00
|
|
|
echo " No window system specified. Looking for X11."
|
1992-04-01 10:45:51 +00:00
|
|
|
window_system=none
|
1993-03-23 06:54:37 +00:00
|
|
|
if [ -r /usr/lib/libX11.a \
|
|
|
|
-o -d /usr/include/X11 \
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
-o -d /usr/X386/include \
|
|
|
|
-o -d ${x_includes}/X11 ]; then
|
1992-04-01 10:45:51 +00:00
|
|
|
window_system=x11
|
1992-04-19 08:53:55 +00:00
|
|
|
fi
|
1992-04-01 10:45:51 +00:00
|
|
|
;;
|
|
|
|
* )
|
1993-05-25 18:07:11 +00:00
|
|
|
echo "Don't specify the window system more than once." >&2
|
1992-04-01 10:45:51 +00:00
|
|
|
exit 1
|
|
|
|
;;
|
|
|
|
esac
|
|
|
|
|
|
|
|
case "${window_system}" in
|
|
|
|
x11 )
|
1992-12-06 22:12:37 +00:00
|
|
|
HAVE_X_WINDOWS=yes
|
|
|
|
HAVE_X11=yes
|
1992-11-20 17:08:32 +00:00
|
|
|
echo " Using X11."
|
1992-04-01 10:45:51 +00:00
|
|
|
;;
|
|
|
|
x10 )
|
1992-12-06 22:12:37 +00:00
|
|
|
HAVE_X_WINDOWS=yes
|
|
|
|
HAVE_X11=no
|
1992-11-20 17:08:32 +00:00
|
|
|
echo " Using X10."
|
1992-04-01 10:45:51 +00:00
|
|
|
;;
|
|
|
|
none )
|
1992-12-06 22:12:37 +00:00
|
|
|
HAVE_X_WINDOWS=no
|
|
|
|
HAVE_X11=no
|
1992-11-20 17:08:32 +00:00
|
|
|
echo " Using no window system."
|
1992-04-01 10:45:51 +00:00
|
|
|
;;
|
|
|
|
esac
|
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
### If we're using X11, we should use the X menu package.
|
|
|
|
HAVE_X_MENU=no
|
|
|
|
case ${HAVE_X11} in
|
1992-10-17 22:07:23 +00:00
|
|
|
yes )
|
1992-12-06 22:12:37 +00:00
|
|
|
HAVE_X_MENU=yes
|
1992-04-01 10:45:51 +00:00
|
|
|
;;
|
|
|
|
esac
|
|
|
|
|
1993-03-23 06:54:37 +00:00
|
|
|
### Check for XFree386. It needs special hacks.
|
1993-05-03 02:07:41 +00:00
|
|
|
lib_havexbsd=no
|
|
|
|
]
|
|
|
|
AC_HAVE_LIBRARY( Xbsd , have_libxbsd=yes , have_libxbsd=no )
|
|
|
|
[
|
|
|
|
if [ -n "${x_libraries}" ] && [ -f ${x_libraries}/libXbsd.a ]; then
|
|
|
|
have_libxbsd=yes
|
|
|
|
fi
|
|
|
|
|
1993-03-23 06:54:37 +00:00
|
|
|
case ${window_system} in
|
|
|
|
x11 )
|
1993-05-03 02:07:41 +00:00
|
|
|
if [ -d /usr/X386/include ] && [ "${have_libxbsd}" = "yes" ]; then
|
1993-03-23 06:54:37 +00:00
|
|
|
HAVE_XFREE386=yes
|
|
|
|
if [ "${C_SWITCH_X_SITE}" = "" ]; then
|
1993-03-25 01:14:34 +00:00
|
|
|
C_SWITCH_X_SITE="-I/usr/X386/include"
|
1993-03-23 06:54:37 +00:00
|
|
|
fi
|
|
|
|
fi
|
|
|
|
;;
|
|
|
|
esac
|
1992-12-06 22:12:37 +00:00
|
|
|
|
|
|
|
#### Extract some information from the operating system and machine files.
|
1992-04-24 08:11:54 +00:00
|
|
|
|
1992-04-19 08:53:55 +00:00
|
|
|
echo "Examining the machine- and system-dependent files to find out"
|
|
|
|
echo " - which libraries the lib-src programs will want, and"
|
|
|
|
echo " - whether the GNU malloc routines are usable."
|
1993-05-03 02:07:41 +00:00
|
|
|
|
|
|
|
### It's not important that this name contain the PID; you can't run
|
|
|
|
### two configures in the same directory and have anything work
|
|
|
|
### anyway.
|
|
|
|
tempcname="conftest.c"
|
|
|
|
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
echo '
|
|
|
|
#include "'${srcdir}'/src/'${opsysfile}'"
|
|
|
|
#include "'${srcdir}'/src/'${machfile}'"
|
1992-04-01 10:45:51 +00:00
|
|
|
#ifndef LIBS_MACHINE
|
|
|
|
#define LIBS_MACHINE
|
|
|
|
#endif
|
|
|
|
#ifndef LIBS_SYSTEM
|
|
|
|
#define LIBS_SYSTEM
|
|
|
|
#endif
|
1993-03-25 03:51:20 +00:00
|
|
|
#ifndef C_SWITCH_SYSTEM
|
|
|
|
#define C_SWITCH_SYSTEM
|
1993-03-25 03:57:51 +00:00
|
|
|
#endif
|
1992-04-24 08:11:54 +00:00
|
|
|
@configure@ libsrc_libs=LIBS_MACHINE LIBS_SYSTEM
|
1993-03-25 03:51:20 +00:00
|
|
|
@configure@ c_switch_system=C_SWITCH_SYSTEM
|
1993-05-28 04:33:12 +00:00
|
|
|
|
|
|
|
#ifdef UNEXEC
|
|
|
|
@configure@ unexec=UNEXEC
|
|
|
|
#else
|
|
|
|
@configure@ unexec=unexec.o
|
|
|
|
#endif
|
|
|
|
|
1992-04-19 08:53:55 +00:00
|
|
|
#ifdef SYSTEM_MALLOC
|
1992-04-24 08:11:54 +00:00
|
|
|
@configure@ system_malloc=yes
|
1992-04-19 08:53:55 +00:00
|
|
|
#else
|
1992-04-24 08:11:54 +00:00
|
|
|
@configure@ system_malloc=no
|
1992-04-19 08:53:55 +00:00
|
|
|
#endif
|
1993-06-08 07:18:40 +00:00
|
|
|
|
|
|
|
#ifndef C_DEBUG_SWITCH
|
|
|
|
#define C_DEBUG_SWITCH -g
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifndef C_OPTIMIZE_SWITCH
|
|
|
|
#define C_OPTIMIZE_SWITCH -O
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifdef __GNUC__
|
|
|
|
@configure@ CFLAGS=C_DEBUG_SWITCH C_OPTIMIZE_SWITCH
|
|
|
|
#else
|
|
|
|
@configure@ CFLAGS=C_DEBUG_SWITCH
|
|
|
|
#endif
|
1992-04-19 08:53:55 +00:00
|
|
|
' > ${tempcname}
|
1993-05-04 14:17:07 +00:00
|
|
|
# The value of CPP is a quoted variable reference, so we need to do this
|
|
|
|
# to get its actual value...
|
|
|
|
foo=`eval "echo $CPP"`
|
|
|
|
eval `${foo} ${tempcname} \
|
1993-03-23 06:54:37 +00:00
|
|
|
| grep '@configure@' \
|
|
|
|
| sed -e 's/^@configure@ \([^=]*=\)\(.*\)$/\1"\2"/'`
|
1992-04-19 08:53:55 +00:00
|
|
|
rm ${tempcname}
|
|
|
|
|
1993-05-28 04:33:12 +00:00
|
|
|
### Compute the unexec source name from the object name.
|
|
|
|
UNEXEC_SRC="`echo ${unexec} | sed 's/\.o/.c/'`"
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
|
1992-04-19 08:53:55 +00:00
|
|
|
# Do the opsystem or machine files prohibit the use of the GNU malloc?
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
# Assume not, until told otherwise.
|
|
|
|
GNU_MALLOC=yes
|
1992-04-19 08:53:55 +00:00
|
|
|
if [ "${system_malloc}" = "yes" ]; then
|
1992-12-06 22:12:37 +00:00
|
|
|
GNU_MALLOC=no
|
|
|
|
GNU_MALLOC_reason="
|
1992-11-20 17:08:32 +00:00
|
|
|
(The GNU allocators don't work with this system configuration.)"
|
1992-04-19 08:53:55 +00:00
|
|
|
fi
|
|
|
|
|
1993-05-31 21:28:10 +00:00
|
|
|
if [ x"${REL_ALLOC}" = x ]; then
|
1992-12-06 22:12:37 +00:00
|
|
|
REL_ALLOC=${GNU_MALLOC}
|
1992-04-19 08:53:55 +00:00
|
|
|
fi
|
1992-04-01 10:45:51 +00:00
|
|
|
|
1992-12-06 22:12:37 +00:00
|
|
|
LISP_FLOAT_TYPE=yes
|
|
|
|
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
|
|
|
|
#### Find out which version of Emacs this is.
|
|
|
|
version=`grep 'defconst[ ]*emacs-version' ${srcdir}/lisp/version.el \
|
|
|
|
| sed -e 's/^.*"\([0-9][0-9]*\.[0-9][0-9]*\)\..*$/\1/'`
|
1993-05-31 21:28:10 +00:00
|
|
|
if [ x"${version}" = x ]; then
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
echo "${progname}: can't find current emacs version in
|
|
|
|
\`${srcdir}/lisp/version.el'." >&2
|
|
|
|
exit 1
|
|
|
|
fi
|
|
|
|
|
|
|
|
|
The GNU coding standards specify that CFLAGS should be left for
users to set.
* Makefile.in (CFLAGS): Let configure determine the default value
for this. Don't
have it default to DEFS.
(${SUBDIR}): Pass CFLAGS down to submakes, not DEFS.
(lib-src/Makefile, src/Makefile): Edit the default value for
CFLAGS into these files, not DEFS.
* configure.in (CFLAGS): Choose a default value for this - "-g"
normally, or "-g -O" if we're using GCC. Edit it into the
top-level Makefile.
* configure.in: When scanning the machine and system description
#include files, write their names to conftest.c properly.
* configure.in: In configuration name case for Apallos running
Domainios, set opsys, not opsysfile.
* configure.in: Use the autoconf AC_CONFIG_HEADER macro to produce
src/config.h, instead of AC_OUTPUT; the latter overwrites
src/config.h even when it hasn't changed, puts a makefile-style
comment at the top even though it's C code, and produces a
config.status script which doesn't do the job right.
* configure.in: Add AC_LN_S test, so we can tell whether or not we
can use a symbolic link to get the X Menu library into src.
* Makefile.in (LN_S): New variable.
(src/Makefile): Edit the value of LN_S into this makefile.
1993-05-09 22:52:15 +00:00
|
|
|
#### Specify what sort of things we'll be editing into Makefile and config.h.
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
]
|
|
|
|
AC_SUBST(configuration)
|
|
|
|
AC_SUBST(version)
|
|
|
|
AC_SUBST(srcdir)
|
|
|
|
AC_SUBST(c_switch_system)
|
|
|
|
AC_SUBST(libsrc_libs)
|
1993-05-03 02:07:41 +00:00
|
|
|
AC_SUBST(rip_paths)
|
|
|
|
AC_SUBST(inst_paths)
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
AC_SUBST(LD_SWITCH_X_SITE)
|
|
|
|
AC_SUBST(C_SWITCH_X_SITE)
|
The GNU coding standards specify that CFLAGS should be left for
users to set.
* Makefile.in (CFLAGS): Let configure determine the default value
for this. Don't
have it default to DEFS.
(${SUBDIR}): Pass CFLAGS down to submakes, not DEFS.
(lib-src/Makefile, src/Makefile): Edit the default value for
CFLAGS into these files, not DEFS.
* configure.in (CFLAGS): Choose a default value for this - "-g"
normally, or "-g -O" if we're using GCC. Edit it into the
top-level Makefile.
* configure.in: When scanning the machine and system description
#include files, write their names to conftest.c properly.
* configure.in: In configuration name case for Apallos running
Domainios, set opsys, not opsysfile.
* configure.in: Use the autoconf AC_CONFIG_HEADER macro to produce
src/config.h, instead of AC_OUTPUT; the latter overwrites
src/config.h even when it hasn't changed, puts a makefile-style
comment at the top even though it's C code, and produces a
config.status script which doesn't do the job right.
* configure.in: Add AC_LN_S test, so we can tell whether or not we
can use a symbolic link to get the X Menu library into src.
* Makefile.in (LN_S): New variable.
(src/Makefile): Edit the value of LN_S into this makefile.
1993-05-09 22:52:15 +00:00
|
|
|
AC_SUBST(CFLAGS)
|
1993-05-23 04:42:38 +00:00
|
|
|
AC_SUBST(prefix)
|
1993-05-24 06:08:13 +00:00
|
|
|
AC_SUBST(exec_prefix)
|
The GNU coding standards specify that CFLAGS should be left for
users to set.
* Makefile.in (CFLAGS): Let configure determine the default value
for this. Don't
have it default to DEFS.
(${SUBDIR}): Pass CFLAGS down to submakes, not DEFS.
(lib-src/Makefile, src/Makefile): Edit the default value for
CFLAGS into these files, not DEFS.
* configure.in (CFLAGS): Choose a default value for this - "-g"
normally, or "-g -O" if we're using GCC. Edit it into the
top-level Makefile.
* configure.in: When scanning the machine and system description
#include files, write their names to conftest.c properly.
* configure.in: In configuration name case for Apallos running
Domainios, set opsys, not opsysfile.
* configure.in: Use the autoconf AC_CONFIG_HEADER macro to produce
src/config.h, instead of AC_OUTPUT; the latter overwrites
src/config.h even when it hasn't changed, puts a makefile-style
comment at the top even though it's C code, and produces a
config.status script which doesn't do the job right.
* configure.in: Add AC_LN_S test, so we can tell whether or not we
can use a symbolic link to get the X Menu library into src.
* Makefile.in (LN_S): New variable.
(src/Makefile): Edit the value of LN_S into this makefile.
1993-05-09 22:52:15 +00:00
|
|
|
|
1993-05-21 21:33:34 +00:00
|
|
|
AC_DEFINE_UNQUOTED(config_machfile, "\"${machfile}\"")
|
|
|
|
AC_DEFINE_UNQUOTED(config_opsysfile, "\"${opsysfile}\"")
|
|
|
|
AC_DEFINE_UNQUOTED(LD_SWITCH_X_SITE, ${LD_SWITCH_X_SITE})
|
|
|
|
AC_DEFINE_UNQUOTED(C_SWITCH_X_SITE, ${C_SWITCH_X_SITE})
|
1993-05-28 04:33:12 +00:00
|
|
|
AC_DEFINE_UNQUOTED(UNEXEC_SRC, ${UNEXEC_SRC})
|
1993-05-21 21:33:34 +00:00
|
|
|
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
[
|
|
|
|
if [ "${HAVE_X_WINDOWS}" = "yes" ] ; then
|
|
|
|
] AC_DEFINE(HAVE_X_WINDOWS) [
|
|
|
|
fi
|
|
|
|
if [ "${HAVE_X11}" = "yes" ] ; then
|
|
|
|
] AC_DEFINE(HAVE_X11) [
|
|
|
|
fi
|
|
|
|
if [ "${HAVE_XFREE386}" = "yes" ] ; then
|
|
|
|
] AC_DEFINE(HAVE_XFREE386) [
|
|
|
|
fi
|
|
|
|
if [ "${HAVE_X_MENU}" = "yes" ] ; then
|
|
|
|
] AC_DEFINE(HAVE_X_MENU) [
|
|
|
|
fi
|
|
|
|
if [ "${GNU_MALLOC}" = "yes" ] ; then
|
|
|
|
] AC_DEFINE(GNU_MALLOC) [
|
|
|
|
fi
|
|
|
|
if [ "${REL_ALLOC}" = "yes" ] ; then
|
|
|
|
] AC_DEFINE(REL_ALLOC) [
|
|
|
|
fi
|
|
|
|
if [ "${LISP_FLOAT_TYPE}" = "yes" ] ; then
|
|
|
|
] AC_DEFINE(LISP_FLOAT_TYPE) [
|
|
|
|
fi
|
1992-04-01 10:45:51 +00:00
|
|
|
|
1992-10-27 09:07:23 +00:00
|
|
|
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
#### Report on what we decided to do.
|
|
|
|
echo "
|
1992-10-27 09:07:23 +00:00
|
|
|
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
Configured for \`${configuration}'.
|
1992-11-20 17:08:32 +00:00
|
|
|
|
1993-05-28 04:52:55 +00:00
|
|
|
Where should the build process find the source code? ${srcdir}
|
1992-11-20 17:08:32 +00:00
|
|
|
What operating system and machine description files should Emacs use?
|
|
|
|
\`${opsysfile}' and \`${machfile}'
|
The GNU coding standards specify that CFLAGS should be left for
users to set.
* Makefile.in (CFLAGS): Let configure determine the default value
for this. Don't
have it default to DEFS.
(${SUBDIR}): Pass CFLAGS down to submakes, not DEFS.
(lib-src/Makefile, src/Makefile): Edit the default value for
CFLAGS into these files, not DEFS.
* configure.in (CFLAGS): Choose a default value for this - "-g"
normally, or "-g -O" if we're using GCC. Edit it into the
top-level Makefile.
* configure.in: When scanning the machine and system description
#include files, write their names to conftest.c properly.
* configure.in: In configuration name case for Apallos running
Domainios, set opsys, not opsysfile.
* configure.in: Use the autoconf AC_CONFIG_HEADER macro to produce
src/config.h, instead of AC_OUTPUT; the latter overwrites
src/config.h even when it hasn't changed, puts a makefile-style
comment at the top even though it's C code, and produces a
config.status script which doesn't do the job right.
* configure.in: Add AC_LN_S test, so we can tell whether or not we
can use a symbolic link to get the X Menu library into src.
* Makefile.in (LN_S): New variable.
(src/Makefile): Edit the value of LN_S into this makefile.
1993-05-09 22:52:15 +00:00
|
|
|
What compiler should emacs be built with? ${CC} ${CFLAGS}
|
1992-12-06 22:12:37 +00:00
|
|
|
Should Emacs use the GNU version of malloc? ${GNU_MALLOC}${GNU_MALLOC_reason}
|
|
|
|
Should Emacs use the relocating allocator for buffers? ${REL_ALLOC}
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
What window system should Emacs use? ${window_system}${x_includes+
|
1993-03-23 05:31:35 +00:00
|
|
|
Where do we find X Windows header files? }${x_includes}${x_libraries+
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
Where do we find X Windows libraries? }${x_libraries}
|
1992-04-04 05:55:39 +00:00
|
|
|
|
* configure: Don't make the top-level Makefile read-only - people
may want to edit the values of the path variables.
Make path specification conform to GNU coding standards.
* configure (long_usage): Remove all traces of old arguments from
usage messages, and document the options we do accept in more
detail: -with-x... and --srcdir.
(options, boolean_opts): Deleted; we don't have enough options to
make this worthwhile.
(prefix, bindir, lisppath, datadir, libdir, lockdir): Deleted,
along with the code which supported them; these should be set as
arguments to the top-level make.
(config_h_opts): Since this no longer doubles as a list of option
names, make them upper case; this simplifies the code which uses
them to build the sed command to edit src/config.h. Change the
code which sets them.
(cc, g, O): Don't allow the user to set these using options; they
should be specified using `CC=' and `CFLAGS=' arguments to the
top-level make. Just choose reasonable default values for them,
and edit them into Makefile.in's default CC and CONFIG_CFLAGS
values.
(gnu_malloc, rel_alloc): Don't allow the user to set these using
options; use them whenever the configuration files say they're
possible.
Simplify the argument processing loop. Don't accept abbreviations
for option names; these might conflict with other configuration
options in the future.
Add some support for the `--srcdir' option. Check for the sources
in . and .. if `--srcdir' is omitted. If the directories we will
compile in don't exist yet, create them under the current directory.
Note that the rest of the build process doesn't really support
this.
Edit only the top Makefile. That should edit the others. Edit
into the makefile: `version', from lisp/version.el, `configname'
and `srcdir' from the configuration arguments, `CC' and
`CONFIG_CFLAGS' as guessed from the presence or absence of GCC in
the user's path, and LOADLIBES as gleaned from the system
description files.
Simplify the report generated; it doesn't need to include any
description of paths now.
Make `config.status' exec configure instead of just calling it, so
there's no harm in overwriting `config.status'.
* Makefile.in (version, configname): New variables, used to choose
the default values for datadir and libdir.
Path variables rearranged into two clearer groups:
- In the first group are the variables specified by the GNU coding
standards (prefix, bindir, datadir, statedir, libdir, mandir,
manext, infodir, and srcdir).
- In the second are the variables actually used for Emacs's paths
(lispdir, locallisppath, lisppath, buildlisppath, etcdir, lockdir,
archlibdir), which depend on the first category.
datadir and libdir default to directories under
${prefix}/lib/emacs instead of ${prefix}/emacs, by popular
demand.
etcdir and lispdir default to subdirectories of datadir.
archlibdir defaults to libdir.
The new installation tree is a bit deeper than it used to be, so
use the new make-path program in lib-src to build them all.
Always build a new src/paths.h.tmp and then move-if-change it to
src/paths.h, to avoid unnecessary rebuilds while responding to the
right changes.
Remove all mention of arch-lib. Run utility commands from
lib-src, and let the executables be copied into archlibdir when
Emacs is installed.
Add targets for src/Makefile, lib-src/Makefile, and
oldXMenu/Makefile, editing the values of the path variables into
them.
Let lib-src do its own installation.
(datadir): Default to putting data files under
${prefix}/lib/emacs/${version}, not /usr/local/emacs.
(emacsdir): Variable deleted; it would only be confusing to use.
(lispdir, etcdir): Default to ${datadir}/lisp.
(mkdir): Use make-path for this.
(lockdir): Do this in mkdir.
(Makefile): New target.
* configure (usage_message): Rename to long_usage.
1992-12-12 15:42:53 +00:00
|
|
|
"
|
* configure.in: When checking for X windows, search for an X11
subdirectory of ${x_includes}.
* configure.in: Check for gettimeofday function, for getdate.y.
Change `configure' to a mixture of custom code and autoconf stuff.
autoconf can't derive all the information we need, but we'd really
like to be able to take advantage of some of its tests, and its
file-editing facilities.
* configure.in: Renamed from configure.
Quote the sections of shell script we want copied literally to
the configure script.
(compile): Initialize this to make the autoconf macros' code happy.
Use AC_PROG_CC, AC_CONST, and AC_RETSIGTYPE instead of writing out
code to do their jobs.
Use autoconf to produce Makefile and src/config.h.
Remove the Makefile-style comment that autoconf places at the top
of src/config.h.
(config_h_opts): Removed - no longer necessary.
* Makefile.in (configname): Renamed to configuration.
(CONFIG_CFLAGS): Renamed to DEFS.
(CC, DEFS, C_SWITCH_SYSTEM, version, configuration): Adjusted to
get values via autoload @cookies@.
(libsrc_libs): Get this from autoconf. We used to do nothing
about this.
(${SUBDIR}): Pass DEFS to submakes instead of CONFIG_CFLAGS.
1993-04-10 06:02:36 +00:00
|
|
|
]
|
The GNU coding standards specify that CFLAGS should be left for
users to set.
* Makefile.in (CFLAGS): Let configure determine the default value
for this. Don't
have it default to DEFS.
(${SUBDIR}): Pass CFLAGS down to submakes, not DEFS.
(lib-src/Makefile, src/Makefile): Edit the default value for
CFLAGS into these files, not DEFS.
* configure.in (CFLAGS): Choose a default value for this - "-g"
normally, or "-g -O" if we're using GCC. Edit it into the
top-level Makefile.
* configure.in: When scanning the machine and system description
#include files, write their names to conftest.c properly.
* configure.in: In configuration name case for Apallos running
Domainios, set opsys, not opsysfile.
* configure.in: Use the autoconf AC_CONFIG_HEADER macro to produce
src/config.h, instead of AC_OUTPUT; the latter overwrites
src/config.h even when it hasn't changed, puts a makefile-style
comment at the top even though it's C code, and produces a
config.status script which doesn't do the job right.
* configure.in: Add AC_LN_S test, so we can tell whether or not we
can use a symbolic link to get the X Menu library into src.
* Makefile.in (LN_S): New variable.
(src/Makefile): Edit the value of LN_S into this makefile.
1993-05-09 22:52:15 +00:00
|
|
|
AC_OUTPUT(Makefile)
|