[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Still cannot build native windows version of emacs 23.*
From: |
Eric Lilja |
Subject: |
Re: Still cannot build native windows version of emacs 23.* |
Date: |
Tue, 05 Feb 2008 10:15:11 +0100 |
User-agent: |
Thunderbird 2.0.0.9 (Windows/20071031) |
Eli Zaretskii skrev:
From: Eric Lilja <address@hidden>
Date: Mon, 04 Feb 2008 22:28:31 +0100
Can you manually modify admin/unidate/makefile so that Emacs when run
displays the value of load-path? Then show it here.
Well, the only working emacs I have right now is GNU Emacs 22.1.50.1
(i386-mingw-nt5.1.2600) of 2007-10-28 on MINDCOOLER. My stupid script
proceeded to wipe out my working emacs-23 binaries thinking that the
build went okay. Can I use this and what exactly do I need to do? Sorry
for asking dumb questions.
No need to apologize, I didn't explain enough.
What I meant is this: modify the line in admin/unidate/makefile that
says
"../../src/oo-spd/i386/emacs.exe" -Q --multibyte -batch -f
batch-byte-compile unidata-gen.el
to say this instead:
"../../src/oo-spd/i386/emacs.exe" -Q --multibyte -batch --eval "(message
\"%s\" load-path)"
and then see what it displays when you run "make bootstrap".
Note that, since admin/unidate/makefile is generated by "make
bootstrap", you will either need to edit admin/unidate/makefile.w32-in
and then run nt/configure.bat to create admin/unidate/makefile, or
edit admin/unidate/makefile while "make bootstrap" runs, but before it
gets to where it executes the above line.
OK, here's what happened. My admin/unidata/makefile was generated after
the configure step, but I couldn't find the line I was supposed to
change so I read your message a little bit more carefully. Ah, and you
said that it's generated during bootstrap so I thought that it would be
modified during bootstrap and when that had happened I was going to
alter the line you wanted me to alter (before it was used, hopefully).
But the admin/unidata/makefile was never altered after it was initially
created during the bootstrap phase. I'm attaching it (maybe useless but
oh well).
# Start of settings from configure.bat
COMPILER=gcc
MCPU_FLAG=-mtune=pentium4
NODEBUG=1
NOCYGWIN=1
USER_CFLAGS=-I../../xpm-3.5.1-1/include
USER_LDFLAGS=-L../../xpm-3.5.1-1/lib
# End of settings from configure.bat
# -*- Makefile -*- definition file for building GNU Emacs on Windows NT.
# Copyright (C) 2000, 2001, 2002, 2003, 2004,
# 2005, 2006, 2007, 2008 Free Software Foundation, Inc.
#
# 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 3, 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, Inc., 51 Franklin Street, Fifth Floor,
# Boston, MA 02110-1301, USA.
# Ensure 'all' is the default target
all:
# NOTES
#
# I tried to force gmake to use the native shell for simplicity, by
# setting SHELL as below, but this didn't work reliably because of
# various case sensitivity niggles. Specifically, COMSPEC (which is in
# fact usually spelled ComSpec on NT, to make life difficult) typically
# references "cmd.exe" (on NT) when the file is actually called
# "CMD.EXE" on disk for hysterical raisons. As a result, GNU make
# thinks it doesn't exist (unless compiled with a switch to ignore
# case), and so doesn't change which shell it will invoke to execute
# commands.
#
# It would be possible, though very tedious using just gmake facilities,
# to convert the COMSPEC value to uppercase to solve this problem, but
# it isn't worth it. That is partly because, even when using the native
# shell, gmake tends to not be happy with backslashes in command
# strings. The obvious solution is to use forward slashes as much as
# possible, which can be made to work most of the time (putting
# filenames in quotes often helps), but there are still some internal
# cmd.exe commands like `del' and `copy' that won't work with them.
# Although it is possible to convert slashes to backslashes when
# necessary, gmake requires explicitly calling its subst function, which
# nmake does not understand). For this reason, it is simplest to
# mandate that rm and cp be available, so we can use Unix-format file
# names everywhere. (Fortunately both MS and GNU make, and the
# respective compilers, are happy with Unix-format names.)
#
# Since we cannot easily force the choice of a particular shell, we must
# make the effort to cope with whichever shell is being used.
# Fortunately, the only command we need to use that is shell specific is
# the testing of a file's existence for the purpose of working out when
# we are copying files to their original location. That particular
# requirement is abstracted easily enough.
#
# The only other problem area was the change of directory when running
# temacs to dump emacs.exe (where gmake doesn't support cd foo in any
# useful way), but that has been resolved by modifying the Windows
# unexec function slightly to not require the directory change while
# still allowing objects and binaries to be in subdirectories.
# This doesn't work.
#SHELL:=$(COMSPEC)
# Determine whether make is using sh or cmd/command as shell; cmd.exe
# will output "ECHO is on" when echo is given by itself, while sh will
# not produce any output.
sh_output := $(shell echo)
ifeq "$(findstring ECHO, $(sh_output))" "ECHO"
THE_SHELL = $(COMSPEC)$(ComSpec)
SHELLTYPE=CMD
else
USING_SH = 1
THE_SHELL = $(SHELL)
SHELLTYPE=SH
endif
MAKETYPE=gmake
# The following "ifeq" does not appear to DTRT, and therefore breaks
# the build on mingw32. Also the -m option does not exist in many
# (reasonably recent even) versions of Cygwin. These issues need to be
# remedied before putting this cygpath kludge back in.
# Convert CURDIR to native file name, if in Cygwin format
#ifeq "$(shell cygpath $(CURDIR))" "$(CURDIR)"
#CURDIR := $(shell cygpath -m $(CURDIR))
#endif
THISDIR = .
# Cygwin has changed quoting rules somewhat since b20, in a way that
# affects makefiles using sh as the command processor, so we need to
# detect which rules to use.
ifdef USING_SH
sh_output := $(shell echo [Please ignore a syntax error on the next line - it
is intentional] 1>&2)
sh_output := $(shell echo foo")
ifeq "$(sh_output)" ""
NEW_CYGWIN = 1
endif
# By default, newer versions of Cygwin mess with NTFS ACLs in an
# attempt to emulate traditional posix file permissions. This can
# cause bad effects, such as .exe files that are missing the
# FILE_EXECUTE/FILE_GENERIC_EXECUTE permissions when they are created
# with Cygwin commands that don't expect to be creating executable
# files. Then when we later use a non-Cygwin program to create the
# real .exe, the previous Cygwin defined ACL sticks.
CYGWIN=nontsec
export CYGWIN
endif
ALL_DEPS = $^
EMPTY =
SPACE = $(EMPTY) $(EMPTY)
SUBSYSTEM_WINDOWS=-Wl,-subsystem,windows
SUBSYSTEM_CONSOLE=-Wl,-subsystem,console
# INSTALL_DIR is the directory into which emacs will be installed.
#
ifndef INSTALL_DIR
INSTALL_DIR = $(CURDIR)/..
endif
export EMACSLOADPATH
# Determine the architecture we're running on.
# Define ARCH for our purposes;
# Define CPU for use by ntwin32.mak;
# Define CONFIG_H to the appropriate config.h for the system;
#
ifdef PROCESSOR_ARCHITECTURE
# We're on Windows NT
CPU = $(PROCESSOR_ARCHITECTURE)
CONFIG_H = config.nt
OS_TYPE = windowsnt
ifeq "$(PROCESSOR_ARCHITECTURE)" "x86"
ARCH = i386
CPU = i386
else
ifeq "$(PROCESSOR_ARCHITECTURE)" "MIPS"
ARCH = mips
else
ifeq "$(PROCESSOR_ARCHITECTURE)" "ALPHA"
ARCH = alpha
else
ifeq "$(PROCESSOR_ARCHITECTURE)" "PPC"
ARCH = ppc
else
error Unknown architecture type "$(PROCESSOR_ARCHITECTURE)"
endif
endif
endif
endif
else
# We're on Windows 95
ARCH = i386
CPU = i386
CONFIG_H = config.nt
OS_TYPE = windows95
endif
AR = ar -rsc
AR_OUT =
CC = gcc
CC_OUT = -o$(SPACE)
LINK = gcc
LINK_OUT = -o$(SPACE)
RC = windres -O coff
RC_OUT = -o$(SPACE)
RC_INCLUDE = --include-dir$(SPACE)
libc =
baselibs =
O = o
A = a
BASE_LIBS = $(libc) $(baselibs)
ADVAPI32 = -ladvapi32
COMCTL32 = -lcomctl32
COMDLG32 = -lcomdlg32
GDI32 = -lgdi32
MPR = -lmpr
SHELL32 = -lshell32
USER32 = -luser32
WSOCK32 = -lwsock32
WINMM = -lwinmm
WINSPOOL = -lwinspool
OLE32 = -lole32
ifdef NOOPT
DEBUG_CFLAGS = -DEMACSDEBUG
else
DEBUG_CFLAGS =
endif
CFLAGS = -I. -DWIN32_LEAN_AND_MEAN -D_WIN32_WINNT=0x0400
$(ARCH_CFLAGS) -D$(ARCH) \
-D_CRTAPI1=_cdecl \
$(DEBUG_CFLAGS) $(USER_CFLAGS) $(LOCAL_FLAGS)
EMACS_EXTRA_C_FLAGS = -DUSE_CRT_DLL=1
# see comments in allocate_heap in w32heap.c before changing any of the
# -stack, -heap, or -image-base settings.
TEMACS_EXTRA_LINK = -Wl,-stack,0x00800000 -Wl,-heap,0x00100000
-Wl,-image-base,0x01000000 $(SUBSYSTEM_CONSOLE) -Wl,-entry,__start
-Wl,-Map,$(BLD)/temacs.map
ifdef NOOPT
OBJDIR = oo
else
OBJDIR = oo-spd
endif
$(OBJDIR):; -mkdir "$(OBJDIR)"
BLD = $(OBJDIR)/$(ARCH)
stamp_BLD: $(OBJDIR)
-mkdir "$(BLD)"
echo $(BLD) > $@
COMPILER_TEMP_FILES =
CP = cp -f
CP_DIR = cp -rf
DEL = rm
DEL_TREE = rm -r
ifdef USING_SH
IFNOTSAMEDIR = if [ ! -s ../same-dir.tst ] ; then
FOREACH = for f in
FORVAR = $${f}
FORDO = ; do
ENDFOR = ; done
ENDIF = ; fi
ARGQUOTE = '
ifdef NEW_CYGWIN
DQUOTE = "
else
DQUOTE = ""
endif
else
IFNOTSAMEDIR = if not exist ../same-dir.tst
FOREACH = for %%f in (
FORVAR = %%f
FORDO = ) do
ENDFOR =
ENDIF =
ARGQUOTE = "
DQUOTE = \"
endif
ifdef NODEBUG
DEBUG_FLAG =
DEBUG_LINK =
else
DEBUG_FLAG = -gstabs+ -g3
DEBUG_LINK = -gstabs+ -g3
endif
ifdef NOCYGWIN
NOCYGWIN = -mno-cygwin
endif
ifeq "$(ARCH)" "i386"
ifdef NOOPT
ARCH_CFLAGS = -D_X86_=1 -c $(DEBUG_FLAG) $(NOCYGWIN)
else
ARCH_CFLAGS = -D_X86_=1 -c $(DEBUG_FLAG) $(NOCYGWIN) $(MCPU_FLAG) -O2 \
# -fbuiltin \
# -finline-functions \
# -fomit-frame-pointer
endif
ARCH_LDFLAGS = $(SYS_LDFLAGS)
else
ERROR Unknown architecture type "$(ARCH)".
endif
LINK_FLAGS = $(ARCH_LDFLAGS) $(DEBUG_LINK) $(NOCYGWIN) $(USER_LDFLAGS)
export XMFLAGS
.DEFAULT:
$(BLD)/%.o: %.c
$(CC) $(CFLAGS) $(CC_OUT)$@ $<
# arch-tag: 35eb9662-8534-4bcf-b891-0730a09d657f
# makefile.w32-in -- W32 -*- Makefile -*- to generate character property tables.
# Copyright (C) 2005, 2006, 2007, 2008
# National Institute of Advanced Industrial Science and Technology (AIST)
# Registration Number H13PRO009
#
# This file is part of GNU Emacs.
# GNU Emacs is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3, or (at your option)
# any later version.
# GNU Emacs is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
# You should have received a copy of the GNU General Public License
# along with GNU Emacs; see the file COPYING. If not, write to the
# Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
# Boston, MA 02110-1301, USA.
EMACS = ../../src/$(BLD)/emacs.exe
DSTDIR = ../../lisp/international
EMACSLOADPATH = $(CURDIR)/../../lisp
# Quote EMACS so it could be a file name with embedded whitespace
RUNEMACS = "$(EMACS)" -Q --multibyte -batch
all: $(DSTDIR)/charprop.el
.el.elc:
$(RUNEMACS) -f batch-byte-compile $<
unidata.txt: UnicodeData.txt
sed -e \
$(ARGQUOTE)s/\([^;]*\);\(.*\)/(#x\1 $(DQUOTE)\2$(DQUOTE))/$(ARGQUOTE)\
-e $(ARGQUOTE)s/;/$(DQUOTE) $(DQUOTE)/g$(ARGQUOTE) < $< > $@
charprop-SH: unidata-gen.elc unidata.txt
ELC=$(CURDIR)/unidata-gen.elc; \
DATA=$(CURDIR)/unidata.txt; \
cd $(DSTDIR); \
$(RUNEMACS) --load $${ELC} -f unidata-gen-files $${DATA}
charprop-CMD: unidata-gen.elc unidata.txt
$(RUNEMACS) --eval $(ARGQUOTE)(cd
$(DQUOTE)$(DSTDIR)$(DQUOTE))$(ARGQUOTE) --load $(CURDIR)/unidata-gen.elc -f
unidata-gen-files $(CURDIR)/unidata.txt
${DSTDIR}/charprop.el: charprop-$(SHELLTYPE)
clean:
- $(DEL) unidata-gen.elc unidata.txt
- Re: Still cannot build native windows version of emacs 23.*, (continued)
- Re: Still cannot build native windows version of emacs 23.*, Eric Lilja, 2008/02/04
- Re: Still cannot build native windows version of emacs 23.*, Jason Rumney, 2008/02/04
- Re: Still cannot build native windows version of emacs 23.*, Eric Lilja, 2008/02/04
- Re: Still cannot build native windows version of emacs 23.*, Juanma Barranquero, 2008/02/04
- Re: Still cannot build native windows version of emacs 23.*, Eli Zaretskii, 2008/02/04
- Re: Still cannot build native windows version of emacs 23.*, Eli Zaretskii, 2008/02/04
- Re: Still cannot build native windows version of emacs 23.*, Eric Lilja, 2008/02/04
- Re: Still cannot build native windows version of emacs 23.*, Eli Zaretskii, 2008/02/04
- Re: Still cannot build native windows version of emacs 23.*, Eric Lilja, 2008/02/04
- Re: Still cannot build native windows version of emacs 23.*, Eli Zaretskii, 2008/02/04
- Re: Still cannot build native windows version of emacs 23.*,
Eric Lilja <=
- Re: Still cannot build native windows version of emacs 23.*, Eli Zaretskii, 2008/02/05
- Re: Still cannot build native windows version of emacs 23.*, Eric Lilja, 2008/02/05
- Re: Still cannot build native windows version of emacs 23.*, Eli Zaretskii, 2008/02/05
- Re: Still cannot build native windows version of emacs 23.*, Eric Lilja, 2008/02/05
- Re: Still cannot build native windows version of emacs 23.*, Eli Zaretskii, 2008/02/06
- Re: Still cannot build native windows version of emacs 23.*, Eric Lilja, 2008/02/06
- Re: Still cannot build native windows version of emacs 23.*, Eli Zaretskii, 2008/02/06
- Re: Still cannot build native windows version of emacs 23.*, Eric Lilja, 2008/02/06
- Re: Still cannot build native windows version of emacs 23.*, Eric Lilja, 2008/02/07
- Re: Still cannot build native windows version of emacs 23.*, Eli Zaretskii, 2008/02/08