GNU bug report logs - #14603
MS Windows configure needs to be run via msysconfig.sh

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 14603 in the body.
You can then email your comments to 14603 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Dmitry Gutov <dgutov <at> yandex.ru>
To: bug-gnu-emacs <at> gnu.org
Subject: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 01:53:08 +0400

I'm trying to bootstrap the current Emacs trunk in the environment that
worked fine not too long ago, and I'm getting this error:

...
gcc  -std=gnu99 -DHAVE_CONFIG_H -I. -I/c/Users/gutov/vc/emacs-bzr/trunk/lib -I../src -I/c/Users/gutov/vc/emacs-bzr/trunk/src    -mtune=pentium4   -DUSE_CRT_DLL=1 -I /c/Users/gutov/vc/emacs-bzr/trunk/nt/inc    -g3 -O2 -gdwarf-2 -MT u64.o -MD -MP -MF .deps/u64.Tpo -c -o u64.o u64.c
mv -f .deps/u64.Tpo .deps/u64.Po
gcc  -std=gnu99 -DHAVE_CONFIG_H -I. -I/c/Users/gutov/vc/emacs-bzr/trunk/lib -I../src -I/c/Users/gutov/vc/emacs-bzr/trunk/src    -mtune=pentium4   -DUSE_CRT_DLL=1 -I /c/Users/gutov/vc/emacs-bzr/trunk/nt/inc    -g3 -O2 -gdwarf-2 -MT utimens.o -MD -MP -MF .deps/utimens.Tpo -c -o utimens.o utimens.c
mv -f .deps/utimens.Tpo .deps/utimens.Po
gcc  -std=gnu99 -DHAVE_CONFIG_H -I. -I/c/Users/gutov/vc/emacs-bzr/trunk/lib -I../src -I/c/Users/gutov/vc/emacs-bzr/trunk/src    -mtune=pentium4   -DUSE_CRT_DLL=1 -I /c/Users/gutov/vc/emacs-bzr/trunk/nt/inc    -g3 -O2 -gdwarf-2 -MT dup2.o -MD -MP -MF .deps/dup2.Tpo -c -o dup2.o dup2.c
dup2.c:35:0: warning: "WIN32_LEAN_AND_MEAN" redefined [enabled by default]
In file included from ../src/conf_post.h:32:0,
                 from ../src/config.h:1840,
                 from dup2.c:20:
c:/Users/gutov/vc/emacs-bzr/trunk/nt/inc/ms-w32.h:145:0: note: this is the location of the previous definition
dup2.c:38:26: fatal error: msvc-inval.h: No such file or directory
compilation terminated.
Makefile:1428: recipe for target `dup2.o' failed
make[3]: *** [dup2.o] Error 1
make[3]: Leaving directory `/c/Users/gutov/vc/emacs-bzr/trunk/lib'
Makefile:1312: recipe for target `all' failed
make[2]: *** [all] Error 2
make[2]: Leaving directory `/c/Users/gutov/vc/emacs-bzr/trunk/lib'
Makefile:367: recipe for target `lib' failed
make[1]: *** [lib] Error 2
make[1]: Leaving directory `/c/Users/gutov/vc/emacs-bzr/trunk'
Makefile:1002: recipe for target `bootstrap' failed
make: *** [bootstrap] Error 2

Message #8 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Dmitry Gutov <dgutov <at> yandex.ru>
Cc: 14603 <at> debbugs.gnu.org
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 05:43:02 +0300

> From: Dmitry Gutov <dgutov <at> yandex.ru>
> Date: Thu, 13 Jun 2013 01:53:08 +0400
> 
> I'm trying to bootstrap the current Emacs trunk in the environment that
> worked fine not too long ago, and I'm getting this error:
> 
> ...
> gcc  -std=gnu99 -DHAVE_CONFIG_H -I. -I/c/Users/gutov/vc/emacs-bzr/trunk/lib -I../src -I/c/Users/gutov/vc/emacs-bzr/trunk/src    -mtune=pentium4   -DUSE_CRT_DLL=1 -I /c/Users/gutov/vc/emacs-bzr/trunk/nt/inc    -g3 -O2 -gdwarf-2 -MT u64.o -MD -MP -MF .deps/u64.Tpo -c -o u64.o u64.c
> mv -f .deps/u64.Tpo .deps/u64.Po
> gcc  -std=gnu99 -DHAVE_CONFIG_H -I. -I/c/Users/gutov/vc/emacs-bzr/trunk/lib -I../src -I/c/Users/gutov/vc/emacs-bzr/trunk/src    -mtune=pentium4   -DUSE_CRT_DLL=1 -I /c/Users/gutov/vc/emacs-bzr/trunk/nt/inc    -g3 -O2 -gdwarf-2 -MT utimens.o -MD -MP -MF .deps/utimens.Tpo -c -o utimens.o utimens.c
> mv -f .deps/utimens.Tpo .deps/utimens.Po
> gcc  -std=gnu99 -DHAVE_CONFIG_H -I. -I/c/Users/gutov/vc/emacs-bzr/trunk/lib -I../src -I/c/Users/gutov/vc/emacs-bzr/trunk/src    -mtune=pentium4   -DUSE_CRT_DLL=1 -I /c/Users/gutov/vc/emacs-bzr/trunk/nt/inc    -g3 -O2 -gdwarf-2 -MT dup2.o -MD -MP -MF .deps/dup2.Tpo -c -o dup2.o dup2.c
> dup2.c:35:0: warning: "WIN32_LEAN_AND_MEAN" redefined [enabled by default]
> In file included from ../src/conf_post.h:32:0,
>                  from ../src/config.h:1840,
>                  from dup2.c:20:
> c:/Users/gutov/vc/emacs-bzr/trunk/nt/inc/ms-w32.h:145:0: note: this is the location of the previous definition
> dup2.c:38:26: fatal error: msvc-inval.h: No such file or directory
> compilation terminated.
> Makefile:1428: recipe for target `dup2.o' failed

Something was wrong during the configure step, because lib/dup2.c is
not supposed to be compiled in the MinGW build.  My crystal ball says
that somehow the site file nt/mingw-cfg.site was not read by
configure.

Message #11 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Dmitry Gutov <dgutov <at> yandex.ru>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 14603 <at> debbugs.gnu.org
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 11:24:17 +0400

[Message part 1 (text/plain, inline)]

On 13.06.2013 6:43, Eli Zaretskii wrote:
> Something was wrong during the configure step, because lib/dup2.c is
> not supposed to be compiled in the MinGW build.  My crystal ball says
> that somehow the site file nt/mingw-cfg.site was not read by
> configure.

The file is there, and I don't see any errors during configure. Here are 
the logs for it and the bootstrap.

[bootstrap.log (text/plain, attachment)]

[configure.log (text/plain, attachment)]

Message #14 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Dmitry Gutov <dgutov <at> yandex.ru>
Cc: 14603 <at> debbugs.gnu.org
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 16:17:12 +0300

> Date: Thu, 13 Jun 2013 11:24:17 +0400
> From: Dmitry Gutov <dgutov <at> yandex.ru>
> CC: 14603 <at> debbugs.gnu.org
> 
> On 13.06.2013 6:43, Eli Zaretskii wrote:
> > Something was wrong during the configure step, because lib/dup2.c is
> > not supposed to be compiled in the MinGW build.  My crystal ball says
> > that somehow the site file nt/mingw-cfg.site was not read by
> > configure.
> 
> The file is there

Its being there does not automatically guarantee that it will be read.

> and I don't see any errors during configure.

That's because the errors are subtle ;-)

Here's one error:

> checking for a BSD-compatible install... /bin/install -c

It should have had this line before the above one:

  configure: loading site script ./nt/mingw-cfg.site

Here's another error:

> checking whether dup2 works... no

It should have said this instead:

  checking whether dup2 works... (cached) yes

And another one:

> checking for library containing acl_get_file... no

which should be

  checking for library containing acl_get_file... (cached) none required

All these, and quite a few more, are arranged by nt/mingw-cfg.site.

How (with what command line) did you invoke the configure script?

Message #17 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Dmitry Gutov <dgutov <at> yandex.ru>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 14603 <at> debbugs.gnu.org
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 17:23:28 +0400

On 13.06.2013 17:17, Eli Zaretskii wrote:
> That's because the errors are subtle ;-)

I figured so. :)

> How (with what command line) did you invoke the configure script?

No arguments, just ./configure

Message #20 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Dmitry Gutov <dgutov <at> yandex.ru>
Cc: 14603 <at> debbugs.gnu.org
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 16:44:02 +0300

> Date: Thu, 13 Jun 2013 17:23:28 +0400
> From: Dmitry Gutov <dgutov <at> yandex.ru>
> CC: 14603 <at> debbugs.gnu.org
> 
> > How (with what command line) did you invoke the configure script?
> 
> No arguments, just ./configure

That will never work.  You need to follow the instructions in
nt/INSTALL.MSYS.  The correct invocation is

    ./nt/msysconfig.sh --prefix=PREFIX ...

where "..." stands for any options or arguments you want to pass to
configure.

Message #23 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 14603 <at> debbugs.gnu.org, Dmitry Gutov <dgutov <at> yandex.ru>
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 12:15:00 -0400

Eli Zaretskii wrote:

>> No arguments, just ./configure
>
> That will never work. 

Can you make it detect someone has done that and throw an error?
Naively, I would expect quite a lot of people to do that (but I never
configured anything on MS Windows; maybe it's common knowledge).

Message #26 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Glenn Morris <rgm <at> gnu.org>
Cc: 14603 <at> debbugs.gnu.org, dgutov <at> yandex.ru
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 20:50:50 +0300

> From: Glenn Morris <rgm <at> gnu.org>
> Cc: Dmitry Gutov <dgutov <at> yandex.ru>,  14603 <at> debbugs.gnu.org
> Date: Thu, 13 Jun 2013 12:15:00 -0400
> 
> Eli Zaretskii wrote:
> 
> >> No arguments, just ./configure
> >
> > That will never work. 
> 
> Can you make it detect someone has done that and throw an error?

I don't know what to detect.  Looking at $0 is not going to fit the
bill, because it is quite alright to set CONFIG_SITE manually before
invoking ./configure.  You will see that the top-level Makefile.in
does precisely that, when it finds it needs to re-run configure.
Moreover, someone might even copy mingw-cfg.site to
$prefix/share/config.site, and I don't think they deserve to be
punished by throwing an error.  So testing that "$CONFIG_SITE" is
non-empty is also sub-optimal if not outright wrong.

I wrote detailed instructions in nt/INSTALL.MSYS so that people would
know how to avoid these.  If they don't follow, they will realize that
soon enough...

> Naively, I would expect quite a lot of people to do that

I would expect them to read the instructions first.  Maybe it's me who
is naive.

> maybe it's common knowledge

It isn't.

Message #29 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Dmitry Gutov <dgutov <at> yandex.ru>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 14603 <at> debbugs.gnu.org
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 22:02:48 +0400

On 13.06.2013 17:44, Eli Zaretskii wrote:
>> Date: Thu, 13 Jun 2013 17:23:28 +0400
>> From: Dmitry Gutov <dgutov <at> yandex.ru>
>> CC: 14603 <at> debbugs.gnu.org
>>
>>> How (with what command line) did you invoke the configure script?
>>
>> No arguments, just ./configure
>
> That will never work.  You need to follow the instructions in
> nt/INSTALL.MSYS.  The correct invocation is
>
>      ./nt/msysconfig.sh --prefix=PREFIX ...

So it is. Sorry!

I went through the same mistake the first time I built Emacs with MSYS, 
went to the instruction, got it right.

But now, after a few successful builds this way, and after switching 
from Windows to GNU/Linux and back a few times (I use systems of both 
kinds), on the one hand, I felt confident that I knew how to do it, on 
the other hand, the last few times I built Emacs were on Linux, so I 
completely forgot about "msysconfig.sh".

Like Glenn said, it would be great if ./configure aborted when the 
environment is MSYS (would it be too hard to detect that?), or delegated 
transparently to ./nt/msysconfig.sh.

Message #36 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Jan Djärv <jan.h.d <at> swipnet.se>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: Glenn Morris <rgm <at> gnu.org>, 14603 <at> debbugs.gnu.org, dgutov <at> yandex.ru
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 20:06:00 +0200

Hello.

Can't you just set some environment variable in msysconfig.sh and check for it in configure?
If it isn't set, msysconfig.sh has not been run.

	Jan D.

13 jun 2013 kl. 19:50 skrev Eli Zaretskii <eliz <at> gnu.org>:

>> From: Glenn Morris <rgm <at> gnu.org>
>> Cc: Dmitry Gutov <dgutov <at> yandex.ru>,  14603 <at> debbugs.gnu.org
>> Date: Thu, 13 Jun 2013 12:15:00 -0400
>> 
>> Eli Zaretskii wrote:
>> 
>>>> No arguments, just ./configure
>>> 
>>> That will never work. 
>> 
>> Can you make it detect someone has done that and throw an error?
> 
> I don't know what to detect.  Looking at $0 is not going to fit the
> bill, because it is quite alright to set CONFIG_SITE manually before
> invoking ./configure.  You will see that the top-level Makefile.in
> does precisely that, when it finds it needs to re-run configure.
> Moreover, someone might even copy mingw-cfg.site to
> $prefix/share/config.site, and I don't think they deserve to be
> punished by throwing an error.  So testing that "$CONFIG_SITE" is
> non-empty is also sub-optimal if not outright wrong.
> 
> I wrote detailed instructions in nt/INSTALL.MSYS so that people would
> know how to avoid these.  If they don't follow, they will realize that
> soon enough...
> 
>> Naively, I would expect quite a lot of people to do that
> 
> I would expect them to read the instructions first.  Maybe it's me who
> is naive.
> 
>> maybe it's common knowledge
> 
> It isn't.
> 
> 

Message #39 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: rgm <at> gnu.org, dgutov <at> yandex.ru
Cc: 14603 <at> debbugs.gnu.org
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 21:07:09 +0300

> Date: Thu, 13 Jun 2013 20:50:50 +0300
> From: Eli Zaretskii <eliz <at> gnu.org>
> Cc: dgutov <at> yandex.ru, 14603 <at> debbugs.gnu.org
> 
> > From: Glenn Morris <rgm <at> gnu.org>
> > Cc: Dmitry Gutov <dgutov <at> yandex.ru>,  14603 <at> debbugs.gnu.org
> > Date: Thu, 13 Jun 2013 12:15:00 -0400
> > 
> > Eli Zaretskii wrote:
> > 
> > >> No arguments, just ./configure
> > >
> > > That will never work. 
> > 
> > Can you make it detect someone has done that and throw an error?
> 
> I don't know what to detect.  Looking at $0 is not going to fit the
> bill, because it is quite alright to set CONFIG_SITE manually before
> invoking ./configure.  You will see that the top-level Makefile.in
> does precisely that, when it finds it needs to re-run configure.
> Moreover, someone might even copy mingw-cfg.site to
> $prefix/share/config.site, and I don't think they deserve to be
> punished by throwing an error.  So testing that "$CONFIG_SITE" is
> non-empty is also sub-optimal if not outright wrong.

I could perhaps output a warning if "$CONFIG_SITE" is found empty, but
will people pay attention?  After all, this warning:

  configure: WARNING: libacl development library was not found or not usable.
  configure: WARNING: emacs will be built without ACL support.

didn't have any effect on Dmitry.

Message #42 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Dmitry Gutov <dgutov <at> yandex.ru>
Cc: 14603 <at> debbugs.gnu.org
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 21:13:33 +0300

> Date: Thu, 13 Jun 2013 22:02:48 +0400
> From: Dmitry Gutov <dgutov <at> yandex.ru>
> CC: 14603 <at> debbugs.gnu.org
> 
> But now, after a few successful builds this way, and after switching 
> from Windows to GNU/Linux and back a few times (I use systems of both 
> kinds), on the one hand, I felt confident that I knew how to do it, on 
> the other hand, the last few times I built Emacs were on Linux, so I 
> completely forgot about "msysconfig.sh".

You don't need to remember.  Just run "make" all the time, it will
DTRT.  The necessary magic is in GNUmakefile.  You don't need to
invoke msysconfig.sh except once per branch.

> Like Glenn said, it would be great if ./configure aborted when the 
> environment is MSYS (would it be too hard to detect that?), or delegated 
> transparently to ./nt/msysconfig.sh.

Patches welcome.

Message #45 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Jan Djärv <jan.h.d <at> swipnet.se>
Cc: rgm <at> gnu.org, 14603 <at> debbugs.gnu.org, dgutov <at> yandex.ru
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 21:18:27 +0300

> From: Jan Djärv <jan.h.d <at> swipnet.se>
> Date: Thu, 13 Jun 2013 20:06:00 +0200
> Cc: Glenn Morris <rgm <at> gnu.org>,
>  dgutov <at> yandex.ru,
>  14603 <at> debbugs.gnu.org
> 
> Can't you just set some environment variable in msysconfig.sh and check for it in configure?
> If it isn't set, msysconfig.sh has not been run.

That won't do the job, because "make" sometimes needs to run the
configure script, and when it does, it doesn't go through
msysconfig.sh.  See GNUmakefile and the top-level Makefile.in.

Anyway, people should feel free to change these scripts as they see
fit.  There's no need to wait for me to do it.  It's not like I'm the
best shell scripting or autoconf guru here; far from that.  I crafted
all this stuff together so that maintainers who know nothing about
Windows would be able to develop and maintain it.

Message #48 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 14603 <at> debbugs.gnu.org, dgutov <at> yandex.ru
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 14:22:49 -0400

Eli Zaretskii wrote:

> I could perhaps output a warning if "$CONFIG_SITE" is found empty, but
> will people pay attention?  After all, this warning:
>
>   configure: WARNING: libacl development library was not found or not usable.
>   configure: WARNING: emacs will be built without ACL support.
>
> didn't have any effect on Dmitry.

That's a bad warning because it appears in the middle of reams of output
that people do not normally pay attention to. http://debbugs.gnu.org/14612

Yours is a fundamental warning that should be the last (and maybe first
as well) thing configure prints.

Message #51 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 14603 <at> debbugs.gnu.org, Jan Djärv <jan.h.d <at> swipnet.se>,
 dgutov <at> yandex.ru
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Thu, 13 Jun 2013 14:26:03 -0400

Eli Zaretskii wrote:

> Anyway, people should feel free to change these scripts as they see
> fit.  There's no need to wait for me to do it.  It's not like I'm the
> best shell scripting or autoconf guru here; far from that.  I crafted
> all this stuff together so that maintainers who know nothing about
> Windows would be able to develop and maintain it.

It was (really) great that you did this work, and it will mean that MS
Windows gets to benefit from a lot of stuff for free, but don't expect
those of us who don't care about MS Windows to suddenly start doing so
and pick up the MS-Windows specific pieces. It just means we are in a
slightly more informed position to bitch about things. ;)

Message #64 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Glenn Morris <rgm <at> gnu.org>
Cc: 14603 <at> debbugs.gnu.org, jan.h.d <at> swipnet.se, dgutov <at> yandex.ru
Subject: Re: bug#14603: 24.3.50; Error while bootstrapping Emacs on Windows
Date: Fri, 14 Jun 2013 09:43:50 +0300

> From: Glenn Morris <rgm <at> gnu.org>
> Cc: Jan Djärv <jan.h.d <at> swipnet.se>,  dgutov <at> yandex.ru,
>   14603 <at> debbugs.gnu.org
> Date: Thu, 13 Jun 2013 14:26:03 -0400
> 
> It was (really) great that you did this work, and it will mean that MS
> Windows gets to benefit from a lot of stuff for free, but don't expect
> those of us who don't care about MS Windows to suddenly start doing so
> and pick up the MS-Windows specific pieces.

The configure script is not Windows specific.  I don't see why you
would refrain from making changes there that cater to the Windows
build, given user requests and bug reports.

Message #67 received at 14603 <at> debbugs.gnu.org (full text, mbox):

From: Noam Postavsky <npostavs <at> users.sourceforge.net>
To: 14603 <at> debbugs.gnu.org
Subject: Re: bug#14603 MS Windows configure needs to be run via msysconfig.sh
Date: Tue, 31 Jan 2017 12:22:30 -0500

tag 14603 fixed
close 14603 24.4
quit

./configure now works for Windows builds too since msysconfig.sh was
removed by f4f38fad9a6 "Remove nt/msysconfig.sh" (2013-11-13).

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.