12:31 p.m.
Hi,
My work machine crashed during my summer vacation. I guess it needed a
break to! So I have reinstalled it from scratch.
When I got to cygwin I was careful not to include any X11
stuff. However when I configure a XEMacs 21.5, latest and greatest,
build configure complains, and stops, about not finding the X11 header
files. configure seems to be looking specifically for
X11/Intrincis.h. Why!? Need assistance...
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
4:30 p.m.
Hi, Mats -
(Sorry I haven't yet responded to your 21.4 ping, but I had hoped that
I would get to it during my vacation last week. That didn't happen,
but I'm hopeful of generating a 21.4 release before the end of
September.)
On Tue, Sep 7, 2010 at 7:31 AM, Mats Lidell <mats.lidell(a)cag.se> wrote:
My work machine crashed during my summer vacation. I guess it needed
a
break to! So I have reinstalled it from scratch.
When I got to cygwin I was careful not to include any X11
stuff. However when I configure a XEMacs 21.5, latest and greatest,
build configure complains, and stops, about not finding the X11 header
files. configure seems to be looking specifically for
X11/Intrincis.h. Why!? Need assistance...
I don't know how you configured, but I configure my cygwin 21.5 XEmacs
like this:
--without-x --with-dragndrop --pdump --with-dialogs --with-widgets
--with-sound=native
and there is no X dependency (from Installation):
Window System:
Compiling in support for the Microsoft window system.
Using MS-Windows menubars.
Using MS-Windows scrollbars.
Using MS-Windows dialog boxes.
Using MS-Windows native widgets.
Compiling in support for Drag'n'Drop (EXPERIMENTAL).
- Drag'n'Drop prototype: msw.
- Vin
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
9:59 p.m.
>>>> Vin Shelton <acs(a)alumni.princeton.edu> writes:
Vin> (Sorry I haven't yet responded to your 21.4 ping, but I had hoped that
Vin> I would get to it during my vacation last week.)
No, problem. Take the time you need.
Vin> I don't know how you configured, but I configure my cygwin 21.5 XEmacs
Vin> like this:
Vin> --without-x --with-dragndrop --pdump --with-dialogs --with-widgets
Vin> --with-sound=native
OK. The "--without-x" looks interesting. I have used "--with-msw" and
assumed that it would be the right thing. I'll try your configuration
instead.
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
2:12 a.m.
Mats Lidell writes:
OK. The "--without-x" looks interesting. I have used
"--with-msw"
and assumed that it would be the right thing. I'll try your
configuration instead.
I'm a bit surprised that it *requires* the X11 headers.
The rationale for looking for them is that you can build on MS Windows
with *both* X and MSW support[1], which allows you to run XEmacs on a
Windows box and access it from an X terminal (which might not be
capable of running XEmacs, or might not allow installation of XEmacs
for administrative reasons).
Footnotes:
[1] There's no point in building with both X and MSW on non-Windows,
since Windows is a local-only console unless you use a transparent
proxy like VNC.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
5:59 a.m.
Stephen J. Turnbull writes 2010-09-08 03:12:
Mats Lidell writes:
> OK. The "--without-x" looks interesting. I have used
"--with-msw"
> and assumed that it would be the right thing. I'll try your
> configuration instead.
I'm a bit surprised that it *requires* the X11 headers.
Well this is what I get using this configuration:
./configure --prefix=/src/xemacs \
--with-site-includes=/usr/include/noX \
--with-mule \
--with-msw \
--with-package-path=~/.xemacs:/src/xemacs/xemacs-packages:src/xemacs/mule-packages:/src/xemacs/site-packages
\
--with-newgc
...
checking Xm/Xm.h usability... no
checking Xm/Xm.h presence... no
checking for Xm/Xm.h... no
checking for X defines extracted by xmkmf...
checking X11/Intrinsic.h usability... no
checking X11/Intrinsic.h presence... no
checking for X11/Intrinsic.h... no
configure: error: Unable to find X11 header files.
$
So stopping due to this error is what I mean by requires.
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
8:46 a.m.
Mats Lidell writes:
So stopping due to this error is what I mean by requires.
Right, I understood that. Normally, however, when the user doesn't
specifically request a feature and it's missing, we would build
anyway, perhaps with a warning (like when XPM is missing).
Should we change this so that you don't need --without-x to build if
you don't have X11 installed?
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
9:16 a.m.
On Wed, Sep 8, 2010 at 9:46 AM, Stephen J. Turnbull <stephen(a)xemacs.org> wrote:
Mats Lidell writes:
> So stopping due to this error is what I mean by requires.
Right, I understood that. Normally, however, when the user doesn't
specifically request a feature and it's missing, we would build
anyway, perhaps with a warning (like when XPM is missing).
Should we change this so that you don't need --without-x to build if
you don't have X11 installed?
That is already the case, I don't need to use --without-x on cygwin. I
suspect Mats has some other package installed that is confusing
configure into thinking he does have X11. Mats, could you sends me all
the output from configure and config.log, and I can compare to what
happens here?
Regards
Robert
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
9:55 a.m.
>>>> "Stephen" == Stephen J Turnbull
<stephen(a)xemacs.org> writes:
Stephen> Should we change this so that you don't need --without-x to
Stephen> build if you don't have X11 installed?
Looking through configure.ac, you shouldn't need to specify --without-x
in this case, but something odd is happening.
Specifying --with-msw doesn't disable searching for X, but if X can't
be found then the test Matt is failing on shouldn't be run at all (this
is on lines line 2969--2981 of configure.ac). Perhaps the autoconf's
AC_PATH_XTRA has changed its interface and we're always assuming that it
succeeds?
Matt, can you please post your config.log?
Malcolm
--
Malcolm Purvis <malcolmp(a)xemacs.org>
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
4:21 p.m.
>>>> Malcolm Purvis <malcolmp(a)xemacs.org> writes:
Malcolm> Matt, can you please post your config.log?
I already mailed the logs to Robert Pluim who correctly analysed it
suggesting that I actually had som X11 stuff somewhere. Thanks Robert!
And, yes, mea culpa, I did. The gif-devel-libraries, selected by me,
dragged in X11. I also found the xproto package installed. I don't
understand how that got there though.
Anyway using the cygwin setup I was able to unstall all X11 stuff and,
voila, now configure works fine. No --with-msw nor --without-x is
required!
So you could argue that it was my cygwin installation that was broken
by having portions of X11 installed and thus fooling configure.
Maybe a few lines describing how to configure cygwin would be
helpful. I'll see what I can come up with.
Yours
--
%% Mats
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
5:01 a.m.
Mats Lidell writes:
So you could argue that it was my cygwin installation that was
broken by having portions of X11 installed and thus fooling
configure.
Yes, I think I would argue that. I sympathize with, but on balance
disagree with Robert Pluim. It is not obvious that a user with a
partial X installation would want XEmacs to build a TTY-only version
(think of broken Linux distros which don't by default distribute
headers with the libraries they supply -- FVO "broken" == "almost all
Linux distros" with tongue firmly planted in cheek).
I think it's better to barf and have the user contact us or the
platform distro for help.
Maybe a few lines describing how to configure cygwin would be
helpful. I'll see what I can come up with.
Just extracting a few lines of error and explanation and adding to
PROBLEMS under Cygwin (add Cygwin section if needed, like "Linux" or
"Motif" sections) would be a big help, I think. Clean it up a bit and
maybe it can go in the FAQ too.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
9:37 a.m.
On Thu, Sep 9, 2010 at 6:01 AM, Stephen J. Turnbull <stephen(a)xemacs.org> wrote:
Mats Lidell writes:
> So you could argue that it was my cygwin installation that was
> broken by having portions of X11 installed and thus fooling
> configure.
Yes, I think I would argue that. I sympathize with, but on balance
disagree with Robert Pluim. It is not obvious that a user with a
partial X installation would want XEmacs to build a TTY-only version
(think of broken Linux distros which don't by default distribute
headers with the libraries they supply -- FVO "broken" == "almost all
Linux distros" with tongue firmly planted in cheek).
I was thinking more along the lines of continuing on and building an
MSW version even if the X installation was broken, thus only on
cygwin. I have no idea how much autoconf (re)work that would take, and
I'm not sure I want to find out :)
I think it's better to barf and have the user contact us or the
platform distro for help.
For this we should at least make the error message a little clearer.
Something like the following?
diff --git a/configure.ac b/configure.ac
--- a/configure.ac
+++ b/configure.ac
＠＠ -3089,7 +3089,7 ＠＠
dnl make sure we can find Intrinsic.h
AC_CHECK_HEADER(X11/Intrinsic.h, ,
- [AC_MSG_ERROR([Unable to find X11 header files.])])
+ [AC_MSG_ERROR([X autodetected or requested, but unable to find X11
header files.])])
dnl -lXt and -lX11 are required
dnl Some broken systems require the magic "-b i486-linuxaout" flag
Robert
--
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
12:46 p.m.
Robert Pluim writes:
b/configure.ac ＠＠ -3089,7 +3089,7 ＠＠
dnl make sure we can find Intrinsic.h
AC_CHECK_HEADER(X11/Intrinsic.h, ,
- [AC_MSG_ERROR([Unable to find X11 header files.])]) +
[AC_MSG_ERROR([X autodetected or requested, but unable to find X11
header files.])])
Let's make it specific:
X autodetected or requested, but Xt header file Intrinsic.h not found.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
1:40 p.m.
On Thu, Sep 9, 2010 at 1:46 PM, Stephen J. Turnbull <stephen(a)xemacs.org> wrote:
Robert Pluim writes:
> b/configure.ac ＠＠ -3089,7 +3089,7 ＠＠
>
> dnl make sure we can find Intrinsic.h
> AC_CHECK_HEADER(X11/Intrinsic.h, ,
> - [AC_MSG_ERROR([Unable to find X11 header files.])]) +
> [AC_MSG_ERROR([X autodetected or requested, but unable to find X11
> header files.])])
Let's make it specific:
X autodetected or requested, but Xt header file Intrinsic.h not found.
Inline & attached. Regenerating configure showed only that line
changed. Someone with commit privileges will have to commit it.
diff --git a/configure.ac b/configure.ac
--- a/configure.ac
+++ b/configure.ac
＠＠ -3089,7 +3089,7 ＠＠
dnl make sure we can find Intrinsic.h
AC_CHECK_HEADER(X11/Intrinsic.h, ,
- [AC_MSG_ERROR([Unable to find X11 header files.])])
+ [AC_MSG_ERROR([X autodetected or requested, but Xt header file
Intrinsic.h not found.])])
dnl -lXt and -lX11 are required
dnl Some broken systems require the magic "-b i486-linuxaout" flag
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/mailman/listinfo/xemacs-beta
4977
days inactive
4979
days old
12 comments
6 participants
participants (6)
-
Malcolm Purvis -
Mats Lidell -
Robert Pluim -
Stephen J. Turnbull -
The XEmacs Package Smoketest -
Vin Shelton