[GRASS5] configure-problem on linux and irix

GRASS GRASS developer
1

Hi,

there must be some small errors with configure.
configure no longer finds GLwMDrawA.h, which was found with the same
setup once before (file irix-configure contains all switches i used to
compile a perfectly working bin installation on irix 6.5/with gcc).
And there must be some typo around line 3404 in configure.

Please see the attached files.

cu,

Andreas

--
Andreas Lange, 65187 Wiesbaden, Germany, Tel. +49 611 807850
Andreas.Lange@Rhein-Main.de - A.C.Lange@GMX.net

(attachments)

irix-configure.txt (392 Bytes)
config.err.txt (3.21 KB)

2

On Sat, Jan 06, 2001 at 02:00:34PM +0100, Andreas Lange wrote:

Hi,

there must be some small errors with configure.
configure no longer finds GLwMDrawA.h, which was found with the same
setup once before (file irix-configure contains all switches i used to
compile a perfectly working bin installation on irix 6.5/with gcc).
And there must be some typo around line 3404 in configure.

./configure --bindir=/usr/people/andreas/bin \
--enable-postgres=no \
--with-jpeg-includes=/usr/freeware/include \
--with-jpeg-libs=/usr/freeware/lib32 \
--with-tiff-includes=/usr/freeware/include \
--with-tiff-libs=/usr/freeware/lib32 \
--with-tcltk-includes=/usr/freeware/include \
--with-tcltk-libs=/usr/freeware/lib32 \
--with-opengl-includes=/usr/include \

You can leave this off or make it /usr/include/GL and it should work.
It should have never worked before, since "GL" isn't appended when the
directory is specified.

--with-opengl-libs=/usr/lib32

[snip]

checking TIFF include dirs... -I/usr/freeware/include
checking additional TIFF library dirs... -L/usr/freeware/lib32
checking for TIFFOpen in -ltiff... yes
checking Tcl include dirs... -I/usr/freeware/include
./configure[3404]: test: argument expected

I looked, but couldn't find anything. I don't see that error here, even
when I specify --with-tcltk-includes=/usr/include/tcl8.3. All the
"test" calls have arguments (though some, of course, may have empty
values).

checking Tk include dirs... -I/usr/freeware/include
checking additional Tcl and Tk library dirs... -L/usr/freeware/lib32
checking for Tk_MainWindow in -ltk8.0... yes
checking GL/gl.h and GL/GLwMDrawA.h... configure: warning: *** Didn't find GL/GLwMDrawA.h
-I/usr/include and

See above.

checking for glBegin in -lGL... yes
checking for gluBeginCurve in -lGLU... yes
checking for GLwDrawingAreaMakeCurrent in -lGLw... no
checking for GLwCreateMDrawingArea in -lGLwM... no

I'm not sure what's up with these last two. They can fail if X_LIBS
isn't set correctly. It's possible I need to add the X_PRE_LIBS as well
(X_LIBS and X_PRE_LIBS being set by the AC_XTRA macro).

--
Eric G. Miller <egm2@jps.net>

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

3

Hi Eric and Andreas

"Eric G . Miller" wrote:

> checking TIFF include dirs... -I/usr/freeware/include
> checking additional TIFF library dirs... -L/usr/freeware/lib32
> checking for TIFFOpen in -ltiff... yes
> checking Tcl include dirs... -I/usr/freeware/include
> ./configure[3404]: test: argument expected

I looked, but couldn't find anything. I don't see that error here,
even when I specify --with-tcltk-includes=/usr/include/tcl8.3. All
the "test" calls have arguments (though some, of course, may have
empty values).

I found what is causing the problem. The line it refers to is

if test $tkincdir && test -z $HAVE_TK_H ; then

and $HAVE_TK_H is empty at that point. I don't know why it is empty but
this is what is causing the error.

--
Sincerely,

Jazzman (a.k.a. Justin Hickey) e-mail: jhickey@hpcc.nectec.or.th
High Performance Computing Center
National Electronics and Computer Technology Center (NECTEC)
Bangkok, Thailand

People who think they know everything are very irritating to those
of us who do. ---Anonymous

Jazz and Trek Rule!!!

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

4

On Mon, Jan 08, 2001 at 02:23:17PM +0700, Justin Hickey wrote:

Hi Eric and Andreas

"Eric G . Miller" wrote:
> > checking TIFF include dirs... -I/usr/freeware/include
> > checking additional TIFF library dirs... -L/usr/freeware/lib32
> > checking for TIFFOpen in -ltiff... yes
> > checking Tcl include dirs... -I/usr/freeware/include
> > ./configure[3404]: test: argument expected
>
> I looked, but couldn't find anything. I don't see that error here,
> even when I specify --with-tcltk-includes=/usr/include/tcl8.3. All
> the "test" calls have arguments (though some, of course, may have
> empty values).

I found what is causing the problem. The line it refers to is

if test $tkincdir && test -z $HAVE_TK_H ; then

and $HAVE_TK_H is empty at that point. I don't know why it is empty but
this is what is causing the error.

Ahh, yes. Someone other than myself commented out the portion of
configure.in that would've set that variable... Will fix.

--
Eric G. Miller <egm2@jps.net>

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

5

Hi Eric

> I found what is causing the problem. The line it refers to is
>
> if test $tkincdir && test -z $HAVE_TK_H ; then
>
> and $HAVE_TK_H is empty at that point. I don't know why it is empty
> but this is what is causing the error.

Ahh, yes. Someone other than myself commented out the portion of
configure.in that would've set that variable... Will fix.

Sorry, but I still get the error message after updating configure.in,
running autoconf, removing config.cache, and running configure. My
experience has been that variables somtimes need to be inside "" to work
as expected on some systems. What about trying "$HAVE_TK_H"?

--
Sincerely,

Jazzman (a.k.a. Justin Hickey) e-mail: jhickey@hpcc.nectec.or.th
High Performance Computing Center
National Electronics and Computer Technology Center (NECTEC)
Bangkok, Thailand

People who think they know everything are very irritating to those
of us who do. ---Anonymous

Jazz and Trek Rule!!!

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

6

On Tue, Jan 09, 2001 at 12:07:27PM +0700, Justin Hickey wrote:

Hi Eric

> > I found what is causing the problem. The line it refers to is
> >
> > if test $tkincdir && test -z $HAVE_TK_H ; then
> >
> > and $HAVE_TK_H is empty at that point. I don't know why it is empty
> > but this is what is causing the error.
>
> Ahh, yes. Someone other than myself commented out the portion of
> configure.in that would've set that variable... Will fix.

Sorry, but I still get the error message after updating configure.in,
running autoconf, removing config.cache, and running configure. My
experience has been that variables somtimes need to be inside "" to work
as expected on some systems. What about trying "$HAVE_TK_H"?

Hmm, well it would seem such an error would crop up elsewhere as well.
Anyway, why don't you try out some of your ideas. I can't reproduce the
problem here, so it's hard for me to find a solution. That whole
$HAVE_TK_H could probably be removed anyway since the code that would've
set it has been commented out.

--
Eric G. Miller <egm2@jps.net>

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

7

Hi Eric

"Eric G . Miller" wrote:

> What about trying "$HAVE_TK_H"?

Hmm, well it would seem such an error would crop up elsewhere as well.
Anyway, why don't you try out some of your ideas. I can't reproduce
the problem here, so it's hard for me to find a solution. That whole
$HAVE_TK_H could probably be removed anyway since the code that
would've set it has been commented out.

The ""'s worked, but since HAVE_TK_H is commented out I removed them as
well as HAVE_TCL_H. Commiting changes now.

Thanks for the tip.

--
Sincerely,

Jazzman (a.k.a. Justin Hickey) e-mail: jhickey@hpcc.nectec.or.th
High Performance Computing Center
National Electronics and Computer Technology Center (NECTEC)
Bangkok, Thailand

People who think they know everything are very irritating to those
of us who do. ---Anonymous

Jazz and Trek Rule!!!

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

8

Thanks a lot for looking at the problems.

Eric G . Miller wrote:

> checking Tk include dirs... -I/usr/freeware/include
> checking additional Tcl and Tk library dirs... -L/usr/freeware/lib32
> checking for Tk_MainWindow in -ltk8.0... yes
> checking GL/gl.h and GL/GLwMDrawA.h... configure: warning: *** Didn't find GL/GLwMDrawA.h
> -I/usr/include and

See above.

> checking for glBegin in -lGL... yes
> checking for gluBeginCurve in -lGLU... yes
> checking for GLwDrawingAreaMakeCurrent in -lGLw... no
> checking for GLwCreateMDrawingArea in -lGLwM... no

the last two checks fail on linux too with the same message. And i
didn't specify any paths except --with-tcltk-includes=/usr/include
--with-tcltk-libs=/usr/lib (this is still neccessary and i wonder why?).

But i can not check as my CVS doesn't compile any longer.

cu,

Andreas

--
Andreas Lange, 65187 Wiesbaden, Germany, Tel. +49 611 807850
Andreas.Lange@Rhein-Main.de - A.C.Lange@GMX.net

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

9

On Fri, Jan 12, 2001 at 01:23:38PM +0100, Andreas Lange wrote:

Thanks a lot for looking at the problems.

Eric G . Miller wrote:
>
> > checking Tk include dirs... -I/usr/freeware/include
> > checking additional Tcl and Tk library dirs... -L/usr/freeware/lib32
> > checking for Tk_MainWindow in -ltk8.0... yes
> > checking GL/gl.h and GL/GLwMDrawA.h... configure: warning: *** Didn't find GL/GLwMDrawA.h
> > -I/usr/include and
>
> See above.
>
> > checking for glBegin in -lGL... yes
> > checking for gluBeginCurve in -lGLU... yes
> > checking for GLwDrawingAreaMakeCurrent in -lGLw... no
> > checking for GLwCreateMDrawingArea in -lGLwM... no
>
the last two checks fail on linux too with the same message. And i
didn't specify any paths except --with-tcltk-includes=/usr/include
--with-tcltk-libs=/usr/lib (this is still neccessary and i wonder why?).

I don't know why you'd need to specify the tcltk things. Those paths
are both in the default search paths.

As far as the GLw... stuff, my guess is the LDFLAGS needs $X_PRE_LIBS
added but I don't know for sure. It seems most likely that there are
additional linker flags needed for the library check which are currently
not being specified. I just don't know which flags are needed.

--
Eric G. Miller <egm2@jps.net>

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

10

Hi again,

On Fri, Jan 12, 2001 at 03:36:32PM -0800, Eric G . Miller wrote:

On Fri, Jan 12, 2001 at 01:23:38PM +0100, Andreas Lange wrote:
> Eric G . Miller wrote:

[...]

> > > checking GL/gl.h and GL/GLwMDrawA.h... configure: warning: *** Didn't find GL/GLwMDrawA.h
> > > -I/usr/include and
> >
> > See above.
> >
> > > checking for glBegin in -lGL... yes
> > > checking for gluBeginCurve in -lGLU... yes
> > > checking for GLwDrawingAreaMakeCurrent in -lGLw... no
> > > checking for GLwCreateMDrawingArea in -lGLwM... no
> >

[...]

As far as the GLw... stuff, my guess is the LDFLAGS needs $X_PRE_LIBS
added but I don't know for sure. It seems most likely that there are
additional linker flags needed for the library check which are currently
not being specified. I just don't know which flags are needed.

Eric, do you have a hit for me on this one
(r.showdspf.opengl compiled fine for the last month, I can't remember that I
changed anything related on my Linux machine):

src.contrib/GMSL/g3d/src3d/raster/r3.showdspf.openGL > gmake5
[...]
gcc -s -o /home/neteler/ggg/dist.i686-pc-linux-gnu/etc/bin/cmd/r3.showdspf
OBJ.i686-pc-linux-gnu/new_init_graphics.o OBJ.i686-pc-linux-gnu/cap_data.o
OBJ.i686-pc-linux-gnu/init_caps.o OBJ.i686-pc-linux-gnu/debugf.o
OBJ.i686-pc-linux-gnu/do_color_ogl.o OBJ.i686-pc-linux-gnu/get_color_ogl.o
OBJ.i686-pc-linux-gnu/draw_cap_ogl.o OBJ.i686-pc-linux-gnu/draw_polys_ogl.o
OBJ.i686-pc-linux-gnu/draw_cappolys_ogl.o
OBJ.i686-pc-linux-gnu/thresh_array.o OBJ.i686-pc-linux-gnu/make_header.o
OBJ.i686-pc-linux-gnu/r3_data.o OBJ.i686-pc-linux-gnu/r3_find.o
OBJ.i686-pc-linux-gnu/togif.o OBJ.i686-pc-linux-gnu/Ball.o
OBJ.i686-pc-linux-gnu/BallAux.o OBJ.i686-pc-linux-gnu/BallMath.o
OBJ.i686-pc-linux-gnu/main_ogl.o -I/home/neteler/ggg/src/libes/dspf
-I/home/neteler/ggg/src/libes/ogsf -I/usr/include/GL -I/usr/X11R6/include/GL
-I/usr/X11R6/include
/home/neteler/ggg/src/libes/dspf/LIB.i686-pc-linux-gnu/libdsp3.a
/home/neteler/ggg/src/libes/LIB.i686-pc-linux-gnu/libg3d.a
/home/neteler/ggg/src/libes/LIB.i686-pc-linux-gnu/libdatetime.a
/home/neteler/ggg/src/libes/LIB.i686-pc-linux-gnu/libgis.a -lGLU -lGL -lGLw
-lXm -lX11 -lXt -lXext -L/usr/X11R6/lib -L/lib -lm -lz
OBJ.i686-pc-linux-gnu/new_init_graphics.o: In function 'init_graphics':
/home/neteler/ggg/src.contrib/GMSL/g3d/src3d/raster/r3.showdspf.openGL/new_init_graphics.c:570:
undefined reference to 'GLwCreateMDrawingArea'
/home/neteler/ggg/src.contrib/GMSL/g3d/src3d/raster/r3.showdspf.openGL/new_init_graphics.c:629:
undefined reference to 'GLwCreateMDrawingArea'

It must have something to do with configure.

(config.cache removed):
configure
[...]
checking GL/gl.h and GL/GLwMDrawA.h... -I/usr/include/GL and -I/usr/X11R6/include/GL
checking for glBegin in -lGL... yes
checking for gluBeginCurve in -lGLU... yes
checking for GLwDrawingAreaMakeCurrent in -lGLw... yes
checking for GLwCreateMDrawingArea in -lGLwM... no
[...]

grep GLwCreateMDrawingArea /usr/lib/libGLw.so.3.0
Übereinstimmungen in Binärdatei /usr/lib/libGLw.so.3.0.
   |-> matches in binary file

So GLwCreateMDrawingArea is there...

Any ideas?

Markus

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

11

On Sat, Jan 13, 2001 at 09:49:26AM +0000, Markus Neteler wrote:

Eric, do you have a hit for me on this one
(r.showdspf.opengl compiled fine for the last month, I can't remember that I
changed anything related on my Linux machine):

src.contrib/GMSL/g3d/src3d/raster/r3.showdspf.openGL > gmake5
[...]
gcc -s -o /home/neteler/ggg/dist.i686-pc-linux-gnu/etc/bin/cmd/r3.showdspf
OBJ.i686-pc-linux-gnu/new_init_graphics.o OBJ.i686-pc-linux-gnu/cap_data.o
OBJ.i686-pc-linux-gnu/init_caps.o OBJ.i686-pc-linux-gnu/debugf.o
OBJ.i686-pc-linux-gnu/do_color_ogl.o OBJ.i686-pc-linux-gnu/get_color_ogl.o
OBJ.i686-pc-linux-gnu/draw_cap_ogl.o OBJ.i686-pc-linux-gnu/draw_polys_ogl.o
OBJ.i686-pc-linux-gnu/draw_cappolys_ogl.o
OBJ.i686-pc-linux-gnu/thresh_array.o OBJ.i686-pc-linux-gnu/make_header.o
OBJ.i686-pc-linux-gnu/r3_data.o OBJ.i686-pc-linux-gnu/r3_find.o
OBJ.i686-pc-linux-gnu/togif.o OBJ.i686-pc-linux-gnu/Ball.o
OBJ.i686-pc-linux-gnu/BallAux.o OBJ.i686-pc-linux-gnu/BallMath.o
OBJ.i686-pc-linux-gnu/main_ogl.o -I/home/neteler/ggg/src/libes/dspf
-I/home/neteler/ggg/src/libes/ogsf -I/usr/include/GL -I/usr/X11R6/include/GL
-I/usr/X11R6/include
/home/neteler/ggg/src/libes/dspf/LIB.i686-pc-linux-gnu/libdsp3.a
/home/neteler/ggg/src/libes/LIB.i686-pc-linux-gnu/libg3d.a
/home/neteler/ggg/src/libes/LIB.i686-pc-linux-gnu/libdatetime.a
/home/neteler/ggg/src/libes/LIB.i686-pc-linux-gnu/libgis.a -lGLU -lGL -lGLw
-lXm -lX11 -lXt -lXext -L/usr/X11R6/lib -L/lib -lm -lz
OBJ.i686-pc-linux-gnu/new_init_graphics.o: In function 'init_graphics':
/home/neteler/ggg/src.contrib/GMSL/g3d/src3d/raster/r3.showdspf.openGL/new_init_graphics.c:570:
undefined reference to 'GLwCreateMDrawingArea'
/home/neteler/ggg/src.contrib/GMSL/g3d/src3d/raster/r3.showdspf.openGL/new_init_graphics.c:629:
undefined reference to 'GLwCreateMDrawingArea'

It must have something to do with configure.

(config.cache removed):
configure
[...]
checking GL/gl.h and GL/GLwMDrawA.h... -I/usr/include/GL and -I/usr/X11R6/include/GL
checking for glBegin in -lGL... yes
checking for gluBeginCurve in -lGLU... yes
checking for GLwDrawingAreaMakeCurrent in -lGLw... yes
checking for GLwCreateMDrawingArea in -lGLwM... no
[...]

grep GLwCreateMDrawingArea /usr/lib/libGLw.so.3.0
Übereinstimmungen in Binärdatei /usr/lib/libGLw.so.3.0.
   |-> matches in binary file

So GLwCreateMDrawingArea is there...

Any ideas?

Nope. I just did a cvs update a few hours back, did a make distclean,
ran configure, and built it just fine... Like I've said, I think the
problem lies in missing flags for configure (specifically -L and -l
types) for some segment of X and Motif/Lesstif. Whatever they are, I
don't need them so I can't figure out what they are. Try adding the
$X_PRE_LIBS to the $LD_FLAGS=... part of configure.in just before the
AC_CHECK_LIB(GLwM,...). The $X_PRE_LIBS should be defined by the
AC_XTRA macro (it's results are already used in src/CMD/head/head.in).

On a related note, I've noticed many of the G3d public functions don't
have prototypes in G3d.h. This may be related to some other problems
with programs such as r3.out.vis5d. I don't know why some are declared
and others are not -- and I am talking about just the public functions
(as described in the programming manual).

--
Eric G. Miller <egm2@jps.net>

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'