[GRASS5] configure.in Wants testers...

GRASS GRASS developer
1

On Saturday, December 2, 2000, at 12:02 AM, Eric G . Miller wrote:

> Everything looks good here on MacOS X (I don't have to remove the -lm
> switches anymore :smiley: ) except for the following:
>
> configure: warning: *** Failed to find Tcl/Tk libraries [...snip...]
> checking GL/gl.h and GL/GLwMDrawA.h... configure: warning: *** Didn't
> find GL/GLwMDrawA.h
>
> Note that with beta 8 it gave a similar message tcl/tk message,
> however the tcl/t interface works...

Well, those errors really should/do matter now. The OpenGL stuff is
really only important for NVIZ and showdspf. You may need to specify
one of the flags to configure so it knows where to search for your
Tcl/TK includes in libraries. There will be no linker flag for any
compiled code against Tcl/Tk otherwise (again thinking this may only
matter for NVIZ since I think TclTkGrass is all interpreted scripts, so
if wish is found, it's okay).

Hi Eric!

OK, I tried to run

> ./configure --with-tcltk-libs=/usr/local/lib/
                ^^ where my tcl/tk directories are located

And I get the same error. Any thoughts? I would like to get NVIZ complied (mostly for others)...

Thanks,

Jeshua Lacock
Cartographer/Owner
http://SierraMaps.com
http://3dTopoMaps.com
Telephone: (760) 935-4481
----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

2

On Sat, Dec 02, 2000 at 12:56:59AM -0800, Jeshua Lacock wrote:

Hi Eric!

OK, I tried to run

> ./configure --with-tcltk-libs=/usr/local/lib/
                ^^ where my tcl/tk directories are located

And I get the same error. Any thoughts? I would like to get NVIZ complied (mostly for others)...

Yes, it needs to know where the headers are as well. The configure
script "grep's" for either/both TK_VERSION or TCL_VERSION from the
headers for use in building the library name, like ltk8.3. So, that
will fail if it doesn't get the version number trying instead to link to
non-existent libs like "ltk". So find out where tcl.h or tk.h live and
pass the --with-tcltk-includes=[DIR] parameter. A general solution as
well is to use the --with-includes="/path/to/a /path/to/b" and same for
--with-libs=[DIRS]. Multiple directories need to be quoted, but should
then be made available for every other configure check. These two --with
statements are processed before pretty much everything else.

Hope that helps,

--
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.

Configure is now much better for Cygwin builds thank you.

However:

Under Cygwin, autoconfig should choose "bison -y" instead of "yacc".

It also doesn't find the opengl libraries. (Seems to be looking for GL/*
instead of gl/* etc, ie capital letters instead of lower case.)

Cheers

Mike Thomas.

----- Original Message -----
From: Eric G . Miller <egm2@jps.net>
To: <grass5@geog.uni-hannover.de>
Sent: Sunday, 3 December 2000 5:29 AM
Subject: Re: [GRASS5] configure.in Wants testers...

On Sat, Dec 02, 2000 at 12:56:59AM -0800, Jeshua Lacock wrote:
> Hi Eric!
>
> OK, I tried to run
>
> > ./configure --with-tcltk-libs=/usr/local/lib/
> ^^ where my tcl/tk directories are located
>
> And I get the same error. Any thoughts? I would like to get NVIZ

complied (mostly for others)...

Yes, it needs to know where the headers are as well. The configure
script "grep's" for either/both TK_VERSION or TCL_VERSION from the
headers for use in building the library name, like ltk8.3. So, that
will fail if it doesn't get the version number trying instead to link to
non-existent libs like "ltk". So find out where tcl.h or tk.h live and
pass the --with-tcltk-includes=[DIR] parameter. A general solution as
well is to use the --with-includes="/path/to/a /path/to/b" and same for
--with-libs=[DIRS]. Multiple directories need to be quoted, but should
then be made available for every other configure check. These two --with
statements are processed before pretty much everything else.

Hope that helps,

--
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'

----------------------------------------
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, Dec 04, 2000 at 06:52:01AM +1000, Mike Thomas wrote:

Hi Eric.

Configure is now much better for Cygwin builds thank you.

However:

Under Cygwin, autoconfig should choose "bison -y" instead of "yacc".

Well, there should be some method under cygwin where "yacc" can be used
since that is the "official" name of the utility. But I can look into
it.

It also doesn't find the opengl libraries. (Seems to be looking for GL/*
instead of gl/* etc, ie capital letters instead of lower case.)

The headers are often in /usr/include/GL/ on many systems. I don't know
why that isn't the case on Cygwin. Try passing the
--with-opengl-includes=/path/to/gl directive.

--
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

"Eric G . Miller" wrote:

On Mon, Dec 04, 2000 at 06:52:01AM +1000, Mike Thomas wrote:
> Hi Eric.
>
> Configure is now much better for Cygwin builds thank you.
>
> However:
>
> Under Cygwin, autoconfig should choose "bison -y" instead of "yacc".

Well, there should be some method under cygwin where "yacc" can be used
since that is the "official" name of the utility. But I can look into
it.

Cygwin includes "byacc.exe", which is Berkeley's rewrite
of the original AT&T yacc.

--
+=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=+
| Eric B. Mitchell mailto:emitchell@altaira.com |
| tel: (301) 809 - 3534 Altair Aerospace Corporation |
| tel: (800) 7 - ALTAIR 4201 Northview Dr. Suite 410 |
| fax: (301) 805 - 8122 Bowie, MD 20716 |
+=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=+
              ,___
          /"\ / o=\ /"""---===/
         / \_/ \__/ ---===/
         | //\ || /""TT""/ //\ || ||""\
         | // \ || || // \ || ||__/
         | //--==\ |L--/ || //--==\ || || "=,
          \ ---===/
           \____---===/

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

6

Hi Eric,

a report on configure... If *not* runnung autoconf, the
GISBASE = @GISBASE@

is not expanded in head file (tested on Solaris without autoconf).

If running autoconf, it is o.k. (tested on Linux). It seems,
some code needs to be added to configure. Will you look for
that? I have added the "sed" line for GISBASE already to
get an empty variable in head file.

Markus

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

7

On Tue, Dec 05, 2000 at 05:40:51PM +0000, Markus Neteler wrote:

Hi Eric,

a report on configure... If *not* runnung autoconf, the
GISBASE = @GISBASE@

is not expanded in head file (tested on Solaris without autoconf).

If running autoconf, it is o.k. (tested on Linux). It seems,
some code needs to be added to configure. Will you look for
that? I have added the "sed" line for GISBASE already to
get an empty variable in head file.

Hi again,

I have added the missing code now in "configure" (taken from an
autoconf/configure run). Please check if o.k.

Thanks

Markus

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

8

On Tue, Dec 05, 2000 at 05:51:08PM +0000, Markus Neteler wrote:

On Tue, Dec 05, 2000 at 05:40:51PM +0000, Markus Neteler wrote:
> Hi Eric,
>
> a report on configure... If *not* runnung autoconf, the
> GISBASE = @GISBASE@
>
> is not expanded in head file (tested on Solaris without autoconf).
>
> If running autoconf, it is o.k. (tested on Linux). It seems,
> some code needs to be added to configure. Will you look for
> that? I have added the "sed" line for GISBASE already to
> get an empty variable in head file.
>
Hi again,

I have added the missing code now in "configure" (taken from an
autoconf/configure run). Please check if o.k.

Sorry to write again.

On Solaris I get:
creating Makefile
sed: command garbled: s%@VERSION@%5.0beta9%g
creating src/CMD/head/head
sed: command garbled: s%@VERSION@%5.0beta9%g
creating src/scripts/shells/create_fifos.sh
sed: command garbled: s%@VERSION@%5.0beta9%g
creating src/include/config.h

On Linux it expands o.k. (with updated src/CMD/head/head.in).

I have no clue for that! Might be the dot forbidden in
s%@VERSION@%5.0beta9%g
             ^-?
on Solaris/sed?

If so we should remove the VERSION from configure.

Markus

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

9

On Tue, Dec 05, 2000 at 05:58:14PM +0000, Markus Neteler wrote:

On Tue, Dec 05, 2000 at 05:51:08PM +0000, Markus Neteler wrote:
> On Tue, Dec 05, 2000 at 05:40:51PM +0000, Markus Neteler wrote:
> > Hi Eric,
> >
> > a report on configure... If *not* runnung autoconf, the
> > GISBASE = @GISBASE@
> >
> > is not expanded in head file (tested on Solaris without autoconf).
> >
> > If running autoconf, it is o.k. (tested on Linux). It seems,
> > some code needs to be added to configure. Will you look for
> > that? I have added the "sed" line for GISBASE already to
> > get an empty variable in head file.
> >
> Hi again,
>
> I have added the missing code now in "configure" (taken from an
> autoconf/configure run). Please check if o.k.
>
Sorry to write again.

On Solaris I get:
creating Makefile
sed: command garbled: s%@VERSION@%5.0beta9%g
creating src/CMD/head/head
sed: command garbled: s%@VERSION@%5.0beta9%g
creating src/scripts/shells/create_fifos.sh
sed: command garbled: s%@VERSION@%5.0beta9%g
creating src/include/config.h

On Linux it expands o.k. (with updated src/CMD/head/head.in).

I have no clue for that! Might be the dot forbidden in
s%@VERSION@%5.0beta9%g
             ^-?
on Solaris/sed?

If so we should remove the VERSION from configure.

One more from Solaris:

The NAME_VER doesn't work on SUN either:
It is expanded to:
NAME_VER = -n 5

On linux no problem.

I propose to keep the old method to get VERSION and NAME_VER in Makefile.in
for the beta9 release (simple revert in Makefile.in and removeal of sed
expansion on configure[.in]). We have some more time for the stable version
to get the new approach working.

Markus

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

10

Hi Markus and Eric

Markus Neteler wrote:

I propose to keep the old method to get VERSION and NAME_VER in
Makefile.in for the beta9 release (simple revert in Makefile.in and
removeal of sed expansion on configure[.in]). We have some more time
for the stable version to get the new approach working.

I think I have to agree, I also found a problem with configure. I do not
install grass under /usr/local (I'm the only one using the CVS version)
and use the --prefix option. But then GISBASE is not correct, it is
still /usr/local/grass5. I think that GISBASE needs to be tied with
prefix. Not sure though.

--
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'

11

On Tue, Dec 05, 2000 at 06:07:51PM +0000, Markus Neteler wrote:

On Tue, Dec 05, 2000 at 05:58:14PM +0000, Markus Neteler wrote:
> On Tue, Dec 05, 2000 at 05:51:08PM +0000, Markus Neteler wrote:
> > On Tue, Dec 05, 2000 at 05:40:51PM +0000, Markus Neteler wrote:
> > > Hi Eric,
> > >
> > > a report on configure... If *not* runnung autoconf, the
> > > GISBASE = @GISBASE@
> > >
> > > is not expanded in head file (tested on Solaris without autoconf).
> > >
> > > If running autoconf, it is o.k. (tested on Linux). It seems,
> > > some code needs to be added to configure. Will you look for
> > > that? I have added the "sed" line for GISBASE already to
> > > get an empty variable in head file.
> > >
> > Hi again,
> >
> > I have added the missing code now in "configure" (taken from an
> > autoconf/configure run). Please check if o.k.
> >
> Sorry to write again.
>
> On Solaris I get:
> creating Makefile
> sed: command garbled: s%@VERSION@%5.0beta9%g
> creating src/CMD/head/head
> sed: command garbled: s%@VERSION@%5.0beta9%g
> creating src/scripts/shells/create_fifos.sh
> sed: command garbled: s%@VERSION@%5.0beta9%g
> creating src/include/config.h
>
> On Linux it expands o.k. (with updated src/CMD/head/head.in).
>
> I have no clue for that! Might be the dot forbidden in
> s%@VERSION@%5.0beta9%g
> ^-?
> on Solaris/sed?
>
> If so we should remove the VERSION from configure.
>
One more from Solaris:

The NAME_VER doesn't work on SUN either:
It is expanded to:
NAME_VER = -n 5

On linux no problem.

I propose to keep the old method to get VERSION and NAME_VER in Makefile.in
for the beta9 release (simple revert in Makefile.in and removeal of sed
expansion on configure[.in]). We have some more time for the stable version
to get the new approach working.

Markus, the GISBASE needs NAME_VER to be /usr/local/grass5 by default.
I thought putting these things in configure was to easiest way to
propagate them. As far as the "-n 5", guess the -n option to echo isn't
recognized (should still work okay without it I think). I'm not sure
where that earlier sed problem you mention is coming from. For the most
part, I took the sed scripts from Makefile.in verbatim. I think it's
not too big of a deal to get these fixed up pretty quick....

--
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'

12

On Wed, Dec 06, 2000 at 10:22:25AM +0700, Justin Hickey wrote:

Hi Markus and Eric

Markus Neteler wrote:
> I propose to keep the old method to get VERSION and NAME_VER in
> Makefile.in for the beta9 release (simple revert in Makefile.in and
> removeal of sed expansion on configure[.in]). We have some more time
> for the stable version to get the new approach working.

I think I have to agree, I also found a problem with configure. I do not
install grass under /usr/local (I'm the only one using the CVS version)
and use the --prefix option. But then GISBASE is not correct, it is
still /usr/local/grass5. I think that GISBASE needs to be tied with
prefix. Not sure though.

Hi Justin,

I agree: The GISBASE settings are not correct (it was before).
GISBASE has to be identical with --prefix.

Eric, could you please change that?

Markus

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

13

On Wed, Dec 06, 2000 at 09:58:39AM +0000, Markus Neteler wrote:

On Wed, Dec 06, 2000 at 10:22:25AM +0700, Justin Hickey wrote:
> Hi Markus and Eric
>
> Markus Neteler wrote:
> > I propose to keep the old method to get VERSION and NAME_VER in
> > Makefile.in for the beta9 release (simple revert in Makefile.in and
> > removeal of sed expansion on configure[.in]). We have some more time
> > for the stable version to get the new approach working.
>
> I think I have to agree, I also found a problem with configure. I do not
> install grass under /usr/local (I'm the only one using the CVS version)
> and use the --prefix option. But then GISBASE is not correct, it is
> still /usr/local/grass5. I think that GISBASE needs to be tied with
> prefix. Not sure though.
Hi Justin,

I agree: The GISBASE settings are not correct (it was before).
GISBASE has to be identical with --prefix.

Eric, could you please change that?

Sorry it's not working correctly. I thought I had it set-up as we
discussed off list. That is GISBASE=${prefix}/grass${NAME_VER}.

Prefix is by default set to /usr/local with AC_PREFIX_DEFAULT(). Then
VERSION is grepped from ${SRCDIR}/src/CMD/VERSION, then NAME_VER is
sed'ed from VERSION.

So, if you specify --prefix=/opt, you should get /opt/grass5 for
GISBASE. And BINDIR/UNIX_BIN should get set to ${prefix}/bin
eventually; which would be /opt/bin in this case.

I see there had been some problems with VERSION and NAME_VER, which I
think is part of the problem. I will look at it some more...

--
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'

14

Hi Eric,

[...]

Markus, the GISBASE needs NAME_VER to be /usr/local/grass5 by default.
I thought putting these things in configure was to easiest way to
propagate them. As far as the "-n 5", guess the -n option to echo isn't
recognized (should still work okay without it I think). I'm not sure
where that earlier sed problem you mention is coming from. For the most
part, I took the sed scripts from Makefile.in verbatim. I think it's
not too big of a deal to get these fixed up pretty quick....

Thanks for the fix (I am trying Solaris). Only a minor bug left:

head file:
prefix = /data3/grass5
exec_prefix = ${prefix}

GISBASE = /data3/grass5/grass5

I feel that GISBASE get's one grass5 too much. Should be:

prefix = /data3/grass5
exec_prefix = ${prefix}

GISBASE = /data3/grass5

(I have used: configure --prefix=/data3/grass5)

Markus

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

15

Hi again,

concerning the GISBASE:
It seems I was confused by my own settings. Please
ignore my last mail on this issue.

But... There is still a bug (tested on Linux now):

In Makefile it is expanded to:
GISBASE= NONE/grass5

I used plain configure without parameter(s).

Perhaps I find the problem,

Markus

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

16

Hi eric

Markus Neteler wrote:

concerning the GISBASE:
It seems I was confused by my own settings. Please
ignore my last mail on this issue.

But this was a problem for me. I also get the extra grass5 at the end of
GISBASE So am I correctly interpreting how prefix and GISBASE are
related? They are the same with my setup. Should I change my setup? I'm
not sure.

--
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'

17

On Wed, Dec 06, 2000 at 06:45:42PM +0700, Justin Hickey wrote:

Hi eric

Markus Neteler wrote:
> concerning the GISBASE:
> It seems I was confused by my own settings. Please
> ignore my last mail on this issue.

But this was a problem for me. I also get the extra grass5 at the end of
GISBASE So am I correctly interpreting how prefix and GISBASE are
related? They are the same with my setup. Should I change my setup? I'm
not sure.

Well, this is the deal. To get BINDIR from $prefix when no --bindir is
specified, it defaults to ${prefix}/bin. To get GISBASE from $prefix,
we are making ${prefix}/grass${NAME_VER}. So if you specify
--prefix=/opt/grass5, you mess up both of these settings (e.g. you get
BINDIR=/opt/grass5/bin and GISBASE=/opt/grass5/grass5). This is just
how autoconf works. It's set-up with the idea that $prefix ==
'/usr/local' (on many machines), $bindir = ${prefix}/bin, $libdir =
${prefix}/lib, etc... Since GRASS is somewhat "special" with it's
binaries, we stow all the GRASS stuff under a common directory known as
GISBASE. But, everyone wanted not to have to specify --bindir in order to
get the 'grass5', 'gmake5', and 'gmakelinks5' put in a directory that is
in the user's path. See, there's this interaction that's going on based
on some fundamental concepts of the way autoconf was designed. So don't
specify the "grass5" part to --prefix and it should do the right thing.
(At least that's how it *should* be working). If this isn't
a satisfactory design, I give up!

--
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'

18

Hi Eric,

finally I found the bug in configure. Now it expands properly on
Linux and SUN.

I had to move the variable expanding up (before the GISBASE stuff):

cvs diff configure
Index: configure

RCS file: /grassrepository/grass/configure,v
retrieving revision 1.51
diff -r1.51 configure
3570,3572d3569
< test "x$prefix" = xNONE && prefix=$ac_default_prefix
< # Let make expand exec_prefix.
< test "x$exec_prefix" = xNONE && exec_prefix='${prefix}'
4352a4350,4352

test "x$prefix" = xNONE && prefix=$ac_default_prefix
# Let make expand exec_prefix.
test "x$exec_prefix" = xNONE && exec_prefix='${prefix}'

But I don't know where to move in configure.in.
Hope we can manage this until the beta9 release (in a few hours).

Markus

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

19

Hi Eric,

hope (you all) had nice christmas days!

If you (Eric), allow, another "configure" report:

I tested on my Linux laptop:
checking Tcl include dirs... -I/usr/include
checking Tk include dirs... -I/usr/include
checking Tcl/Tk library dirs... -L/usr/lib
checking for Tk_MainWindow in -ltk8.3... (cached) no
configure: warning: *** Failed to find Tcl/Tk libraries

I have:
ls -la /usr/lib/libtcl
libtcl8.3.a libtcl8.3.so libtclstub8.3.a libtclx8.2.a

ls -la /usr/lib/libtk
libtk8.3.a libtk8.3.so libtkstub8.3.a libtkx8.2.a

Somehow the Tk_MainWindow routine is not yet fully working...
Can I assist somehow to fix this?

Markus

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

20

On Wed, Dec 27, 2000 at 10:06:07AM +0000, Markus Neteler wrote:

Hi Eric,

hope (you all) had nice christmas days!

If you (Eric), allow, another "configure" report:

I tested on my Linux laptop:
checking Tcl include dirs... -I/usr/include
checking Tk include dirs... -I/usr/include
checking Tcl/Tk library dirs... -L/usr/lib
checking for Tk_MainWindow in -ltk8.3... (cached) no
configure: warning: *** Failed to find Tcl/Tk libraries

This seems to happen when the test program doesn't have all the required
link -l<lib> flags. It seems to vary by system. Most likely it needs
some of the X11 flags for the test to succeed (-lX11, -lXt, etc...).
Perhaps you could help me determine which should be added?

Create a small program like:

------------------------<cut here>----------------------------------
char *Tk_MainWindow();

int main (void)
{
   Tk_MainWindow();
   return 0;
}
-----------------------<cut here>----------------------------------

Then try compiling it with various options:

$ gcc tktest.c -ltk8.3

$ gcc tktest.c -ltk8.3 -L/usr/X11R6/lib -lXt -lX11

etc...

The test program should *not* include "tk.h", so that the redefinition
of Tk_MainWindow() will not cause a compilation error. On my system, I
only need the "-ltk8.3" flag, but I've seen where X libs are needed
elsewhere to overcome similar problems (like -lGLwM). The program only
needs to compile to indicate successful linking (it will obviously fail
to run properly due to wrong/missing arguments to Tk_MainWindow()).

Also, if you changed some options to configure, you might need to delete
the "config.cache" file so the test is actually performed (it won't be
if the cache already has success/failure stored).

--
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'