[GRASS5] GRASS 5 beta5b internal available

Hi Bill!

I have checked all Gmakefiles for

-lm -> $(MATHLIB)
and missing $(XDRLIB).

> don't know how to make -lsun (bu42).

-> This will the a wrong statement in
src/CMD/head/head
Please check the settings for XDRLIB here.
I need -lnsl instead of -lsun

> don't know how to make -lm (bu42).

This is the math lib. Should be there on SGI!

I can fix these by changing Gmakefiles like this:
------OLD------
LIBRARIES = $(GISLIB) $(DATETIMELIB) $(XDRLIB)

PGM = $(BIN_CONTRIB_CMD)/s.territory

OFILES = main.o \
  territory.o \
  load.o

$(PGM): $(OFILES) $(LIBRARIES)
  $(CC) -o $@ $(LDFLAGS) $(OFILES) $(LIBRARIES) -lm
  @rm -f a.out
------NEW------
LIBRARIES = $(GISLIB) $(DATETIMELIB)

PGM = $(BIN_CONTRIB_CMD)/s.territory

OFILES = main.o \
   territory.o \
         load.o

$(PGM): $(OFILES) $(LIBRARIES)
  $(CC) -o $@ $(LDFLAGS) $(OFILES) $(LIBRARIES) $(XDRLIB) -lm
  @rm -f a.out

Funny, should be the same?! (Berhard, Bill H., any ideas?)

There were also failures of various sorts on:
> src/general/g.help
> src/mapdev/v.digit
> src/mapdev/v.geom
> src/raster/r.mapcalc
> src/raster/r.tiff
> src.contrib/CERL/imagery/i.rvi.prediction2
> src.contrib/CERL/raster/r.rvi
> src.contrib/GMSL/NVIZ2.2
> src.contrib/GMSL/SG3d

The "qsort" warning I also get.

I also had some trouble with make - I had to comment out
INSTALL_DATA_DIR= install -d -m 755
from the Makefile because IRIX install doesn't recognize -d.
What is INSTALL_DATA_DIR used for?

There was a suggestion in the mailing list. It would require
a change in "configure" or "Makefile.in".

I ran the new grass5.0beta5a and it ran fine, as did tcltkgrass. But
when I went back to use my grass5.0beta2, which is my last stable
version, I get errors when I try to start a monitor - it starts OK,
but cannot lock so I can't "select" it. It was running fine before I
compiled the beta5 grass. Anybody know why this would happen? Or
how to fix it? The beta5 and beta2 binaries are in completely different
directory trees.

Sorry, no idea. The XDRIVER was updated to full 24bit between
beta2 and beta3 but this should not matter the grass-5.0b/locks/.

Another question: Does the current configuration script still allow
me to compile the same source tree on several different platforms? I
noticed that it still creates arch-specific subdirectories for object files,
but the binary tree no longer has the arch-specific directory.

Then check the ARCH settings in src/CMD/head/head. Of course it should
create different obj-directories.E.g
LIB.i586-linux-elf/
OBJ.i586-linux-elf/

For the created binaries you can store them into different places
with parameter. The 'make install' command should install GRASS in the
main filesystem. If you want to change this, use a sequence like this:

        ./configure --prefix=/opt/grass5.0 --with-bindir=/usr/bin/
        make install

This will install the GRASS modules, help, and other need libraries
etc. in the directory /opt/grass5.0 and the startup file 'grass5.0'
in the directory /usr/bin.

Hope this (partly) helps

Markus

----------------------------------------
If you want to unsubscribe from GRASS Development
Team internal mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'
length: 3683
max: 0

Hi all I am going to compile beta5b on the alpha today.
I just want to signal that I had the same problems reported
by Bill when compiling under DU alpha, namely in many
Gmakefiles libraries in place of dependences for targets.

Markus Neteler wrote:

>
> I can fix these by changing Gmakefiles like this:
> ------OLD------
> LIBRARIES = $(GISLIB) $(DATETIMELIB) $(XDRLIB)
>
> PGM = $(BIN_CONTRIB_CMD)/s.territory

> ------NEW------

> LIBRARIES = $(GISLIB) $(DATETIMELIB)
>
> PGM = $(BIN_CONTRIB_CMD)/s.territory
>
> $(PGM): $(OFILES) $(LIBRARIES)
> $(CC) -o $@ $(LDFLAGS) $(OFILES) $(LIBRARIES) $(XDRLIB) -lm
> @rm -f a.out

Funny, should be the same?! (Berhard, Bill H., any ideas?)

No !!! in my understanding of make it is not the same !!!
In the first version XDRLIB appears to be a dependency of $PGM, it means
that PGM cannot be compiled if the object XDRLIB is not obtained somewhere
before (but no target XRLIB is present in the Gmakefile).
In the second version instead (correctly) the XDRLIB appears only as
a linker option to the compiler.
This is how I view it, but Gmake is not make :), so I may be wrong (???)

> I also had some trouble with make - I had to comment out
> INSTALL_DATA_DIR= install -d -m 755
> from the Makefile because IRIX install doesn't recognize -d.
> What is INSTALL_DATA_DIR used for?

The same problem is encountered under alpha DU. DU install

does not like the -d option, and as far as I can read the

man does not have any similar option. SO I just had to

substitute install -d -m 755 with mkdir -m 755 that is

similar (but not identical !!!) to install.

I'll let you know on beta5b

bye

LP

--
--> Luca Palmeri - Dip. Proc. Chimici dell'Ingengeria
    Universita' di Padova TEL: +390498275527 -------
------------------------- FAX: +390498275528 -------
--> Think holistic @ the end of the 2nd millenium <--

----------------------------------------
If you want to unsubscribe from GRASS Development
Team internal mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'
length: 3281
max: 0

Hi all,

     the following patch applies to s.surf.idw/cmd/main.c and

corrects a segfaulting that appens when the surface interpolation is

executed with a mask set.

bye

LP

--
--> Luca Palmeri - Dip. Proc. Chimici dell'Ingengeria
    Universita' di Padova TEL: +390498275527 -------
------------------------- FAX: +390498275528 -------
--> Think holistic @ the end of the 2nd millenium <--

(attachments)

s.surf.idw.patch (113 Bytes)