[GRASSLIST:5614] Re: GRASS 5.0.1 compilation issues

[Glynn Clements]

Wayne Gibson wrote:

> curses library
> --------------
> At least on our system, I switched to libcurses(/usr/lib/libcurses.a or 
> /usr/ccs/libcurses.a) and not libncurses.  This fixed the issues with the 
> compile of "set_data".  Don't exactly know the ramifications of that decision 
> but I got a clean compile and so far GRASS seems to work.  To do this, I 
> modified the file src/CMD/head/head directly before compiling.  Is there some 
> way that the configure script could be modified to use a different library for 
> SUN's (assuming this was the correct fix)?

There isn't any practical way to prevent configure from detecting a
library if it's in the linker's search path.

If you have a version of libncurses which doesn't work, and it's in a
directory which is in the linker's search path, the only solution is
to remove it.

OTOH, the problem may have been due to both libcurses and libncurses
being used. The most likely reason for that would be the bogus
"termlib" check, which has since been removed.

> Environmental Variables
> -----------------------
> One additional behavior I've noticed is that the GRASS_WIDTH,
> GRASS_HEIGHT, and other xwindow variables are not being picked up in
> the GRASS environment. If I start d.mon, the values are not
> recognized. I went into the grass5 script that is installed in
> ${prefix}/bin and added the following:

GRASS_{WIDTH,HEIGHT} are environment variables, not GRASS variables. 
Setting them in $GISRC won't work, and isn't meant to work (it would
have worked in previous versions because the contents of $GISRC were
copied into the environment; that no longer happens).

-- 
Glynn Clements <glynn.clements at virgin.net>