Wallace, Beverly T wrote:
I recently installed GRASS 5.0.0. I configured it for fifos and had two
problems:
Just out of curiosity, was there any particular reason for using
FIFOs? Is this a system which simply doesn't support sockets, or is
there a problem with the socket code?
(1) Configuring with "--enable-fifo" is not enough, "--disable-socket" is
also needed. Is that how it is supposed to work? The configure help says
that "--enable-fifo" will use the FIFO XDriver instead of sockets, which
suggests that sockets are automatically disabled.
OK; this can be fixed easily enough.
(2) The monitorcap file did not work until I added full paths.
This is due to:
date: 2002/01/04 17:04:17; author: andreas; state: Exp; lines: +14 -14
removed absolute paths in monitorcap file
in:
src/display/devices/HTMLMAP/Gmakefile
src/display/devices/HTMLMAP/moncap.entry
src/display/devices/PNGdriver/Gmakefile
src/display/devices/PNGdriver/moncap.entry
src/display/devices/monitorcap/Gmakefile
src/display/devices/monitorcap/monitorcap
Andreas: was there a specific reason for this change? It won't work
with FIFOs (and the field is ignored when using sockets).
There are two solutions here:
1. (the easy way): revert the change.
2. (the hard way): modify the code to prepend $GISBASE to the paths.
This could be done either in src/libes/raster/io.c (for clients) and
src/display/devices/lib/connect_fifo.c (for drivers), or by modifying
R_parse_monitorcap().
The only reason to choose option 2 is so that the GRASS directory can
be moved easily.
--
Glynn Clements <glynn.clements@virgin.net>