create_fifos.sh, monitors, NVIZ woes...
neteler at geog.uni-hannover.de
Mon Nov 8 04:09:57 EST 1999
> There seems to be some discrepencies between what the instructions say
> for the directory tree name (/usr/local/grass5 - README at Baylor ftp
> site) and what some programs expect (/usr/local/grass5.0b).
In my opinion we should call this standard path:
Is that o.k. with everybody? Then we change it generally.
> For instance, create_fifos.sh has these lines:
> # create_fifos.sh.in
This is changed automatically using my install script...
It should be in the same directory as the binaries (at least
at Hannover site).
> When I changed this line, and remade the fifos because I couldn't start
> monitors and thought that might be the problem, I discovered that
> somewhere the monitors also expected to live under
> /usr/local/grass-5.0b. I haven't tracked down where this expectation is
Same thing: This is changed automatically using my install script...
> coming from, but I suspect it's hard coded into the files. So, I renamed
No, it isn't. Just a path variable in etc/monitorcap. But read above.
> Now, if I could get NVIZ to work...It dies at some point when talking to
> tk.tcl. Here's the relevant part of strace, if it'll help.
> open("/usr/lib/tk8.0/tk.tcl", O_RDONLY) = 5
> fcntl(5, F_SETFD, FD_CLOEXEC) = 0
> ioctl(5, TCGETS, 0xbfffda84) = -1 ENOTTY (Inappropriate ioctl
> for dev
> read(5, "# tk.tcl --\n#\n# Initialization s"..., 4096) = 4096
> read(5, "vent add <<Clear>> <Clear>\n }"..., 4096) = 1601
> close(5) = 0
> brk(0x8110000) = 0x8110000
> --- SIGSEGV (Segmentation fault) ---
> +++ killed by SIGSEGV +++
> The problem seems to be that ioctl(5,TCGETS,...). Perhaps a system bug?
Sorry, here it works. Anybody else having such problems?
More information about the grass-user