NAME

SYNOPSIS

DESCRIPTION

Typical usage is:

someprogram | togeomview    (to send commands) or
someprogram-generating-OOGL-data | togeomview -g  (to send geometry) 

Togeomview uses a named pipe in the directory /tmp/geomview to communicate with geomview. If unspecified, the pipe's default name is "OOGL". When sending geometry (-g), a geomview object with the same name as the pipe appears in geomview's object browser.

By default, when no suitable copy of geomview is running, togeomview invokes "geomview" with arguments specifying the appropriate named pipe. A different command may be specified as in:

togeomview OOGL gv -wpos 300x300 -c my_startup_script

which communicates through a pipe named OOGL, and (if necessary) invokes the given gv command. The pipe name is required if a command is specified.

After togeomview has created it, the named pipe may be written as an ordinary file. For example, one could use

togeomview pipename < /dev/null

to invoke a listening copy of geomview, and then run a program which simply wrote to /tmp/geomview/pipename.

FILES

BUGS

If no copy of geomview is reading from the pipe, or if geomview gets far enough behind, a program writing data to ``togeomview'' will be forced to block after sending a few kilobytes.

Because of the buffering in the pipe, the sender may be substantially ahead of the geomview display.

If geomview exits, the sending program receives a write-on-broken-pipe (SIGPIPE) signal, which will kill it unless measures are taken to catch or ignore that signal.

Only one copy of geomview can read from a given pipe at a time. If a second copy attempts to read from it, both will probably fail. It's fine to have multiple copies of geomview reading from different pipes.

Note that togeomview will invoke geomview if no extant copy is listening to the relevant pipe; it can't connect to an existing copy of geomview started by other means.

SEE ALSO