|
|
Magic Mailing List |
|
From: cfk (cfk AT pacbell DOT net) Date: Mon Mar 24 2003 - 22:00:24 EST
I'll let the discussion continue on the technology file, euclidean vs
manhattan and the like and go on to another point with the intention of
coming back to the Mosis issue a little later. Magic-7.2.31 compiles and
runs just fine with 'make tcl' & 'make install-tcl'. I am able to do a 'load
/usr/lib/libspice' and a ":spice::version" just fine. Good job, Tim.
I have a little further on BLT. From a bash prompt (magic not loaded), I can
do:
wish
package require BLT
namespace import blt::*
set x {1 2 3}
set y {4 5 6}
graph .g -title "killroy was here"
.g element create x -label "x killroy data" -xdata $x -ydata $y
pack .g
and get a graph just fine. Once magic has started up, I get a blank tclspice
graph with y axis going from 0 to 1 in 0.1 increments and x axis going from
0 to 1 in 100.0m increments, but I cant write anything to either that graph
or a new graph from the tcl prompt in magic. It also seems tclspice is
giving me an error to the bash prompt screen that is only seen when magic
exits that is coming from inside its graph.c when it cannot get access to
the graph object. An asciiplot works fine from within magic, just the BLT
graphics seem thwarted at this moment by some logical battle going on
between a couple of subroutines.
The good news is spice works from a magic tcl prompt and will extract
geometry from a magic layout of an inverter, and gate and nor gate. Again, I
am concentrating on the standard cells from
http://www.rsl.ukans.edu/~mlinhard/magic until the Mosis technology file
issue gets a little further along.
|
|
|
|