MAGIC Magic Mailing List
 
 

From: cfk (cfk AT pacbell DOT net)
Date: Sun Jun 08 2003 - 15:44:17 EDT

  • Next message: cfk: "tclspice-0.2.10 and magic"

    Well, of course, the moment I send out one of these whining e-mails, I end
    up answering at least part of the question immediately. It turns out that
    /usr/lib/tclX8.3/pkgIndex.tcl in my RH8 installation has a statement that
    says "ifneeded Tclx 8.3 load ../libtclx8.3.so", so I went up a level and
    found no libtclx8.3, so I then made a soft link as follows:
    
        ln libtclx.so.0 libtclx8.3.so
    
    What puzzles me is that I shouldnt have to do that sort of thing as I would
    have expected RH8 to set up these correctly.
    
    In any case, I can now go:
    
    magic -w invlay        #a basic cmos inverter
    irsim
    package require spice
    spice_init_gui testx_trans.cir    #a circuit powering & pulsing invlay's
    extraction
    spice::run
    spicewish::plot in out
    graphnode in
    graphnode out 1
    
    
    Now I can admire the new spicewish "STOP, Run, Plot, Edit, Save Pro, Quit"
    window. Thats the good news. The bad news, is that pressing STOP, says
    "spice not running" and the magic layout, irsim (yes I have an irsim graph
    because of graphnode) and plot windows cease updating and magic is frozen.
    
    So as long as I dont click the spicewish window buttons, things progress,
    otherwise they are a bit dormant. I dont really know what to do with the
    spicewish menu yet, so I'll ignore it until it bytes me in the bottom
    (later).
    
    Charles
    
    
    ----- Original Message -----
    From: "cfk" <cfk AT pacbell DOT net>
    To: <ngspice-users AT lists DOT sourceforge.net>
    Cc: <magic-dev AT csl DOT cornell.edu>
    Sent: Sunday, June 08, 2003 12:06 PM
    Subject: tclspice-0.2.9 after DAC
    
    
    > So, DAC is over and I'm trying to get things working again with magic and
    > tclspice. I must be dense or something, as I cannot seem to get the plot
    > running in magic now. Let me start with our last e-mail exchange back a
    > month or so ago, which is below.
    >
    > Basically, from within magic, I can go "package require spice" and the tk
    > shell returns 0.2.9 as I would expect. But the troubles begin with the
    next
    > line "spice_init_gui example.cir". At this point, I get a message saying
    > that spice couldnt load "/usr/lib/tclX8.3../libtclx8.3.so". Now the
    curious
    > part, is the installation I am testing on has a "/usr/lib/tclX8.3" and
    there
    > is a libtclx.so soft link to libtclx.so.0. This is all default RH8 stuff.
    I
    > am trying to keep my magic & tclspice contained within a default RH8
    > installation and it used to work just fine.
    >
    > I can see that the statement in tclspice_source.tcl at line 22 saying
    > "package require Tclx" appears to be the culprit, but my lack of tcl
    > knowledge precludes me from going much further.
    >
    > On the good news side, I can run my circuit with "spice::run" from within
    > magic and I can do a "spice::asciiplot" from with magic. As you might
    > recall, the issue is plotting from within magic as when I do a
    > "spice::plot", I get a "cant open viewport for graphics" for error.
    >
    > I am basically thinking that tclspice and magic are making two different
    > assumptions about the tcl/tk environment here and we have the two
    > assumptions battling with each other for control of the screen, viewport
    or
    > *something*.
    >
    > I dont want to cause a great amount of effort for anyone, either on the
    > tclspice or the magic side, but I keep coming back to this one and it just
    > seems to be in the way again. I am hoping that we can get the two
    codebases
    > playing a bit more nicely with each other.
    >
    > With highest regards, Charles
    >
    >
    


  •  
     
    Questions? Contact Rajit Manohar
    cornell logo