|
|
Magic Mailing List |
|
From: Anand K Ojha (aojha AT du DOT edu) Date: Fri Mar 26 2004 - 20:05:29 EST
A few days ago, I saw the this new version, and loaded and installed, and was very pleased to see that we didn't have to modify a single line for installing it on our Sparc/Solaris 8. Congratulations Tim, et al. In previous versions, we had to do two modifications in the mak.defs file. 1. changing the line 'CC = cc' to 'CC = gcc' 2. providing full path for 'ar' Yes, as mentioned by Tim in one of his posts, the Solaris ld doesn't work. In fact, for all previous versions, we had to first install GNU make, ld, m4, .. (binutils )from sunfreeware.com, and rename sun's utilities (ld to ld_sun) to be able to compile Magic. We currently installed Magic 7-2.64 using all GNU utilities, so we don't know if Solaris utilities work on this newer version or not; but we don't have a need to try this out. However, the tcl version of IRSIM (in Magic 7-2.64) continues to cause segmentation fault when we use the analyzer, but choosing the non-tcl/tk/blt options during installation lets us use the stand-alone non-tcl IRSIM. It's only that the non-tcl IRSIM is not well integrated, but since it works well, we have postponed looking into the tcl-IRSIM seg fault after our quarter ends in May. Our students are using tcl-Magic and non-tcl IRSIM from Magic 7-2.64 version, and we don't want to disturb our system. I would also like to mention that our tclspice also seg faults while ngspice, which is non-tcl, works fine. So, there appears something not right with our tcl/tk's interaction with solaris or with some of the '.so' files. This seems odd because John Osterhaut, the inventor of tcl/tk, later worked for Sun Microsystems. Anyhow, if someone has been able to make tcl-ed IRSIM or tclspice work on solaris, please let me know. Once again, thanks to all who are working so hard to make all these CAD tools available to us. --Anand Ojha, University of Denver.
|
|
|
|