SERRER@NRCM3.NRC.CA (Martin Serrer - Systems Analyst) (03/02/90)
Hi,
Has anyone managed to install the Xw widget library on an IRIS. I am running
IRIX 3.2 and I have the SGI X development environment.
When I try to build the widget library, at the point where the file Manager.c
is compiled I get the most helpful message...
Fatal error in /usr/include/uopt - core dumped
Signal 139
Isn't UNIX wonderful :-( What the %&%^#!@$@#^*& does one do at such a
point.
I apologize for the tone but I have been fighting with this code for 3 days
and this is just the latest problem I've encountered.
Martin
+-----------------------------------------------------------------------------+
| Martin Serrer Systems Lab., Bldg. M3, Montreal Rd.|
| 613-993-9442 (Bell) National Research Council of Canada,|
| serrer@syslab.nrc.ca (BITNET) Ottawa, Ontario, Canada K1A-0R6 |
+----------Software Rusts...------------------Rust never Sleeps...------------+moraes@cs.toronto.edu (Mark Moraes) (03/02/90)
SERRER@NRCM3.NRC.CA (Martin Serrer - Systems Analyst) writes: > Has anyone managed to install the Xw widget library on an IRIS. I am running >IRIX 3.2 and I have the SGI X development environment. Yes, but not with the SGI X development environment. We built MIT X. MIT X.V11R4 comes with an sgi config file and a much fixed set of Xw widgets in contrib/toolkits/Xw. (WARNING: Do not try to use the Xw widgets in the MIT R3 source with the MIT R3 includes or libXt. It doesn't work, for reasons documented in the R3 doc/tutorials/r3widgets.ms document) > When I try to build the widget library, at the point where the file Manager.c >is compiled I get the most helpful message... >Fatal error in /usr/include/uopt - core dumped > Signal 139 /usr/include? /usr/lib/uopt is the MIPS optimizer, which dumps core on a few files in Xw when compiling with -O. (if it's any consolation, it did this on the DECstation 3100 as well) Compile with make -k then go back to Xw and use make CDEBUGFLAGS=-g to recompile the files that caused cc to dump core. Alternatively, you can opt for the lower performance, high correctness side of the world and compile everything without -O, which is what we did. (When compiling something the size of X, even looking through the output of the make to find errors is exhausting. And then there's programs that work when compiled without -O but die when compiled with -O, like contrib/clients/pbmplus/ppm/giftoppm.c) Perhaps someone from SGI/MIPS would care to compile all of X.V11R4 before Irix3.3 is released -- X makes a great compiler test suite. > Isn't UNIX wonderful :-( Yes.