jhc@irwin.uucp (James H. Coombs) (11/02/90)
Thanks! CCROOTDIR was indeed the problem with lex++/yacc++. Has anyone been using dbx successfully? As soon as I load a file, I get dbx: fatal error: I/O error I have CCROOTDIR set. I am explicitly invoking CC/sun4/dbx. --Jim
dwf@hope.lanl.gov (David W. Forslund) (11/02/90)
In article <55044@brunix.UUCP> jhc@irwin.uucp (James H. Coombs) writes:
Thanks! CCROOTDIR was indeed the problem with lex++/yacc++.
Has anyone been using dbx successfully? As soon as I load a file, I get
dbx: fatal error: I/O error
I have CCROOTDIR set. I am explicitly invoking CC/sun4/dbx.
--Jim
The problem is with SunOS4.1. The C++2.0 dbx as provided will not
work under SunOS4.1. The bug is docmented in the RTF, I
believe, and there is a patch provided on the SunOS4.1 tape which
replaces the dbx with one that works.
--
David Forslund
Advanced Computing Laboratory
MS B287
Los Alamos National Laboratory
Los Alamos, NM 87545
Voice:(505) 665-1907
FAX: (505) 665-4939
EMAIL: dwf@lanl.govtruett@cup.portal.com (Truett Lee Smith) (11/02/90)
Dbx frequently gives I/O errors while reading either the symbols or when doing the initial processing of a core. We found it necessary to make sure that the workstation server had allocated swap space for the workstation running dbs which was at least four times the size of either the core of the executable. In my case, I was working with an 8 MB executable that had made a 9 MB core image, so a 40 MB swap space was insufficient but 50 MB was. It had to read in 103,000 symbols (C++ generates *LOTS* of symbols) but dbx was able to trace the program after taking about 20 minutes to get set up!!! Truett Smith, Sunnyvale, CA truett@cup.portal.com OR truett@tdd.sj.nec.com