home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: gnu.gdb.bug
- Path: sparky!uunet!cis.ohio-state.edu!zippysun.math.uakron.edu!wurgler
- From: wurgler@zippysun.math.uakron.edu (Tom Wurgler)
- Subject: GDB and F77 on Sun revisited
- Message-ID: <9207211419.AA02532@zippysun.math.uakron.edu>
- Sender: gnulists@ai.mit.edu
- Organization: GNUs Not Usenet
- Distribution: gnu
- Date: Tue, 21 Jul 1992 14:19:22 GMT
- Approved: bug-gdb@prep.ai.mit.edu
- Lines: 63
-
- Hello again.
-
- A thousand pardons. I sent email on July 9th saying that I was having
- trouble with GDB and F77. Well, I was (and I still am) but my
- assessment of the situation was WRONG. Here is the opening salvo from
- my last email:
-
- > I am trying to debug a F77 program on a Sun SPARCstation 1. It pretty much
- > works. There are just certain times when I get an error (in GDB) that I do
- > not know how to get around. GDB seems to have no problems until I attempt
- > to reference a variable that is used in a subroutine call to an outside
- > package (such as SunPHIGS). Here is a bit to show you what I mean:
-
- And I went on to say that I could not print variables used in calls to my
- outside package (here SunPHIGS). Well that is true but not completely true.
- I saw last week that GDB v4.6 was released and so I thought that I would grab
- it and see if it would fix my problems. When it too failed I investigated
- even further.
-
- Here is my new assessment of what is happening. I had always noticed that,
- for the most part, I couldn't print variables that were used in the calls to
- the SunPHIGS package. I was doing more testing and noticed that I got
- "$1 = Cannot access memory at address xxxxxxxx." when I did "info locals".
-
- In almost every subroutine in which I have these calls to the SunPHIGS
- package, I use an INCLUDE file that come with SunPHIGS. This INCLUDE file is
- made up of INTEGER and PARAMETER statements (1107 lines worth). I found that
- if I didn't have the INCLUDE file, "info locals" worked. When I used the
- INCLUDE file, "info locals" choked. So I jammed the contents of the INCLUDE
- file into my code, commented out all of the PARAMETER statements and GDB
- didn't choke. So then I uncommented out groups of PARAMETER statements until
- GDB failed. It appears now that when I reach a certain limit of variables
- defined by PARAMETER statements, that GDB chokes (ie, "info locals" chokes)
- with the "Cannot access memory at address..." error.
-
- Here is my environment:
-
- Sun workstation, SPARCstation 1.
-
- SunOS 4.1.1 (SunOS Release 4.1.1 (GENERIC) #1: Thu Oct 11 10:25:14 PDT 1990)
-
- GDB versions 4.4 and 4.6. I used "sun4os4" when I configured.
-
- I used Sun's cc the first time. Then I ran
- "make CC=gcc" somewhere at the top and rebuilt.
- I used gcc version 2.2.2.
-
- F77 version 1.4. I usually use the "-g -C" switches but I get the same
- results using just "-g".
-
- The "outside" package that I am using is SunPHIGS version 1.3.
-
- Thanks for any light that you can shed upon this...
-
- I apologize again for my earlier stupidity. I hope that I didn't cause too
- much trouble.
-
- Please let me know if you need more information or if there is anything that I
- can do to clarify what I have said here.
-
- Bill Benedetto
- wurgler@zippysun.math.uakron.edu
-
-