🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: gnu.gcc.help
Article #42390

line pointer for undefs [Was: -g option]

#42390
From: Ian T Zimmerman
Date: Mon, 18 Dec 1995 00:00
45 lines
2638 bytes
In article <43282f$84m@ncar.ucar.edu> jadams@sage.cgd.ucar.edu (James Adams) writes:

> 
> In article <4322kd$1am@ncar.ucar.edu>, jadams@sage.cgd.ucar.edu (James Adams) writes:
> >Hello,
> >
> >	I'd like to build a version of gcc which supports the creation of
> >a symbol table for debugging.  Version 2.6.3 as well as 2.7.0 both appear
> >not to support this.  What is the latest version which does support this
> >(the -g option) ?
> >
> 
> 	I've think I figured this one out myself.  I am trying to
> compile gcc on an SGI, IRIX 5.3.  Apparently the GNU C compiler
> does not support the -g option (symbol table creation) on the
> SGI's.  As far as I can tell the reason for this is that gcc can
> only produce debugging information for certain types of native
> formats, and the format used by IRIX, ELF, is not one of these.
> 

I am on a Linux 1.3.8 system with ELF as native object format, with
gcc 2.7.0 and friends. -g works OK for me, it generates stabs format
info (I think). I think that the executable format and the debug info
format are to some degree independent of each other. Also, I think it
was one of the purposes of ELF to increase that degree of
independence.

So far, just so much babble (I hope some expert can correct me). Now,
however, I have a question of my own. Before this upgrade, I had a.out
Linux version 1.0.9 and 1.2.5, with gcc 2.5.8. It used to be that when
I compiled and linked with -g and I had undefined external references,
the link phase error message contained the _source line_ of the
reference. That meant I could jump right to the problem spot in emacs
the same way as with compile time errors.

In fact I remember using this as an example of the superiority of free
tools over behemoth `IDEs'.

But now after this move to ELF and 2.7.0, this feature seems to be
gone! Why, or am I missing something?
-- 
Ian T Zimmerman            +-------------------------------------------+
P.O. Box 13445             I    With so many executioners available,   I
Berkeley, California 94712 I suicide is a really foolish thing to do.  I
USA  <itz@rahul.net>       +-------------------------------------------+
Technical Details
Message-ID: <ITZ.95Dec18092520@kronstadt.rahul.net>
Path: arc1-ca.pugleaf.net!unknown!not-for-mail
References: <4322kd$1am@ncar.ucar.edu> <43282f$84m@ncar.ucar.edu>

Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 11318c,a5fcea5934855cf3
X-Google-Attributes: gid11318c,public
From: itz@rahul.net (Ian T Zimmerman)
Subject: line pointer for undefs [Was: -g option]
Date: 1995/12/18
Message-ID: <ITZ.95Dec18092520@kronstadt.rahul.net>
X-Deja-AN: 122223934
distribution: world
references: <4322kd$1am@ncar.ucar.edu> <43282f$84m@ncar.ucar.edu>
organization: a2i network
newsgroups: gnu.gcc.help