home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.unix.aix
- Path: sparky!uunet!newsgate.watson.ibm.com!yktnews!admin!yktnews!prener
- From: prener@watson.ibm.com (Dan Prener)
- Subject: Re: xlc -g option
- Sender: news@watson.ibm.com (NNTP News Poster)
- Message-ID: <PRENER.92Sep12032836@encap2.watson.ibm.com>
- In-Reply-To: jpennell@axion.bt.co.uk's message of 11 Sep 92 13:58:42 GMT
- Date: Sat, 12 Sep 1992 08:28:36 GMT
- Disclaimer: This posting represents the poster's views, not necessarily those of IBM
- References: <1992Sep11.145842@axion.bt.co.uk>
- Nntp-Posting-Host: encap2.watson.ibm.com
- Organization: IBM T.J. Watson Research Center, Hawthorne, New York
- Lines: 19
-
- In article <1992Sep11.145842@axion.bt.co.uk> jpennell@axion.bt.co.uk (John Pennell) writes:
-
- > Thanks to all who mailed me regarding the problems I was having using
- > -O on the xlc compiler. A new set of binaries are winging their way
- > to me from IBM. Now, I have another problem. Somewhere I have read
- > that it is inadvisable to use -g with -O. Why? I am asking because I
- > wish to use PURIFY, a memory allocation tool and it would be useful if
- > the -g switch could be used. Am I correct in assuming the only reason
- > it shouldn't be used is because the output (post-optimisation) may not
- > be reliable?
-
- You can use -g and -O together. I do it frequently. However, you should
- be aware of the limitations on debugging such code. You will not be able
- to count on the values of variables displayed at arbitrary points in the
- code, since the variables may be held in registers, rather than in their
- storage locations, or may not exist at all. Values displayed at the
- start of non-inlined functions will be correct.
- --
- Dan Prener (prener@watson.ibm.com)
-