home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: gnu.gcc.help
- Path: sparky!uunet!convex!linac!pacific.mps.ohio-state.edu!cis.ohio-state.edu!che.utexas.EDU!jwe
- From: jwe@che.utexas.EDU (John Eaton)
- Subject: Re: Segmentation Fault
- Message-ID: <9211161710.AA03192@schoch.che.utexas.edu>
- Sender: daemon@cis.ohio-state.edu
- Organization: GNUs Not Usenet
- References: jwe@che.utexas.EDU (John Eaton)
- Distribution: gnu
- Date: Mon, 16 Nov 1992 17:10:28 GMT
- Lines: 32
-
- : I do not believe that compiler crashes usually have anything to do
- : with a particular line in the source.
-
- This may now be true for most of the gcc crashes, but with g++, which
- seems to crash a lot more frequently than gcc, I have found that many
- of the crashes can be avoided by changing the user's code. Most of
- the time, it is something very simple.
-
- Even just knowing that the compiler is working on something related to
- a given line of source would be much more helpful than getting
- `cc1plus got fatal signal xxx', especially when there may be thousands
- of lines of code in the current file.
-
- I would even be happy if the message said something like:
-
- cc1plus: I don't think line 13 has anything to do with what's
- cc1plus: going to happen next, but I was just working on
- cc1plus: something vaguely related to that line.
- cc1plus: Now I'm bailing out from line 1313 of cp-tree.c.
-
- I agree that the best thing is to fix the compiler so that it doesn't
- crash. Unfortunately, many of us are not able to do that, and we are
- trying to use these compilers for real projects. Having cc1plus tell
- me what line(s) it is working on when it crashes would be helpful.
-
- If RMS is unwilling to add this feature to cc1plus, does anyone out
- there have patches to do it?
-
- --
- John W. Eaton | 4.3BSD is not perfect. -- Leffler, et al. (1989).
- jwe@che.utexas.edu |
-
-