This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: What does "internal error #20000524"?
- From: Joe Buck <jbuck at synopsys dot COM>
- To: zack at codesourcery dot com (Zack Weinberg)
- Cc: rodrigc at mediaone dot net (Craig Rodrigues), gcc at gcc dot gnu dot org
- Date: Wed, 23 Jan 2002 09:50:04 -0800 (PST)
- Subject: Re: What does "internal error #20000524"?
> On Tue, Jan 22, 2002 at 08:08:05PM -0500, Craig Rodrigues wrote:
> >
> > What do we gain by using the my_friendly_abort() macro vs. doing
> > just a plain abort() and getting the same "Internal compiler error"
> > messages in the C part of gcc?
>
> In my opinion, not much, and new code should use plain abort().
The advantage of my_friendly_abort is that the user gets an indication
of the point in his/her code where the compiler blew up. Often it is
possible to work around the compiler bug by rewriting the code at that
point in the program.
It is true that the indication of the location in the *compiler* of the
crash doesn't add much, since it's only of use to the compiler developer
who can find it just as easily if there is an abort. But the location
in the user's code is another matter, and saved me much work when I
used g++ extensively in the early 90s (when it was a seething mass of
bugs that ICEd if you looked at it funny).
Now, if we can make ICEs so rare that they almost never occur, it doesn't
matter.