This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: Internal compiler error (cygwin)
On Sun, Apr 01, 2001 at 01:29:48AM +0100, Dave Korn wrote:
>
>----- Original Message -----
>From: "Zack Weinberg" <zackw@stanford.edu>
>To: "Thomas Henlich" <henlich@mmers1.mw.tu-dresden.de>
>Cc: <gcc-bugs@gcc.gnu.org>
>Sent: Friday, March 30, 2001 7:56 AM
>Subject: Re: Internal compiler error (cygwin)
>
>
>> On Thu, Mar 29, 2001 at 05:46:53PM +0200, Thomas Henlich wrote:
>> > I got an error message about 'Internal compiler error' using gcc
>2.95.3-2
>> > on cygwin (1.1.8, Windows NT-4.0 i386)
>
>> There isn't an abort on line 6111 of cccp.c in my source tree, which
>> makes me wonder if your copy of the program has been modified from the
>> official 2.95.3. Can any Cygwin hackers comment please?
>>
>> I wonder why it says 'in file included from spvw.d' when the source
>> file is spvw.c... That suggests memory corruption to me.
>
> I'm not a cygwin hacker, but I know this much: the cygwin version of gcc
>is indeed modified, and it's up to version 2.95.2-5 or perhaps even -6 by
>now, so it would be worth seeing if this bug has been fixed in the more
>recent versions.
The last 2.95.2 version of gcc was -9.
The current cygwin version of gcc is actually 2.95.3-1. 2.95.3-2 is a
"testing release". There are no bug fixes in this version beyond
reapplied patches from Mumit Khan that were in the 2.95.2 releases and
my changes in cygwin.h for getting -mno-cygwin to work better.
There are additions in cccp.c, though, so the line numbers won't match.
I haven't added an abort, though.
Here are the lines in question:
/* Scan thru the replacement list, ignoring comments and quoted
strings, picking up on the macro calls. It does a linear search
thru the arg list on every potential symbol. Profiling might say
that something smarter should happen. */
if (end < buf)
abort (); <<<---HERE
This is in collect_expansion ().
I wonder if this is a #!!&# CRLF problem.
cgf