Re: New failing tests for g++

[Robert Lipe <robertl at dgii dot com>]

Klaus-Georg Adams wrote:

> I have a couple of simple testcases for egcs which fail for me
> (i686-pc-linux-gnulibc1). I had submitted two of them already to
> egcs-bugs, but you didn't include them in g++.robertl. Maybe you can
> take a look.

I only harvested "obvious" failures from the epoch of EGCS through
Janruary of this year.  I still have a couple hundred (sigh) messages in
that batch that may or may not represent reasonable cases to be boiled
down.  Of course, people didn't stop submitting bugs at that time,
either, so there may be some submitted since then.

I've added comments showing the submitter, EGCS release, message-id, and
triplet where the problem was seen in my local copy.  (Can you tell I'm
raising the standards a little? :-) They will become eb129.C and eb130.C
unless someone else has beaten me to those numbers.

But neither of these tests compile for me.   I'd like to work that out before
doing the commit.   They both fail on my Linux system (RH 5, probably 5.1 by
the end of the week) with errors of the form:

Executing on unix: /home/robertl/tmp/negcs/gcc/testsuite/../xgcc -B/home/robertl
/tmp/negcs/gcc/testsuite/../ /net/rjlhome/home/play/egcs/gcc/testsuite/g++.old-d
eja/g++.robertl/eb129.C  -ansi -pedantic-errors  -I/net/rjlhome/home/play/egcs/l
ibio -I/home/robertl/tmp/negcs/libraries/libio -I/net/rjlhome/home/play/egcs/lib
stdc++ -I/net/rjlhome/home/play/egcs/libstdc++/stl -I/home/robertl/tmp/negcs/gcc
/testsuite/../include -c  -o /home/robertl/tmp/negcs/gcc/testsuite/eb129.o
compiler exited with status 1
output is:
In file included from /net/rjlhome/home/play/egcs/libstdc++/stl/stl_alloc.h:73,
                 from /net/rjlhome/home/play/egcs/libstdc++/stl/list:31,
                 from /net/rjlhome/home/play/egcs/gcc/testsuite/g++.old-deja/g++
.robertl/eb129.C:5:
/usr/include/pthread.h:438: parse error before `*'
/net/rjlhome/home/play/egcs/gcc/testsuite/g++.old-deja/g++.robertl/eb129.C: In f
unction `int main()':
/net/rjlhome/home/play/egcs/gcc/testsuite/g++.old-deja/g++.robertl/eb129.C:14: p
arse error before `>'
/net/rjlhome/home/play/egcs/gcc/testsuite/g++.old-deja/g++.robertl/eb129.C:20: `
it' undeclared (first use this function)
/net/rjlhome/home/play/egcs/gcc/testsuite/g++.old-deja/g++.robertl/eb129.C:20: (
Each undeclared identifier is reported only once





> 
> Thanks, kga
> -- 
> -------------------------------------------------------------------------
> Klaus-Georg Adams        Email: Klaus-Georg.Adams@chemie.uni-karlsruhe.de
> Institut f. Anorg. Chemie, Lehrstuhl II            Tel: 49(0)721 608 3485
> Universität Karlsruhe, D-76128 Karlsruhe
> -------------------------------------------------------------------------
> ---------snip----------------
> // Build don't link:
> // Gives ICE 109
> int main()
> {
> 	try {
> 	}
> 	catch (bad_alloc) { // ERROR - parse error
> 		return 1;
> 	}
> 	return 0;
> }
> 
> ---------snip----------------
> // Build don't link:
> // Gives ICE
> #include <list>
> #include <functional>
> #include <algorithm>
> #include <cassert>
> int main()
> {
> 	list<int> l;
> 	l.push_back(1);
> 	l.push_back(2);
> 
> 	list<int>::iterator it = 
> 		find_if( l.begin(), l.end(),
> 			 // This is a typo, it should be bind2nd, but an
> 			 // ICE is not a very helpful diagnostic!
> 			 binder2nd( equal_to<int>(), 2 ) ); // ERROR - 
> 	assert( *(it) == 2 );
> }
> 
> ---------snip----------------
> // Build don't link:
> // Gives spurious warnings when compiled with -Wall
> 
> #include <hash_set>
> std::hash_set<int> foo;
> ---------snip----------------
> 

-- 
Robert Lipe       http://www.dgii.com/people/robertl       robertl@dgii.com
              (WEB ADDRESS MAY BE TEMPORARILY UNAVAILABLE)