This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: PR libstdc++/6750

From: Paolo Carlini <pcarlini at unitus dot it>
To: Peter Schmid <schmid at snake dot iap dot physik dot tu-darmstadt dot de>
Cc: libstdc++ at gcc dot gnu dot org, bkoz at redhat dot com
Date: Tue, 21 May 2002 10:09:02 +0200
Subject: Re: PR libstdc++/6750
References: <Pine.LNX.4.30.0205210219380.14745-100000@snake.iap.physik.tu-darmstadt.de>

Peter Schmid wrote:

>I investigated PR libstdc++/6750 and found out that reverting
>the patch
>2002-05-18  Benjamin Kosnik  <bkoz@redhat.com>
>
>	PR libstdc++/6518
>	* include/bits/ostream.tcc (ostream::operator<<(const char*)): Fix
>	for null case.
>	(ostream::operator<<(const _CharT*)): Same.
>	(ostream<char>::operator<<(const char*)): Same.
>	* testsuite/27_io/ostream_inserter_char.cc (test07): Add test.
>fixes the io related problems.
>
>Therefore, I ask to revert this patch which causes severe io related
>problems.
>
Hi Peter,

indeed, the testcase you submitted as part of libstdc++/6750 is 
sensitive to this recent change, since, after terminating a getline with 
a single delim char ('\n') the buffer turns out to be empty and triggers 
the new code checked in by Benjamin.

If I understand well libstdc++/6518 commit, the new code is tought to 
deal more gracefully with the undefined behaviour (avoiding a core 
dump), but still sets the badbit. A few other implementations, however, 
don't set it.

Benjamin, what about not setting the badbit in v3 too?

Ciao, Paolo.

References:
- PR libstdc++/6750
  - From: Peter Schmid

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]