From:	pavenis AT lanet DOT lv
To:	salvador <salvador AT inti DOT gov DOT ar>, djgpp-workers AT delorie DOT com
Date:	Sat, 29 Sep 2001 15:17:22 +0300
MIME-Version:	1.0
Subject:	Re: GCC 3.01 Streams problem
Message-ID:	<3BB5E602.16332.35BCB7@localhost>
In-reply-to:	<3BB4D5C1.12443.6387C@localhost>
References:	<3BB47CE0 DOT 106BF5AF AT inti DOT gov DOT ar>
X-mailer:	Pegasus Mail for Win32 (v3.12c)
Reply-To:	djgpp-workers AT delorie DOT com
Errors-To:	nobody AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

On 28 Sep 2001, at 19:55, pavenis AT lanet DOT lv wrote:

> Verified with following versions:
> 
> i586-pc-msdosdjgpp	gcc-3.0.1 release
> i586-pc-msdosdjgpp	gcc-3.1 20010927 (experimental)
> i686-pc-linux-gnu			      gcc-3.0.2 20010927 (prerelease)
> 
> This problem were present in all 3 cases (so also under Linux).
> So I think it's a serious regression against gcc-2.95.X not only
> for DJGPP but also under Linux

Looked in draft of C++ standard
	|ftp://research.att.com/dist/c++std/WP/

It's said there (27.6.1.3):
	If the function inserts no characters, it calls 		
	setstate(failbit), which may throw
	ios_base::failure (27.4.5.3).

I think it's the same behaviour as we see.

Maybe it's better to use getline instead of get and clean
trailing LF after reading (getline should not have this problem
as it reads delimiting char, so no zero bytes on empty lines)

Andris

- Raw text -