RE: libstdc++/9533: Regression: Can't read from tty with ifstream

RE: libstdc++/9533: Regression: Can't read from tty with ifstream

[Pétur Runólfsson]

The following reply was made to PR libstdc++/9533; it has been noted by GNATS.

From: =?iso-8859-1?Q?P=E9tur_Run=F3lfsson?= <peturr02@ru.is>
To: <paolo@gcc.gnu.org>,
	<gcc-bugs@gcc.gnu.org>,
	<gcc-prs@gcc.gnu.org>,
	<nobody@gcc.gnu.org>,
	=?iso-8859-1?Q?P=E9tur_Run=F3lfsson?= <peturr02@ru.is>,
	<gcc-gnats@gcc.gnu.org>
Cc:  
Subject: RE: libstdc++/9533: Regression: Can't read from tty with ifstream
Date: Mon, 3 Mar 2003 10:44:37 -0000

 >     Something is wrong, agreed, but I cannot reproduce the
 >     regression thing: for me (*), 3.2.2 asks the number but
 >     then hangs upon <newline>. Same behaviour 3.0.4.
 
 Just to clarify, this is of course correct behaviour since the
 input stream is in fully buffered mode (*). The "hang" happens
 because filebuf::underflow() blocks until either BUFSIZ
 characters are entered or EOF occurs.
 
 Petur
 
 * The reason the test case uses fully buffered is that
 unbuffered doesn't work (libstdc++/9024).

Re: libstdc++/9533: Regression: Can't read from tty with ifstream

[Paolo Carlini]

The following reply was made to PR libstdc++/9533; it has been noted by GNATS.

From: Paolo Carlini <pcarlini@unitus.it>
To: =?ISO-8859-1?Q?P=E9tur_Run=F3lfsson?= <peturr02@ru.is>
Cc: paolo@gcc.gnu.org,  gcc-bugs@gcc.gnu.org,  nobody@gcc.gnu.org, 
 gcc-gnats@gcc.gnu.org,  libstdc++@gcc.gnu.org
Subject: Re: libstdc++/9533: Regression: Can't read from tty with ifstream
Date: Mon, 03 Mar 2003 14:16:37 +0100

 P=E9tur Run=F3lfsson wrote:
 
 >>    Patching: indeed, reverting this hunk of 6746 commit:
 >>   =20
 >>    +
 >>    +	    // Set input to nonblocking for fifos.
 >>    +	    if (__mode & ios_base::in)
 >>    +	      fcntl(this->fd(), F_SETFL, O_NONBLOCK);
 >>    +
 >>   =20
 >>    Fixes the problem without regressing on 6746.
 >>   =20
 >>
 >Interesting. I would have thought that removing this would
 >cause the call to underflow() in basic_filebuf::open() to
 >block, causing the test case to wait for input *before*
 >printing out the prompt (this seems to be the reason that
 >non-blocking input is specified).
 >
 Yes, you are right. My bad.
 I didn't notice that since was using a further reduced :( testcase.
 Also, we got a regression elsewhere by simply reverting that hunk (not=20
 on 6746): see
 the libstdc++ list.
 
 >So far, this is what I think is going on:
 >* non-blocking input is specified so underflow() doesn't
 >  block (which is wrong, underflow() should block).
 >* underflow() must not block because it is called from
 >  open() which must not block.
 >* open() calls underflow() so that in_avail() will return
 >  non-zero after open() (libstdc++/6746).
 >
 Agreed.
 
 >However, I don't see how calling underflow() from open()
 >fixes libstdc++/6746.
 >
 I concur, but in fact it does.
 
 > Unless I am missing something,
 >readsome will return up to BUFSIZ characters, and then
 >return 0 until the end of time.
 > =20
 >
 Paolo.
 

RE: libstdc++/9533: Regression: Can't read from tty with ifstream

[Pétur Runólfsson]

The following reply was made to PR libstdc++/9533; it has been noted by GNATS.

From: =?iso-8859-1?Q?P=E9tur_Run=F3lfsson?= <peturr02@ru.is>
To: <paolo@gcc.gnu.org>,
	<gcc-bugs@gcc.gnu.org>,
	<nobody@gcc.gnu.org>,
	<gcc-gnats@gcc.gnu.org>
Cc: <libstdc++@gcc.gnu.org>
Subject: RE: libstdc++/9533: Regression: Can't read from tty with ifstream
Date: Mon, 3 Mar 2003 13:09:06 -0000

 >     Patching: indeed, reverting this hunk of 6746 commit:
 >    =20
 >     +
 >     +	    // Set input to nonblocking for fifos.
 >     +	    if (__mode & ios_base::in)
 >     +	      fcntl(this->fd(), F_SETFL, O_NONBLOCK);
 >     +
 >    =20
 >     Fixes the problem without regressing on 6746.
 
 Interesting. I would have thought that removing this would
 cause the call to underflow() in basic_filebuf::open() to
 block, causing the test case to wait for input *before*
 printing out the prompt (this seems to be the reason that
 non-blocking input is specified).
 
 So far, this is what I think is going on:
 * non-blocking input is specified so underflow() doesn't
   block (which is wrong, underflow() should block).
 * underflow() must not block because it is called from
   open() which must not block.
 * open() calls underflow() so that in_avail() will return
   non-zero after open() (libstdc++/6746).
 
 However, I don't see how calling underflow() from open()
 fixes libstdc++/6746. Unless I am missing something,
 readsome will return up to BUFSIZ characters, and then
 return 0 until the end of time.
 
 Petur

Re: libstdc++/9533: Regression: Can't read from tty with ifstream

[paolo]

Synopsis: Regression: Can't read from tty with ifstream

Responsible-Changed-From-To: unassigned->paolo
Responsible-Changed-By: paolo
Responsible-Changed-When: Mon Mar  3 12:25:22 2003
Responsible-Changed-Why:
    Patching: indeed, reverting this hunk of 6746 commit:
    
    +
    +	    // Set input to nonblocking for fifos.
    +	    if (__mode & ios_base::in)
    +	      fcntl(this->fd(), F_SETFL, O_NONBLOCK);
    +
    
    Fixes the problem without regressing on 6746.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=9533

RE: libstdc++/9533: Regression: Can't read from tty with ifstream

[Pétur Runólfsson]

The following reply was made to PR libstdc++/9533; it has been noted by GNATS.

From: =?iso-8859-1?Q?P=E9tur_Run=F3lfsson?= <peturr02@ru.is>
To: <paolo@gcc.gnu.org>,
	<gcc-bugs@gcc.gnu.org>,
	<nobody@gcc.gnu.org>,
	=?iso-8859-1?Q?P=E9tur_Run=F3lfsson?= <peturr02@ru.is>,
	<gcc-gnats@gcc.gnu.org>
Cc:  
Subject: RE: libstdc++/9533: Regression: Can't read from tty with ifstream
Date: Sun, 2 Mar 2003 18:23:24 -0000

 >     Something is wrong, agreed, but I cannot reproduce the
 >     regression thing: for me (*), 3.2.2 asks the number but
 >     then hangs upon <newline>. Same behaviour 3.0.4.
 
 Yes, that is the same behaviour I get (with 3.2.1). However,
 mainline fails in a different manner (the program just exits).
 I believe the culprit is this code in basic_file_stdio.cc
 (__basic_file<char>::open), which is not present on the 3.2
 branch:
 
 #if defined (F_SETFL) && defined (O_NONBLOCK)
 	    // Set input to nonblocking for fifos.
 	    if (__mode & ios_base::in)
 	      fcntl(this->fd(), F_SETFL, O_NONBLOCK);
 #endif
 
 This causes fread to return 0 immediately instead of waiting
 for input, which in turn causes underflow to return eof().
 
 Also, I think this counts as a regression, since the hang
 (when compiled with 3.2.x) can be avoided by setting the
 buffer size to 1.
 
 Petur

Re: libstdc++/9533: Regression: Can't read from tty with ifstream

[paolo]

Synopsis: Regression: Can't read from tty with ifstream

State-Changed-From-To: open->analyzed
State-Changed-By: paolo
State-Changed-When: Sun Mar  2 18:08:02 2003
State-Changed-Why:
    Something is wrong, agreed, but I cannot reproduce the
    regression thing: for me (*), 3.2.2 asks the number but
    then hangs upon <newline>. Same behaviour 3.0.4.
    
    (*) x86, glibc2.3.1, linux2.4.20.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=9533

libstdc++/9533: Regression: Can't read from tty with ifstream

[peturr02]

>Number:         9533
>Category:       libstdc++
>Synopsis:       Regression: Can't read from tty with ifstream
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Feb 01 12:16:00 UTC 2003
>Closed-Date:
>Last-Modified:
>Originator:     peturr02@ru.is
>Release:        gcc version 3.3 20030106 (prerelease)
>Organization:
>Environment:
Red Hat Linux 8.0
>Description:
If /dev/tty is opened with an ifstream, nothing can be read. The attached program exits immediately when compiled with gcc-20030106, but outputs the prompt and then waits for input when compiled with gcc-3.2.1.

[petur@interim ttybug]$ g++321 -v
Reading specs from /usr/local/lib/gcc-lib/i686-pc-linux-gnu/3.2.1/specs
Configured with: ../gcc-3.2.1/configure --program-suffix=321 --enable-threads=posix --enable-version-specific-runtime-libs --enable-languages=c,c++
Thread model: posix
gcc version 3.2.1
[petur@interim ttybug]$ g++0106 -v
Reading specs from /usr/local/lib/gcc-lib/i686-pc-linux-gnu/3.3/specs
Configured with: ../gcc-20030106/configure --program-suffix=0106 --enable-threads=posix --enable-version-specific-runtime-libs --enable-languages=c,c++ --enable-clocale=gnu
Thread model: posix
gcc version 3.3 20030106 (prerelease)
[petur@interim ttybug]$ rm ttybug && make CXX=g++321
g++321   -static  ttybug.cc   -o ttybug
[petur@interim ttybug]$ ./ttybug < /dev/zero > /dev/null
Enter a number: 4
You entered: 4
[petur@interim ttybug]$ rm ttybug && make CXX=g++0106
g++0106   -static  ttybug.cc   -o ttybug
[petur@interim ttybug]$ ./ttybug < /dev/zero > /dev/null
Enter a number: You entered: -1073743396
>How-To-Repeat:
See attachment.
>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted:
----gnatsweb-attachment----
Content-Type: text/plain; name="ttybug.cc"
Content-Disposition: inline; filename="ttybug.cc"

#include <fstream>

int main()
{
	using namespace std;

	ifstream in ("/dev/tty");
	ofstream out ("/dev/tty");
	in.tie(&out);

	out << "Enter a number: ";
	int i;
	in >> i;
	out << "You entered: " << i << endl;
	return 0;
}