Date: Tue, 30 Jan 2001 17:14:00 -0500
Message-Id: <200101302214.RAA24294@envy.delorie.com>
X-Authentication-Warning: envy.delorie.com: dj set sender to dj AT envy DOT delorie DOT com using -f
From: DJ Delorie <dj AT delorie DOT com>
To: djgpp-workers AT delorie DOT com
In-reply-to: <200101302125.WAA29699@father.ludd.luth.se> (message from Martin
	Str|mberg on Tue, 30 Jan 2001 22:25:49 +0100 (MET))
Subject: Re: Bug000323
References:  <200101302125 DOT WAA29699 AT father DOT ludd DOT luth DOT se>
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
Precedence: bulk


> -extern char *__file_handle_modes;
> +extern unsigned short *__file_handle_modes;

Could this in any way affect already-compiled libraries or programs?

> Index: src/libc/ansi/stdio/doprnt.c
> +
> +  if (fp->_flag & _IOAPPEND)
> +  {
> +    if ( llseek(fileno(fp), 0, SEEK_END) == -1 )
> +    {
> +      return (EOF);
> +    }
> +  }

This shouldn't be needed; the flsbuf() call should move to EOF.

> -      n = _write(fileno(f), base, rn);
> +    {
> +      if ( f->_flag & _IOAPPEND )
> +      {
> +	int save_errno = errno; /* We don't want llseek()'s setting 
> +				   errno to remain. */
> +	if( llseek(fileno(f), 0, SEEK_END) == -1 )
> +	{
> +	  errno = save_errno;
> +	  n = 0;
> +	}
> +	else
> +	{
> +	  n = _write(fileno(f), base, rn);
> +	}
> +      }
> +      else
> +      {
> +	n = _write(fileno(f), base, rn);
> +      }
> +    }

Can we remove all the else cases, and have llseek return a failure if
we can't seek?  Then we'd have just one _write() call left (the
original one, left there).

> @@ -56,7 +57,9 @@ fopen(const char *file, const char *mode
>      return NULL;
>  
>    if (*mode == 'a')
> -    lseek(fd, 0, SEEK_END);
> +  {
> +    f->_flag |= _IOAPPEND;
> +  }

Is this needed?  write() would seek there anyway.