Re: [Gurusamy Sarathy <gsar@activestate.com>] ferror() after fread() on a FILE* opened for write

Re: [Gurusamy Sarathy <gsar@activestate.com>] ferror() after fread() on a FILE* opened for write

[H.J. Lu]

> 
> hjl@lucon.org (H.J. Lu) writes:
> 
> |> How about this patch?
> 
> I have already checked this in for libc-2.2.  It doesn't change anything
> for fread anyway, only fgetc.
> 

What do you mean by that? Have you tried my patch with your program?

-- 
H.J. Lu (hjl@gnu.org)

Re: [Gurusamy Sarathy <gsar@activestate.com>] ferror() after fread() on a FILE* opened for write

[Andreas Schwab]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 554 bytes --]

hjl@lucon.org (H.J. Lu) writes:

|> > 
|> > hjl@lucon.org (H.J. Lu) writes:
|> > 
|> > |> How about this patch?
|> > 
|> > I have already checked this in for libc-2.2.  It doesn't change anything
|> > for fread anyway, only fgetc.
|> > 
|> 
|> What do you mean by that? Have you tried my patch with your program?

What do you mean with "your program"?

-- 
Andreas Schwab                                  "And now for something
schwab@suse.de                                   completely different."
SuSE GmbH, Schanzäckerstr. 10, D-90443 Nürnberg

Re: [Gurusamy Sarathy <gsar@activestate.com>] ferror() after fread() on a FILE* opened for write

[H.J. Lu]

This patch should pass the test. The problem is _IO_file_xsgetn
and _IO_XXX_file_xsputn may call read/write directly.


-- 
H.J. Lu (hjl@gnu.org)
----
Wed Jul 21 08:03:47 1999  H.J. Lu  <hjl@gnu.org>

	* libio/fileops.c (_IO_new_file_underflow): Set the error flag 
	if fp is not opened for read.
	* libio/oldfileops.c (_IO_old_file_underflow): Likewise.

	* libio/fileops.c (_IO_new_file_xsputn): Set the error flag and
	return 0 if f is not opened for write.
	* libio/oldfileops.c (_IO_old_file_xsputn): Likewise.

	* libio/fileops.c (_IO_file_xsgetn): Set the error flag and
	return 0 if f is not opened for read.

Index: libio/fileops.c
===================================================================
RCS file: /work/cvs/gnu/glibc-2.1/libio/fileops.c,v
retrieving revision 1.10
diff -u -p -r1.10 fileops.c
--- libio/fileops.c	1999/05/01 22:41:30	1.10
+++ libio/fileops.c	1999/07/21 14:58:22
@@ -349,6 +349,7 @@ _IO_new_file_underflow (fp)
   if (fp->_flags & _IO_NO_READS)
     {
       __set_errno (EBADF);
+      fp->_flags |= _IO_ERR_SEEN;
       return EOF;
     }
   if (fp->_IO_read_ptr < fp->_IO_read_end)
@@ -745,6 +746,13 @@ _IO_new_file_xsputn (f, data, n)
   int must_flush = 0;
   _IO_size_t count;
 
+  if (f->_flags & _IO_NO_WRITES)
+    {
+      __set_errno (EBADF);
+      f->_flags |= _IO_ERR_SEEN;
+      return 0;
+    }
+
   if (n <= 0)
     return 0;
   /* This is an optimized implementation.
@@ -833,6 +841,13 @@ _IO_file_xsgetn (fp, data, n)
   register _IO_size_t want, have;
   register _IO_ssize_t count;
   register char *s = data;
+
+  if (fp->_flags & _IO_NO_READS)
+    {
+      __set_errno (EBADF);
+      fp->_flags |= _IO_ERR_SEEN;
+      return 0;
+    }
 
   want = n;
 
Index: libio/oldfileops.c
===================================================================
RCS file: /work/cvs/gnu/glibc-2.1/libio/oldfileops.c,v
retrieving revision 1.2
diff -u -p -r1.2 oldfileops.c
--- libio/oldfileops.c	1999/05/01 22:41:30	1.2
+++ libio/oldfileops.c	1999/07/21 14:59:15
@@ -313,6 +313,7 @@ _IO_old_file_underflow (fp)
   if (fp->_flags & _IO_NO_READS)
     {
       __set_errno (EBADF);
+      fp->_flags |= _IO_ERR_SEEN;
       return EOF;
     }
   if (fp->_IO_read_ptr < fp->_IO_read_end)
@@ -668,6 +669,13 @@ _IO_old_file_xsputn (f, data, n)
   int must_flush = 0;
   _IO_size_t count;
 
+  if (f->_flags & _IO_NO_WRITES)
+    {
+      __set_errno (EBADF);
+      f->_flags |= _IO_ERR_SEEN;
+      return 0;
+    }
+
   if (n <= 0)
     return 0;
   /* This is an optimized implementation.

Re: [Gurusamy Sarathy <gsar@activestate.com>] ferror() after fread() on a FILE* opened for write

[Ulrich Drepper]

hjl@lucon.org (H.J. Lu) writes:

> This patch should pass the test. The problem is _IO_file_xsgetn
> and _IO_XXX_file_xsputn may call read/write directly.

This is not necessary.  Don't add it.

-- 
---------------.      drepper at gnu.org  ,-.   1325 Chesapeake Terrace
Ulrich Drepper  \    ,-------------------'   \  Sunnyvale, CA 94089 USA
Cygnus Solutions `--' drepper at cygnus.com   `------------------------

Re: [Gurusamy Sarathy <gsar@activestate.com>] ferror() after fread() on a FILE* opened for write

[Andreas Jaeger]

>>>>> Ulrich Drepper writes:

Uli> hjl@lucon.org (H.J. Lu) writes:
>> This patch should pass the test. The problem is _IO_file_xsgetn
>> and _IO_XXX_file_xsputn may call read/write directly.

Uli> This is not necessary.  Don't add it.

Are you sure?  Check the current Posix draft of the Austin Group for
fread/fgetc and fwrite/fputc.  They explicitly mention EBADF for:
  
  The filedescriptor underlying stream is not a valid file descriptor
  open for reading (writing with fwrite/fputc).

But this is marked as a new addition.

Andreas
-- 
 Andreas Jaeger   aj@arthur.rhein-neckar.de    jaeger@informatik.uni-kl.de
  for pgp-key finger ajaeger@aixd1.rhrk.uni-kl.de

Re: [Gurusamy Sarathy <gsar@activestate.com>] ferror() after fread() on a FILE* opened for write

[Andreas Schwab]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 919 bytes --]

Andreas Jaeger <aj@arthur.rhein-neckar.de> writes:

|> >>>>> Ulrich Drepper writes:
|> 
|> Uli> hjl@lucon.org (H.J. Lu) writes:
|> >> This patch should pass the test. The problem is _IO_file_xsgetn
|> >> and _IO_XXX_file_xsputn may call read/write directly.
|> 
|> Uli> This is not necessary.  Don't add it.
|> 
|> Are you sure?  Check the current Posix draft of the Austin Group for
|> fread/fgetc and fwrite/fputc.  They explicitly mention EBADF for:
|>   
|>   The filedescriptor underlying stream is not a valid file descriptor
|>   open for reading (writing with fwrite/fputc).

If you use std{in,out,err} then the underlying fd _is_ valid for both
read and write (until you change that with freopen).

Andreas.

-- 
Andreas Schwab                                  "And now for something
schwab@suse.de                                   completely different."
SuSE GmbH, Schanzäckerstr. 10, D-90443 Nürnberg

Re: [Gurusamy Sarathy <gsar@activestate.com>] ferror() after fread() on a FILE* opened for write

[Ulrich Drepper]

Andreas Jaeger <aj@arthur.rhein-neckar.de> writes:

> Are you sure?  Check the current Posix draft of the Austin Group for
> fread/fgetc and fwrite/fputc.  They explicitly mention EBADF for:

OK, it's nowadays referring to getc.  (We are working through the XSH
stuff only tomorrow so I haven't read it).  In this light the changes
should be added.

-- 
---------------.      drepper at gnu.org  ,-.   1325 Chesapeake Terrace
Ulrich Drepper  \    ,-------------------'   \  Sunnyvale, CA 94089 USA
Cygnus Solutions `--' drepper at cygnus.com   `------------------------