Re: Ping[3]: [PATCH][BZ #14543] Fix fseek behaviour when called inwide mode

[Jeff Law <law at redhat dot com>]

On 09/24/2012 06:08 AM, Siddhesh Poyarekar wrote:
> Ping again!
I haven't seen any comments on this submission since Joseph's testsuite 
comment.


First, I put this code into f18 & rawhide about 3 weeks ago and there 
haven't been any complaints.  And it does fix a real customer bug.

Second, shouldn't adjust_wide_data's block comment mention its return 
value?  Something as simple as "returns zero upon success and a nonzero 
value in the event of a failure" should probably be sufficient.

I'm a bit confused by the comment at the bottom of adjust_wide_data; it 
says:

+  /* Now seek to the end of the read buffer.  */
+  fp->_wide_data->_IO_read_ptr = fp->_wide_data->_IO_read_end;

Is seeking to the end of the read buffer really the right thing to do 
here?  I realize we're not actually seeking to the end of the buffer, 
that's handled at the "dumb" label within _IO_wfile_seekoff.  So it 
seems to me the comment could be written a bit better.

The ChangeLog entry should mention the BZ.  Here's an example of the 
canonical way to do this:

2012-09-19  Dmitry V. Levin  <ldv@altlinux.org>

        [BZ #14579]
        * elf/rtld.c (dl_main): Limit the check for self loading to normal
        mode only.
        * elf/tst-rtld-load-self.sh: New test.
        * elf/Makefile: Run it.

You should also mention that 14543 is fixed in the NEWS file.

Otherwise it looks good to me.  If you could update those nits it'd be 
appreciated.

jeff