Re: mingw lseek bug

[Bruno Haible]

Eric Blake wrote:
> > You said that you got an fflush failure on mingw. Could you please
> > show a test program that failed before?
> 
> The m4 test suite detected failures on mingw due to this code that got
> invoked prior to the syscmd macro, in an attempt to follow the POSIX rule
> that for seekable streams, child processes should see the stream at the
> next character left unprocessed by the parent:
> 
>   if (lseek (STDIN_FILENO, 0, SEEK_CUR) >= 0
>       && fflush (stdin) == 0)
>     {
>       fseeko (stdin, 0, SEEK_CUR);
>     }
> 
> The failures in the m4 testsuite were due to lost data since the fflush
> discarded what was on the pipe, even though the fflush should never have
> been reached.

Thanks for explaining. I thought you were talking about the lseek call
inside fflush, not about an lseek call before it.

Since loss of data is a severe condition, I'm adding a note about it:


2007-05-28  Bruno Haible  <address@hidden>

        * lib/stdio_.h (fflush): More comments.

--- lib/stdio_.h        24 May 2007 16:59:21 -0000      1.27
+++ lib/stdio_.h        28 May 2007 13:52:52 -0000
@@ -270,7 +270,11 @@
 # if @REPLACE_FFLUSH@
 #  define fflush rpl_fflush
   /* Flush all pending data on STREAM according to POSIX rules.  Both
-     output and seekable input streams are supported.  */
+     output and seekable input streams are supported.
+     Note! LOSS OF DATA can occur if fflush is applied on an input stream
+     that is _not_seekable_ or on an update stream that is _not_seekable_
+     and in which the most recent operation was input.  Seekability can
+     be tested with lseek(fileno(fp),0,SEEK_CUR).  */
   extern int fflush (FILE *gl_stream);
 # endif
 #elif defined GNULIB_POSIXCHECK