Re: new module 'fpurge'

[Eric Blake]

Eric Blake <ebb9 <at> byu.net> writes:

> 
> Eric Blake <ebb9 <at> byu.net> writes:
> 
> > > #elif defined __sferror             /* FreeBSD, NetBSD, OpenBSD, MacOS X, 
> > Cygwin */
> > >   fp->_p = fp->_bf._base;
> > >   fp->_r = 0;
> > >   fp->_w = ((fp->_flags & (__SLBF | __SNBF) == 0) /* fully buffered? */
> 
> Order of operations.  I'm glad to know I'm not the only one that makes this 
> mistake.  Checking this in as obvious; unfortunately, it still doesn't solve 
> the fpurge crash of cygwin.

Actually, I found the real problem, with a bit more looking at the code.  With 
this patch, test-fpurge passes on cygwin (but there may still be a memory leak 
if the caller does ungetc before fpurge).

Committing this instead.

2007-04-16  Eric Blake  <address@hidden>

        Fix fpurge for cygwin.
        * lib/fpurge.c (fpurge): Fix order of operation flub, and return a
        value.
        * modules/fpurge-tests (Depends-on): Clean up trash.

diff --git a/lib/fpurge.c b/lib/fpurge.c
index dc56c8f..c4572fa 100644
--- a/lib/fpurge.c
+++ b/lib/fpurge.c
@@ -33,9 +33,10 @@ fpurge (FILE *fp)
 #elif defined __sferror             /* FreeBSD, NetBSD, OpenBSD, MacOS X, Cygwi
n */
   fp->_p = fp->_bf._base;
   fp->_r = 0;
-  fp->_w = ((fp->_flags & (__SLBF | __SNBF) == 0) /* fully buffered? */
+  fp->_w = ((fp->_flags & (__SLBF | __SNBF)) == 0 /* fully buffered? */
            ? fp->_bf._size
            : 0);
+  return 0;
 #elif defined _IOERR                /* AIX, HP-UX, IRIX, OSF/1, Solaris, mingw 
*/
   fp->_ptr = fp->_base;
   if (fp->_ptr != NULL)
diff --git a/modules/fpurge-tests b/modules/fpurge-tests
index c80a5eb..fc4da8a 100644
--- a/modules/fpurge-tests
+++ b/modules/fpurge-tests
@@ -8,4 +8,4 @@ configure.ac:
 Makefile.am:
 TESTS += test-fpurge
 check_PROGRAMS += test-fpurge
-
+MOSTLYCLEANFILES += t-fpurge.tmp