Eric Blake <ebb9 <at> byu.net> writes:
>
> Larry Jones <lawrence.jones <at> siemens.com> writes:
>
> >
> > CVS nightly testing has been failing on BSD/OS and I've finally tracked
> > the problem down to a bug in fseeko -- it doesn't clear the EOF flag
> > after doing its magic:
>
> Just to make sure other platforms aren't affected, I'm installing this.
Sure enough, mingw is affected. I'm installing this.
From: Eric Blake <[EMAIL PROTECTED]>
Date: Thu, 13 Dec 2007 12:09:31 -0700
Subject: [PATCH] Fix fseeko on mingw.
* lib/fseeko.c (rpl_fseeko) [_IOERR]: Reset EOF flag on successful
seek.
Signed-off-by: Eric Blake <[EMAIL PROTECTED]>
---
ChangeLog | 4 ++++
lib/fseeko.c | 2 ++
2 files changed, 6 insertions(+), 0 deletions(-)
diff --git a/ChangeLog b/ChangeLog
index b91f640..dd3d5f2 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,5 +1,9 @@
2007-12-13 Eric Blake <[EMAIL PROTECTED]>
+ Fix fseeko on mingw.
+ * lib/fseeko.c (rpl_fseeko) [_IOERR]: Reset EOF flag on successful
+ seek.
+
Beef up fseek tests.
* tests/test-fseek.c (main): Also test eof handling.
* tests/test-fseeko.c (main): Likewise.
diff --git a/lib/fseeko.c b/lib/fseeko.c
index e495226..97dcf6d 100644
--- a/lib/fseeko.c
+++ b/lib/fseeko.c
@@ -112,6 +112,8 @@ rpl_fseeko (FILE *fp, off_t offset, int whence)
fp->_offset = pos;
fp->_flags |= __SOFF;
fp->_flags &= ~__SEOF;
+#elif defined _IOERR /* AIX, HP-UX, IRIX, OSF/1, Solaris, mingw
*/
+ fp->_flag &= ~_IOEOF;
#endif
return 0;
}
--
1.5.3.5
