this patch fixes it... Also -
should it really DELONCLOSE by default? I don't think that's consistent
with most implementations of mkstemp - worst case we should document it
(and since the consumer can pass their own flags its not a big deal).
sterling
Index: srclib/apr/file_io/unix/mktemp.c
===================================================================
RCS file: /home/cvspublic/apr/file_io/unix/mktemp.c,v
retrieving revision 1.19
diff -u -r1.19 mktemp.c
--- srclib/apr/file_io/unix/mktemp.c 8 Jan 2002 12:29:38 -0000 1.19
+++ srclib/apr/file_io/unix/mktemp.c 16 Jan 2002 18:22:55 -0000
@@ -207,7 +207,8 @@
#ifdef HAVE_MKSTEMP
int fd;
#endif
- flags = (!flags) ? APR_READ | APR_WRITE | APR_EXCL | APR_DELONCLOSE :
flags;
+ flags = (!flags) ? APR_CREATE | APR_READ | APR_WRITE |
+ APR_EXCL | APR_DELONCLOSE : flags;
#ifndef HAVE_MKSTEMP
return gettemp(template, fp, flags, p);
#else
