hi all
we've been tracking down a discrepancy between the atime reported by
finfo->atime and stat [1]. as it turns out, the documented behavior for
GetFileInformationByHandle's ftLastAccessTime [2] is very different from
that of a raw stat call.
anyway, just in case it helps others debug things in the future, here's a
brief note.
--Geoff
[1] http://marc.theaimsgroup.com/?l=apache-modperl-dev&m=107158889515733&w=2
[2]
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/fileio/base/by_handle_file_information_str.asp
Index: file_io/win32/filestat.c
===================================================================
RCS file: /home/cvspublic/apr/file_io/win32/filestat.c,v
retrieving revision 1.83
diff -u -r1.83 filestat.c
--- file_io/win32/filestat.c 19 Nov 2003 00:25:17 -0000 1.83
+++ file_io/win32/filestat.c 16 Dec 2003 16:09:01 -0000
@@ -350,6 +350,9 @@
memset(finfo, '\0', sizeof(*finfo));
+ /* note that ftLastAccessTime is documented to return the
+ * the proper date but time rounded to midnight
+ */
FileTimeToAprTime(&finfo->atime, &wininfo->ftLastAccessTime);
FileTimeToAprTime(&finfo->ctime, &wininfo->ftCreationTime);
FileTimeToAprTime(&finfo->mtime, &wininfo->ftLastWriteTime);
