On Mon, 11 Jun 2007, Otto Moerbeek wrote:

> On Sun, 10 Jun 2007, Stuart Henderson wrote:
>
> > I just tried "dump -0uf -|restore rf -" on -current (Jun 9, #270),
> > plain FFS1, and noticed that file creation times of the restored files
> > were all set to the epoch.
>
> I can reproduce that...
>
> > If I restore a dump created by older dump using a newer restore,
> > I'm seeing the file creation time restored as the current time
>
> But that not. A new restore reading an old dump restores my timestamps
> ok. Both for dirs and files.
>
> >
> > If I restore a dump created by newer dump using an older restore,
> > I'm seeing the file creation time restored as the epoch
>
> Have to try that one.
>

Try this,

-Otto

Index: traverse.c
================================================== =================
RCS file: /cvs/src/sbin/dump/traverse.c,v
retrieving revision 1.21
diff -u -p -r1.21 traverse.c
--- traverse.c 3 Jun 2007 20:16:08 -0000 1.21
+++ traverse.c 11 Jun 2007 15:40:39 -0000
@@ -480,9 +480,9 @@ dumpino(union dinode *dp, ino_t ino)
if (sblock->fs_magic == FS_UFS1_MAGIC) {
spcl.c_mode = dp->dp1.di_mode;
spcl.c_size = dp->dp1.di_size;
- spcl.c_atime = (time_t)dp->dp1.di_atime;
+ spcl.c_old_atime = (time_t)dp->dp1.di_atime;
spcl.c_atimensec = dp->dp1.di_atimensec;
- spcl.c_mtime = (time_t)dp->dp1.di_mtime;
+ spcl.c_old_mtime = (time_t)dp->dp1.di_mtime;
spcl.c_mtimensec = dp->dp1.di_mtimensec;
spcl.c_birthtime = 0;
spcl.c_birthtimensec = 0;