ctime of mounted files is incorrect about fusecompress HOT 5 CLOSED

I actually ran into this very bug with rsync last night, thanks for finding it. 
I'll
apply the change and write a test when I have the time.

Speaking of rsync, I found another related issue.  rsync sets the mtime/atime 
for
symbolic links to the same value as the source link.  The current fusecompress 
uses
utime() which sets the target file's atime/mtime, not the link's.  At first I 
thought
this might be an unsolvable problem since fuse only provides a utime interface 
(not
lutimes()) but then when I watched the DEBUG messages I saw that fuse 
dereferences,
or doesn't dereference, the link before calling fusecompress as appropriate.  
So, if
rsync tried to set mtime for the link, fuse calls fuse_compress.utime() with 
the link
filename.  If rsync tries to set the mtime for the target file, fuse calls
fuse_compress.utime() with the target filename.
Anyway, I found that using lutimes() instead of utime() fixes the problem.  
My testing seems to show that the behavior is what you would expect.  If you 
use the
"touch" command, which always modifies the target not the link, on the link the
target is modified.  However, if you run lutimes() against the link (I have a 
little
C program that's useful testing, and I'll attach it) the link is updated.  I 
believe
that this change is correct but, of course, it should be tested more thoroughly.

This diff includes the first change too:
--- fusecompress.c  (revision 90)
+++ fusecompress.c  (working copy)
@@ -112,7 +112,7 @@
    // (tar checks this item and it is loudly when the result
    // is different than what it exepects)
    //
-   stbuf->st_ctime = stbuf->st_mtime;
+   //stbuf->st_ctime = stbuf->st_mtime;

    UNLOCK(&file->lock);
    return 0;
@@ -436,12 +436,25 @@
 static int fusecompress_utime(const char *path, struct utimbuf *buf)
 {
    const char *full;
-   
+   struct timeval timesval[2];
+   struct timeval *timesbuf=NULL;
+
    full = fusecompress_getpath(path);

-   if (utime(full, buf) == -1)
+   DEBUG_("('%s')", full);
+
+   if (buf != NULL)
+   {
+       timesval[0].tv_usec = 0;
+       timesval[1].tv_usec = 0;
+       timesval[0].tv_sec = buf->actime;
+       timesval[1].tv_sec = buf->modtime;
+       timesbuf=timesval;
+   }
+       
+   if (lutimes(full,timesbuf) == -1)
        return -errno;
-
+   
    return 0;
 }

This issue was closed by r96.

Patch added in r96. Thanks for your contribution.