adding on branch

What is the logic behind the fact that when you have a project
checked out on a branch, and then you go:

cvs add filename.c
cvs commit -m "." filename.c

it gets put into the Attic directory instead of the normal directory?

Is this because a checkout on the "head" would otherwise pick
up this branch "automatically"/"accidentally"?

But when a subsequent "cvs add" is done, not on a branch, then
it is taken out of the Attic (but unfortunately Attic is not removed -
perhaps this code is what is required to do that?).

Index: rcs.c
===================================================================
RCS file: /cvs/ccvs/src/rcs.c,v
retrieving revision 1.262
diff -c -r1.262 rcs.c
*** rcs.c 3 Jan 2003 04:24:10 -0000 1.262
--- rcs.c 25 May 2003 07:47:15 -0000
***************
*** 700,705 ****
--- 700,707 ----
       free (newpath);
       return 1;
   }
+  *p = '\0';
+  rmdir (rcs->path);
      }

      free (rcs->path);


Ok, but an import is another way of adding a "head" version.
So the import should also go through the same logic, moving a
file out of the Attic and putting it into the normal spot, and
(hopefully) also removing the Attic directory.

Is something like this the right place...

Index: import.c
===================================================================
RCS file: /cvs/ccvs/src/import.c,v
retrieving revision 1.133
diff -c -r1.133 import.c
*** import.c 24 Sep 2002 20:10:38 -0000 1.133
--- import.c 25 May 2003 08:02:31 -0000
***************
*** 694,699 ****
--- 694,704 ----
   letter = 'U';
      add_log (letter, vfile);

+     if (inattic)
+     {
+         RCS_setattic(vers->srcfile, 0);
+     }
+
      freevers_ts (&vers);
      return (0);
  }

All diffs against cvs1-11-5.

BFN.  Paul.