Changeset 13255


Ignore:
Timestamp:
Dec 5, 2009, 3:39:57 AM (10 years ago)
Author:
gb
Message:

revert r13254: 'deref(base,1)' and 'deref(n,1)' should have exactly
the same effect.

One could argue that mark_root() and rmark() should (for some reason)
be consistent about this; since 'base' is a dnode-aligned pointer, it's
at least slightly faster to say:

base[1] = lisp_nil;

and not use 'deref' to clear tag bits that're already clear, but the
same word is being set to NIL pre- and post-r13254, so it's misleading
for that change to claim that it causes "the right offset" to be cleared.

File:
1 edited

Legend:

Unmodified
Added
Removed
Note: See TracChangeset for help on using the changeset viewer.