[comp.lang.clos] saving instances on file -- update

mdb@zero.arc.ab.ca (Mark Brinsmead) (04/11/91)

In article <1991Apr8.194957.25531@arc.ab.ca>, mdb@zero.arc.ab.ca (Mark
Brinsmead) writes:
|> 
...
|>    It just so happens that I *do* have such an animal.  It is basically
|> a set of methods which will traverse an arbitrary tree of LISP/CLOS
|> structures (including instances, hash-tables, arrays, strings, symbols,
|> and numbers)  and save them to a file in the form of lisp source code 
|> which, when executed, will re-create the structure(s) fully preserving
|> EQ'ness throughout the hierarchy.
|> 
|>    It works (should work) for circular structures, and should be portable
|> to almost any common lisp using an up-to-date version of PCL.  Because it
|> is based on PCL, there is no support for saving instances of structures 
|> defined with defstruct.  The code has been used (a bit) and should work,
|> but comes without any sort of warrantee or promises, WHATSOEVER.
|> 
|>    The only hitch is that I'm not sure whether I am allowed to release
|> this code to the public.  I shall begin investigating that issue now;
|> in the meantime, anyone interested in getting a copy should e-mail me
|> so I can decide whether there is sufficient interest to justify posting
|> this to the net. (Assuming, of course, that I am allowed to!)
|> 
|>    In the worst case, I can probably make the *design* available.  That
|> ought to give you enough information to build one of these beasties 
|> yourself.
|> 
|> 

    Just a brief update to the above.

    I have done a little bit of checking, and it appears that there is some
 doubt as to whether I can release this code to the public.  I might even
 have to consult the (*gasp*) corporate lawyers.

    There appears to be enough general interest to justify posting to the
 net, but it could be ages before I can get an official legal opinion on
 this.  (Of course, it could be tomorrow, too.)  Hang in there!

-- Mark Brinsmead

(by the way, this code we're talking about is not especially difficult;
all it takes is a little imagination and blind trust in hash-tables. In
the worst case I can post instructions on how to write one yourself, but
I'm sure it would be nice not to have to debug it.)