Re: Object Serialization (again)

[ptg@ccs.neu.edu (Paul T. Graunke)]

I don't know a good way to do it.  I've replaced structure definitions
with my own vector based implementations to be able to serialize them
before.  You may get by with inspectors and struct->vector.  Serializing
closures and continuations is harder.  CSPing and defunctionalizing provides
a heavy handed solution.  That's what we used for
  Automated Software Engineering 2001
  http://www.ccs.neu.edu/scheme/pubs/index.html#papers

Paul

> 
> Some time back on this list someone asked a question about
> 'easy' object serialization.
> 
> I'm curious about the same thing.  I want to store some
> set of persistent objects, possibly as data in a database,
> possibly as binary data on disk -- it doesn't really
> matter.
> 
> I don't really want to be so heavy-weight as to require
> SrPersist/ODBC and a Full RDBMS for my little application.
> Is there a simple way to store the data as a binary
> image that can be re-loaded at runtime?
> 
> I could just spew the stuff out as XML into a text
> file, but I'd like to avoid the time spent reparsing and
> initializing the data.
> 
> One idea that comes to mind immediately is to create a
> special module for the data, and just recompile it to a
> *.zo each time the program shuts down...
> 
> Anyone else solved this problem before?
> 
> -Brent
>