Re: [Boost-users] [Serialization] serializing handles to pointers, and serializing as garbage collection

From: Robert Ramey (ramey_at_[hidden])
Date: 2007-07-05 11:47:44

I looked at the email and really didn't understand what you wanted the 
library to do that it doesn't already do.  I could have starting a back and 
forth trying to figure this out, but I think it would better just to 
summarize what the library does do.

The library saves and restores any data structures which can be represented 
in C++.

Soooooooo.  If you have some dataset which cannot fit into memory all at 
once, its not really representable as a C++ data structure.  What you would 
have is a C++ data structure for handling data too large to fit all at once 
in memory.  Presumably you already have some sort of infrastructure for 
handling this data (or plan to build one).  If what you are asking is 
whether serialization library provides such an infrastructure, the answer 
would be No - it does not.

Howeever, it might still be helpful.  Suppose you have some sort set up some 
sort of code for paging, on-demand load or what ever implemented in C++ 
which pointed to or managed some sort of off-line storage, the C++ 
implemenation could save its state and restore it which seems to me would be 
what you want.

Its also possible that the serialization library might be used to save and 
restore pieces at a time.

Sooooo, I don't think the serialization library addresses what you want, but 
it may (or may not) be helpful in achieving your goal.

Robert Ramey

"Sean Cavanaugh" <worksonmymachine_at_[hidden]> wrote in message 
news:3ae31ecb0707050116w698de581pf1420f0b7ddce612_at_mail.gmail.com...
I'm a bit new to boost and am trying to determine if I can coax the existing 
boost serialization system to work for a fairly standard asset loading 
system.  Assets in my domain of game programming are things such as bitmaps, 
model geometry, audio data, shaders, and various chunks of meta data 
attached to them.  The quantity of such data should always be assumed to 
vastly exceed the ability to load it all at once, which means getting sneaky 
when dealing with persistence.  From my point of view the 
boost::serialization code is missing a lot of important features, so I am 
kind of surprised that they are not there.