Re: [boost] [Serialization] Bizarre bug

Subject: Re: [boost] [Serialization] Bizarre bug
From: Robert Ramey (ramey_at_[hidden])
Date: 2009-08-11 12:59:38

Jarl Lindrud wrote:
> Robert Ramey <ramey <at> rrsd.com> writes:
>
>>
>> That is not relevant.  The ar >> and ar << don't have to be in
>> the same program.  They just have to be consistent with each
>> other.
>
> Thanks for the clarification. So the program in the OP (let me call
> it A0 :) ), is a valid test case then, and a simpler one at that,
> than A1, A2 and B.

>> In the IPC case, the program could only occur when
>> the recieving program is a later version of the sending version.
>> That is the same problem - reading an old archive version.
>
> Well, the problem also manifests in single programs, as A0 indicates.
> There's no versioning going on there.

This program contains a dummy function whose only purpose
is to illustrate a problem that could occur under very infrequent
circumstances.  Whether or not it should be characterised as a valid
usage of the library is a question I'll leave unaddressed.

Basically there were and are three options here.

a) include class information for all types inside of all archives.
b) eliminate the option "track_selectively"
c) do nothing.

There are trade-offs associated with each option.  Given
the trade-offs, I still believe that c) is the best choice.  And
I also believe most people with a good understanding of
these trade-offs would agree with me.

Having said that I'm considering emitting a compile time warning for
types to which all the following apply
a) use "track_selectively"
b) don't store class data in the archive
c) serialialize through a pointer.

This warning wouldn't come up very often, (though it would in
your test case) and almost all the time it would be spurious.
It would be legitimate only in my previous versioning
example and your test case.  So then  I would have to explain what this
means either in the documentation or repeatedly on this list
- probably both.  Then someone would (legitimately) want to be
able to turn it off.  So even this "simple" idea entails a lot
of work.

I've also considered - as I have in the past - emitting a
compiler time error if versioning is used on type which
doesn't store class information in the archive.  I tried
this before but I couldn't get it implemented without creating
other problems - but I think I might have method now.
This is only peripheraly related to this discussion, I just
mention it so that no one will think this is a total waste
of time.

Robert Ramey

> Regards,
> Jarl.
>
>
> _______________________________________________
> Unsubscribe & other changes:
> http://lists.boost.org/mailman/listinfo.cgi/boost