Posted by
Eric Ehlers-2 on
URL: http://quantlib.414.s1.nabble.com/Extending-QLXL-and-serialization-issue-tp13901p14575.html
Hi Gerardo,
Apologies, I only just noticed the message that you sent in February...
On 2013-02-11 11:52, Ballabio Gerardo wrote:
> -----Messaggio originale-----
> Da: Eric Ehlers [mailto:
[hidden email]]
>> We had long debates about how to guarantee compatibility across
>> releases. We considered solutions such as never deleting any types
>> from the serialization framework, even if the type is deprecated. But
>> even then you have the problem of existing types changing and so you
>> have to version each type. It's just a burden of using
>> boost::serialization.
>
> Changing or deleting existing types is an API break. When you're
> doing that, it's reasonable to expect that you may break
> compatibility
> of serialized files too. But compatibility could at least be
> preserved
> between non-API-breaking releases.
>
> How difficult would it be to modify gensrc so that it reads the list
> of types from a file and registers them in the order as they are
> written, instead of alphabetically? Then new types could always be
> added at the end of the list without changing the numbers of existing
> types, while deletions and reorderings would be reserved to
> API-breaking releases.
You and I are talking about the same thing. When I said "never
deleting any types from the serialization framework", I meant what you
said... instead of numbering the types alphabetically, we maintain a
master list of all types and we number the types in the sequence in
which they appear in that list. The id number of each type never
changes and new types are added to the end of the list. This would not
be hard to implement.
Kind Regards,
Eric
--
===================================================
Eric Ehlers
nazcatech sprl | Brussels |
http://www.nazcatech.be* Distributed computing for pricing analytics
* Use Microsoft Excel as a client to the Grid
------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users