[Q] Testing SPrevayler / Developing SPrevayler

Marco Paga mail at marco-paga.de
Fri Mar 14 08:17:48 UTC 2003


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thursday 13 March 2003 21:26, Adrian Lienhard wrote:
> Hi Marco
>
> As I mentioned in my last mail (on 03/03/03), I'd like to help to improve
> and write tests.
>
> I don't know if you are aware of this:
[SNIP]
> And these will all be logged... A solution may be to use IDs as
> Avi mentioned earlier.

I know what you mean, but if you have a look at the documentation of 
ReferenceStream or of Sixx you can see that this support is built in. You can 
see it for example wiith Sixx if you fire out many small commands and open 
afterwards the last commandlog file. you can see that the seconds value is 
just written ones and then referenced by the following commands that were 
fired out in the same second.

>
> Generally, I like the idea of Prevayler to have a very simple persistence
> for not too large models.

I think SPrevayler can handle large systems, too. We should just go ahead and 
test it.

> I started a very simple implementation of a bank account system with tests.

That's great. It would be great to post it when you're done.

> How do you want to go on?

At the moment I'm thinking about fault tolerance. What should happen if an 
error occures when the command is not correctly written. And what happens if 
the last snapshot is corrupted. After that I want to start work on 
replication.

> There may be other interested people willing to help a little bit?

If there would be people that would be great.

>
> Adrian
>
> _____________________
> Adrian Lienhard
> www.adrian-lienhard.ch
> www.netstyle.ch

regards
marco
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE+cZAxlI77o82OdhIRAv6YAKChvfLnmrGDYWXADx8u4vsNmx4D/gCfRs6M
vVujYbz0d640Owoznx8+TeI=
=cZSp
-----END PGP SIGNATURE-----



More information about the Squeak-dev mailing list