[Alexandria-list] Book Providers and Tests

[Mathieu Leduc-Hamel]

Joseph:

Yah it would be great, I'm already mark as developer on the alexandria
website but i think you need enable something to give access to the
subversion repository...

For the book providers, yes everything should be enable by default,
but selecting by language could be implemented by adding a language
flag to each book providers.

maybe we could add status flag to bookproviders too. just to disable
them who are not so stable?

L.C. Karssen:

Yeah we can think of many option like that on the command line and I'm
planning to offer the more options available in alexandria right now,
on the command line, export would a great candidate.



Mathieu Leduc-Hamel

On 2/2/07, Joseph Method <tristil at gmail.com> wrote:
> Mathieu, I thought of the command-line thing too. Yes, that's a great
> idea. But instead of sending me a diff, why not become a developer and
> commit it yourself? I just need your Rubyforge account name.
>
> At first I didn't know what you meant by being able to select book
> providers by language, but yes, if you mean being able to narrow the
> selections. The default should be all languages, though. Take a look
> at what's already covered in the preferences for book providers and
> how they're implemented on the actual book provider. It would be great
> to be able to turn off the book providers without having to go into
> preferences, just by clicking a check box on the book provider.
>
> On 2/2/07, Mathieu Leduc-Hamel <marrakis at gmail.com> wrote:
> > Hi all,
> >
> > I'm sorry Joseph, I hadn't any time to code on alexandria last week
> > and you had achieve all what I would do on the side of the book
> > providers unit test. I already start to split the amazon.com provider
> > in multiples book_provider, separated by language and country.
> >
> > I had some interrogation the current book provider window. I think we
> > should have a way to select book providers by language and to have the
> > possibility to disable the one we don't want. I will work on that.
> >
> > Another think, for my test, it would cool to have a command-line
> > interface for alexandria, I already implement something in my own
> > repository but I think it's a good improvement, particularly if we'll
> > implement a network interface for alexandria. And by the way we could
> > use that to force log level and the log file location, or another
> > option.
> >
> > I will send you some patch shortly. But what do you think of that? Is
> > this good idea? Do You see any other command-line options that should
> > be available?
> >
> > See ya
> >
> > Mathieu Leduc-Hamel
> >
> > On 2/1/07, Joseph Method <tristil at gmail.com> wrote:
> > > The most recent commit has the book providers all passing their tests,
> > > except for Barnes & Noble and MCU. To tell the truth, I got tired of
> > > fixing book providers and commented out their tests and requires. If
> > > someone else would like to fix these two, that would be great, and
> > > they can be re-included for the release. Essentially, all of these
> > > book providers didn't work because their regexes didn't match anything
> > > and other basic issues (like amadeusbuch.at now being thalia.at!). As
> > > things stand, the tests only certify that the book providers find a
> > > *given isbn*. Additional tests have to be written to cover titles,
> > > authors and keywords. While I'm on the topic, these are other tests
> > > that need to be written:
> > >
> > > * Alexandria starts up completely and then shuts down when given the
> > > close signal
> > > * import an isbn list
> > > * check that encoding is correct within book fields
> > >
> > > My impression is that unittests are mainly for testing backend
> > > functionality. Is this true, or is it feasible to erect a full
> > > Alexandria scaffold that will allow data entry through the Gtk
> > > widgets? Is there a Ruby framework for this, and is it worth the
> > > effort? Would anyone like to become the Testing Czar???
> > >
> > > In any case, what I want to highlight about the current passing tests
> > > is that 1) they don't tell you as much as you may think (they're only
> > > as good as the assertions they make) and 2) at least within the
> > > current release cycle, *tests must pass before a commit or a diff will
> > > be accepted*. This means, for example, that any additional book
> > > providers that don't pass tests will be rejected. As new tests are
> > > written, this policy will cover whatever piece of functionality is
> > > being tested.
> > >
> > > I think this is a sound policy for a collaborative project and I hope
> > > that there can be consensus about this, but when I say "policy" I am
> > > actually just talking about what I will work on.
> > >
> > > Oh, and a really boring, but helpful task for anyone who wants it: the
> > > Amazon provider should be broken up into each of the separate language
> > > sites, so people don't have to choose, and so we can see that they all
> > > actually work (jp and de settings don't do anything currently).
> > > Actually, there's more to it, because the providers should be capable
> > > of activation/deactivation, but that's an additional step.
> > > --
> > > -J. Method
> > > _______________________________________________
> > > Alexandria-list mailing list
> > > Alexandria-list at rubyforge.org
> > > http://rubyforge.org/mailman/listinfo/alexandria-list
> > >
> >
>
>
> --
> -J. Method
>