[Pulp-dev] Bindings' limitations

[Matthias Dellweg]

You nailed it! Thank you.
The bindings, as well as the ansible-modules I am in the process of
developing are trained to be _nice_ to the api. So I would say, they
are ok to test the happy flow for some dedicated workflows.
But when it comes to testing with bad request, you probably need to
handcraft them.

On Wed, 12 Feb 2020 09:25:45 -0500
Grant Gainey <ggainey at redhat.com> wrote:

> The whole point to having generated bindings is to prevent a
> client-tool from passing bogus data into Pulp, failing early and
> loudly so the client-author can *fix their code/assumptions*. Yes,
> that does make it hard to use the generated bindings to test
> bad-data-cases - but to me, that's just the price to be paid for the
> bindings doing the right thing in the real-world case of "I'm trying
> to write a tool to use Pulp, not to break it".
> 
> I would be *very* hesitant to make changes that make it easier to
> sneak "bad data" through the bindings-interfaces.
> 
> Just my $0.02
> G
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <http://listman.redhat.com/archives/pulp-dev/attachments/20200212/e7a9dc6b/attachment.sig>