Idiomatic callbacks

[jamiehodge]

This is certainly a case of pickiness, but is_authorized? and is_conflict? don't feel very idiomatic. I've renamed them to authorized? and conflict?, so that they more closely resemble, for example, forbidden?. I was also tempted to rename generate_etag to etag, but saw that a number of tests currently include :etag.

I realize that there would need to be a deprecation of the existing syntax, but I a) didn't know if you already have a deprecation policy or; b) whether you agreed with the changes to begin with.

Regards,

Jamie

[ghost]

Agree on the idiomacy, but as we're past 1.0, we need to keep the is_ versions around. An alias authorized? => is_authorized? would be enough for that, and would give us time to consider how deprecations could work.

[jamiehodge]

I'll add the aliases.

What are your feelings regarding renaming generate_etag to etag? I understand the callbacks as attributes of the resource, which makes the calculation prefix inappropriate IMHO.

Jamie

[jamiehodge]

I've added aliases for authorized?, conflict? and etag for backwards compatibility. There don't appear to be any tests for the callbacks specifically, so there are no accompanying tests. Perhaps another tasks...

[ghost]

There are specs for Webmachine::Decision::Flow, which is used for gathering facts about a resource. I guess it's best to just grep the codebase for the methods names you changed.

First interesting question about changing the resource API: in the transition phase, do we test the old or new methods? 😈

[jamiehodge]

I have changed all the specs to use the new methods. I hope this is acceptable.

On 05/01/2013, at 22.00, Lars Gierth notifications@github.com wrote:

There are specs for Webmachine::Decision::Flow, which is used for gathering facts about a resource. I guess it's best to just grep the codebase for the methods names you changed.

First interesting question about changing the resource API: in the transition phase, do we test the old or new methods?

—
Reply to this email directly or view it on GitHub.

[bernd]

I guess @seancribbs wanted to keep the callback names close to the ones in basho/webmachine. If that is still a goal, I would oppose to this change to keep the callback API small.

[jamiehodge]

Let's wait for his reaction. I don't see any additional changes, so I would argue that three aliases is not egregious.

[ghost]

I agree that aliases are totally fine, even if I don't like all of the proposed changes (i probably prefer 'generate_etag' over just 'etag'), but having that choice is cool.

[seancribbs]

This is stale, closing. I still maintain that there is value in keeping the callback names as close to the Erlang version, even if they aren't "idiomatic" to Ruby.