fixing deprecation process once for all

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

fixing deprecation process once for all

Stéphane Ducasse
I took 5 min and wrote down that
        http://code.google.com/p/pharo/wiki/DeprecationProcess

comment if you want.

Reply | Threaded
Open this post in threaded view
|

Re: fixing deprecation process once for all

Henrik Sperre Johansen
Sounds good to me, to me there's a missing when/why description though:
Which criterias do we use for when deprecating is warranted?
    - # of callers in image required to deprecate planned for removal (0?)
    - # of callers / age when considering renaming for clarity (low
number of callers and/or age < X years?)
    - new/better API available (always, as long as new place documented?)

As for the how,  since we do have an integration process, would it not
also be desirable with some sort of integration lint tool working on slices?
For deprecations, it could do things like:
- Checking for method removals without corresponding method additions in
a slice aimed at recategorization (happened quite a few times...)
- Deprecations rather than removals of non-private methods

Cheers,
Henry

Den 01.01.2011 12:35, skrev Stéphane Ducasse:
> I took 5 min and wrote down that
> http://code.google.com/p/pharo/wiki/DeprecationProcess
>
> comment if you want.
>
>

Reply | Threaded
Open this post in threaded view
|

Re: fixing deprecation process once for all

Stéphane Ducasse

On Jan 3, 2011, at 9:34 AM, Henrik Johansen wrote:

> Sounds good to me, to me there's a missing when/why description though:
> Which criterias do we use for when deprecating is warranted?
>    - # of callers in image required to deprecate planned for removal (0?)
>    - # of callers / age when considering renaming for clarity (low
> number of callers and/or age < X years?)
>    - new/better API available (always, as long as new place documented?)
>
> As for the how,  since we do have an integration process, would it not
> also be desirable with some sort of integration lint tool working on slices?

yes now I do not time for that... :(

> For deprecations, it could do things like:
> - Checking for method removals without corresponding method additions in
> a slice aimed at recategorization (happened quite a few times...)
        YES!
> - Deprecations rather than removals of non-private methods

This is true that having something more automatic would be good. In fact I would like to have the rbengine available all the time.
but even if I said that I did not make any progress on using metacello to manage core/dev....

Stef

>
> Cheers,
> Henry
>
> Den 01.01.2011 12:35, skrev Stéphane Ducasse:
>> I took 5 min and wrote down that
>> http://code.google.com/p/pharo/wiki/DeprecationProcess
>>
>> comment if you want.
>>
>>
>