Recommendations for version numbers in port names

[Joshua Root]

On 2014-11-16 06:04 , Bradley Giesbrecht wrote:
> 
> On Nov 12, 2014, at 8:53 AM, Joshua Root <jmr at macports.org> wrote:
> 
>> On 2014-11-12 17:14 , Lawrence Velázquez wrote:
>>> On Nov 10, 2014, at 3:02 PM, Bradley Giesbrecht
>>> <pixilla at macports.org> wrote:
>>> 
>>>> I don't know that I understand these variant checks, they
>>>> seem to not produce the same result in all cases:
>>>> 
>>>> http://trac.macports.org/browser/tags/release_2_3_2/base/src/port/port.tcl#L289
>>>>
>>>> 
http://trac.macports.org/browser/tags/release_2_3_2/base/src/port/port.tcl#L1752
>>>> http://trac.macports.org/browser/tags/release_2_3_2/base/src/macports1.0/macports.tcl#L730
>>>>
>>>> 
http://trac.macports.org/browser/tags/release_2_3_2/base/src/port1.0/portutil.tcl#L2289
>>>> 
>>>> Would a proc for testing variant name sanity be worth the
>>>> effort?
>>> 
>>> It would probably be a good idea to decide on an authoritative
>>> specification and consolidate verification in a single place.
>> 
>> I think you mean validation? We don't actually do that anywhere
>> at the moment. All the places above are just parsing.
> 
> If we did create authoritative validation service, would
> portutil.tcl be a good location?

Probably not, that would make it only accessible from portfile
interpreter contexts.

- Josh