1 |
On Thu, Jan 17, 2013 at 12:14 PM, Ciaran McCreesh |
2 |
<ciaran.mccreesh@××××××××××.com> wrote: |
3 |
> On Thu, 17 Jan 2013 12:03:36 -0500 |
4 |
> James Cloos <cloos@×××××××.com> wrote: |
5 |
>> Every current category matches /^[a-z]+-[a-z]+$/. With the possible |
6 |
>> exception of adding moving from [a-z]+ to [a-z0-9]+, that shoud |
7 |
>> remain. |
8 |
> |
9 |
> Untrue. 'virtual' doesn't. If you want the rules for what constitutes a |
10 |
> valid category name, consult PMS. If you want to know what categories |
11 |
> are actually present, consult 'profiles/categories' or your package |
12 |
> mangler. |
13 |
|
14 |
Tend to agree. We should use whatever makes the most sense. I'm not |
15 |
sure how many packages we're actually talking about though - might |
16 |
make sense to introduce a new category when we need it. |
17 |
|
18 |
There are a lot of assumptions people make which aren't backed by PMS. |
19 |
Probably the more common one is the concept that EAPIs are numerical |
20 |
and/or orderable. The whole concept of the "best/newest" EAPI depends |
21 |
on that assumption. |
22 |
|
23 |
Rich |