1 |
On 10/12/2013 01:34, Grant Edwards wrote: |
2 |
> On 2013-12-09, Adam Carter <adamcarter3@×××××.com> wrote: |
3 |
>>> |
4 |
>>> I understand that sometimes a maintainer decides to add a feature that |
5 |
>>> requires some new dependancies, but why three different versions of |
6 |
>>> Ruby all of a sudden? |
7 |
> |
8 |
>> That's the default if you havent specified which version of ruby you |
9 |
>> want, via RUBY_TARGETS in make.conf. |
10 |
> |
11 |
> That seems broken to me. |
12 |
> |
13 |
> Are there packages that require all three versions of Ruby? |
14 |
> |
15 |
|
16 |
|
17 |
|
18 |
No, there are packages that can use all versions but I know of none that |
19 |
require them all. |
20 |
|
21 |
Python is in a similar position with 2.x and 3.x plus all the other |
22 |
implementations too. Portage has no way of knowing what pythons you |
23 |
have, need, or want to use. So the logic is: |
24 |
|
25 |
If you specify in make.conf which versions you want, you get those |
26 |
versions, otherwise you get all versions. Individual ebuilds that use |
27 |
ruby may specify which versions they will be built against, so that you |
28 |
can chop and change the ruby slot in use and your ruby apps still work. |
29 |
|
30 |
Personally, I'm getting a little tired of this eternal fascination with |
31 |
the latest greatest flavour of the day interpreter just because it's |
32 |
shiny and new. And rub devs are amongst the worst I have seen anywhere |
33 |
(worse than php and that's saying something). I know a little of what I |
34 |
talk about, the row of desks behind me has perl, python, php and ruby |
35 |
devs who belong to teams with definite preferences as to the |
36 |
implementation language. Matching code and app quality with experience |
37 |
and language used is an interesting mapping exercise. |
38 |
|
39 |
-- |
40 |
Alan McKinnon |
41 |
alan.mckinnon@×××××.com |