1 |
Ulrich Mueller: |
2 |
>>>>>> On Sun, 27 Jul 2014, Martin Vaeth wrote: |
3 |
> |
4 |
>> Kent Fredric <kentfredric@×××××.com> wrote: |
5 |
>>> -r1.1 weirdness feels like it may cause more problems than it solves. |
6 |
> |
7 |
>> So far, nobody pointed out any problem which would be caused by -r1.1. |
8 |
>> Which is not surprising, since the idea is that -r1.1 cannot be |
9 |
>> distinguished from -r2: |
10 |
>> It is only a hint to the PM that he *may* shortcut certain phases when |
11 |
>> updating from -r1. |
12 |
> |
13 |
> I wonder if it wouldn't be saner to leave our revision syntax |
14 |
> untouched. |
15 |
> |
16 |
> Instead, one could introduce a variable INSTALL_VERSION that would |
17 |
> default to ${PVR} but could be set to the version of a previous ebuild |
18 |
> instead. The PM could compare it against INSTALL_VERSION in the VDB |
19 |
> and skip build and installation if versions match. |
20 |
> |
21 |
> Advantages: |
22 |
> - Support for the variable could be optional. PMs not supporting it |
23 |
> would do a regular revbump instead. |
24 |
> - One could even imagine USE-conditional syntax for the variable. |
25 |
> |
26 |
|
27 |
Aren't we putting a lot of trust in ebuild writers here? If any, I'd say |
28 |
this should definitely be an optional feature for portage as well and |
29 |
default to off. |
30 |
|
31 |
The only time I would really consider using this would be for the very |
32 |
few time-intensive packages I maintain like paraview, rstudio or |
33 |
blender. For the rest, it's too much effort. |
34 |
|
35 |
It will probably also cause confusion for comaintainers and |
36 |
collaborators, especially when INSTALL_VERSION points to a version that |
37 |
has already been removed. |