1 |
On 12 March 2012 20:10, Ciaran McCreesh <ciaran.mccreesh@××××××××××.com> wrote: |
2 |
> On Mon, 12 Mar 2012 20:49:22 +0100 |
3 |
> Ulrich Mueller <ulm@g.o> wrote: |
4 |
>> > That's already not the way things work, since different version |
5 |
>> > strings can be equal versions (and it's illegal to do this), |
6 |
>> > so it's not relevant to the discussion. |
7 |
>> |
8 |
>> This is a design flaw in our versioning system, and it can only occur |
9 |
>> in some corner cases where version components have leading zeros. |
10 |
> |
11 |
> You know, if we had GLEP 55, we could fix that. Although it's debatable |
12 |
> whether it's a flaw, since we're trying to match upstream version |
13 |
> formats where reasonably possible. |
14 |
> |
15 |
> -- |
16 |
> Ciaran McCreesh |
17 |
|
18 |
I'm sure that it's been considered already, but what are the arguments |
19 |
against embedding the EAPI on a per-package (default) or per-version |
20 |
basis in metadata.xml. It IS metadata after all. |
21 |
|
22 |
Something like: |
23 |
<eapi>6</eapi> <!-- Default EAPI unless a versioned tag exists. --> |
24 |
|
25 |
<eapi version="=1.2.*">4</eapi> |
26 |
<eapi version="<1.2">3</eapi> |
27 |
|
28 |
Where the version attribute should presumably have the same syntax as |
29 |
versions in package.* already. |