1 |
>>>>> On Mon, 18 May 2009, Ciaran McCreesh wrote: |
2 |
|
3 |
> On Mon, 18 May 2009 06:59:36 +0200 |
4 |
> Ulrich Mueller <ulm@g.o> wrote: |
5 |
>> AFAICS, there _is_ an ambiguity. You can have the following two |
6 |
>> ebuilds in the tree, simultaneously: |
7 |
|
8 |
>> ${PORTDIR}/app-misc/foo/foo-1a-scm.ebuild |
9 |
>> ${PORTDIR}/app-misc/foo-1a/foo-1a-scm.ebuild |
10 |
|
11 |
[Added some context back to your quotation of my posting.] |
12 |
|
13 |
> There's no ambiguity. It means what we define it to mean. |
14 |
|
15 |
Maybe it's possible to do that for dependencies, but VDB entries and |
16 |
binary packages for above two examples would still collide. |
17 |
|
18 |
So the conclusion still stands: |
19 |
|
20 |
>> The conclusion is that GLEP 54 in its current form is not |
21 |
>> implementable. |
22 |
|
23 |
Hyphens within PV are a Bad Thing, and we should really think about |
24 |
replacing the separator for "scm" by something else, like a period or |
25 |
an underscore. For example, the following two would be unique: |
26 |
|
27 |
${PORTDIR}/app-misc/foo/foo-1a_live.ebuild |
28 |
${PORTDIR}/app-misc/foo-1a/foo-1a-live.ebuild |
29 |
|
30 |
With our current versioning scheme the rule is very simple: ${P} is |
31 |
split into ${PN} and ${PV} at the last hyphen. This can be done in a |
32 |
straight forward way by regexp matching, and I would really hate to |
33 |
lose this nice property. |
34 |
|
35 |
Ulrich |