1 |
On Mon, 31 Jan 2011 19:23:51 +0100 |
2 |
Ulrich Mueller <ulm@g.o> wrote: |
3 |
> PMS currently says: "A whitespace-separated list of versions of this |
4 |
> package (including revision, if specified) that are being replaced |
5 |
> (uninstalled or overwritten) as a result of this install." |
6 |
> |
7 |
> Shouldn't the spec guarantee that it's the empty string (i.e., not |
8 |
> a string containing arbitrary whitespace) if the list is empty? |
9 |
> Otherwise simple binary tests like [[ -n ${REPLACING_VERSIONS} ]] |
10 |
> cannot be done. |
11 |
> |
12 |
> For REPLACED_BY_VERSION, it is explicitly specified: "The single |
13 |
> version ..., or an empty string otherwise." |
14 |
|
15 |
We had an issue with this early on with Paludis: certain eclasses were |
16 |
expecting exact spacing requirements inside A and AA based upon a |
17 |
quirk in how Portage created them, and it was causing weird bugs. I'd |
18 |
be inclined to say that people should write more tolerant code... |
19 |
|
20 |
-- |
21 |
Ciaran McCreesh |