| 1 |
On 04/03/2017 14:33, Peter Humphrey wrote: |
| 2 |
> On Saturday 04 Mar 2017 12:42:38 Alan McKinnon wrote: |
| 3 |
>> On 04/03/2017 11:45, Peter Humphrey wrote: |
| 4 |
>>> Greetings. |
| 5 |
>>> |
| 6 |
>>> This is a KDE ~amd64 box, which still depends on kde-apps/kdepimlibs:4 |
| 7 |
>>> for many things. I've been running happily on =app-crypt/gpgme-1.5.5, |
| 8 |
>>> but today portage wants to upgrade to =app-crypt/gpgme-1.8.0-r2. But |
| 9 |
>>> that version can't be installed while kdepimlibs:4 is still around, as |
| 10 |
>>> I see from the ebuild: |
| 11 |
>>> |
| 12 |
>>> [...] |
| 13 |
>>> RDEPEND="${COMMON_DEPEND} |
| 14 |
>>> |
| 15 |
>>> cxx? ( |
| 16 |
>>> |
| 17 |
>>> !kde-apps/gpgmepp |
| 18 |
>>> !kde-apps/kdepimlibs:4 |
| 19 |
>>> |
| 20 |
>>> )" |
| 21 |
>>> |
| 22 |
>>> [...] |
| 23 |
>>> |
| 24 |
>>> Is it even possible to satisfy that condition? |
| 25 |
>> |
| 26 |
>> No, but you can mask app-crypt/gpgme-1.8.0-r2 and stay on 1.5/1.6 |
| 27 |
> |
| 28 |
> Yes, I've done that, but why has such silliness been allowed out? |
| 29 |
|
| 30 |
Maybe that combination never got tested? It's a ~arch package, so you |
| 31 |
get to be a field tester when you use it :-) |
| 32 |
|
| 33 |
|
| 34 |
-- |
| 35 |
Alan McKinnon |
| 36 |
alan.mckinnon@×××××.com |
Archives