| 1 |
Graham Murray wrote: |
| 2 |
> AllenJB <gentoo-lists@××××××××××.uk> writes: |
| 3 |
> |
| 4 |
>> First of all, a tip: If a portage upgrade is available, do "emerge |
| 5 |
>> portage" first. New versions of portage often have new or improved |
| 6 |
>> features - in this case portage 2.1.6 includes, among other things, |
| 7 |
>> the ability to automatically handle most blockers. |
| 8 |
> |
| 9 |
> Though even the portage2.2 pre-releases do not handle all the cases that |
| 10 |
> should be able to be handled automatically. An example is one which |
| 11 |
> encountered yesterday - foo-x-y-z was already installed and foo-x-y+1-0 |
| 12 |
> was available for update. There are already installed packages which |
| 13 |
> have (R)DEPEND="=foo-x.y*" and others with (R)DEPEND=">=foo-x.0.0". So |
| 14 |
> the already installed foo-x.y.z satisfies all the depends, but the new |
| 15 |
> foo-x.y+1.0 does not. Yet 'emerge -auDv world' flagged a conflict of |
| 16 |
> trying to install two versions of an unslotted package - when the |
| 17 |
> 'obvious' resolution would be keep the already installed version and not |
| 18 |
> upgrade rather than requiring the user to manually mask the new |
| 19 |
> version. |
| 20 |
|
| 21 |
I disagree with you because when I tell emerge to update, it should |
| 22 |
update rather then not update :P |
Archives