| 1 |
On Fri, 2 Jun 2006 01:46:09 +0200, Hemmann, Volker Armin wrote: |
| 2 |
|
| 3 |
> --deep is dangerous! |
| 4 |
> |
| 5 |
> I have stopped using --deep ages ago. |
| 6 |
> As an example: |
| 7 |
> |
| 8 |
> there is an --deep update for libFOO.1 to libFOO.1.1. |
| 9 |
> |
| 10 |
> You make this update which only shows up with --deep |
| 11 |
> |
| 12 |
> Suddenly all apps, linking to libFOO.1 are dead or crashy or acting |
| 13 |
> weired. |
| 14 |
|
| 15 |
Dropping --deep won't stop that happening, only delay it. sooner or |
| 16 |
later, one of your packages will need libFOO.1.1 and it will be |
| 17 |
installed. --deep doesn't cause this problem, it only affects the timing. |
| 18 |
|
| 19 |
> That happened to me several times. I see NO reason to use deep. Ever. |
| 20 |
|
| 21 |
How about this instance? The OP wants all packages affected by the |
| 22 |
profile change to be updated. Without --deep, that won't happen. |
| 23 |
|
| 24 |
> Reduced the occurences where I have to use revdep-rebuilt to almost nil |
| 25 |
> (except that expat tragedy some weeks ago. Man that sucked ;) ). |
| 26 |
|
| 27 |
I do a deep update every day, on various architectures. I run |
| 28 |
revdep-rebuild -p occasionally, just to make sure everything is |
| 29 |
consistent, it rarely picks up anything. |
| 30 |
|
| 31 |
--deep is an option, and I understand why you choose not to use it, but |
| 32 |
on this occasion it is necessary to accomplish the OP's goal. |
| 33 |
|
| 34 |
|
| 35 |
-- |
| 36 |
Neil Bothwick |
| 37 |
|
| 38 |
Machine-independent: Does not run on any existing machine. |
Archives