| 1 |
* On Jun 27, 2002, Jani Monoses <jani@××.ro> said: |
| 2 |
> After looking at why emerge clean wouldn't get rid of older hdparm and openssh I found that |
| 3 |
> the ebuilds for the old versions didn't have a SLOT so portage assumes nothing != 0 and |
| 4 |
> keeps both versions. |
| 5 |
> But aren't all ebuilds supposed to have SLOT defined? |
| 6 |
> Maybe temporarily until all such cases are fixed emerge could presume missing slot means 0? |
| 7 |
|
| 8 |
Generally speaking, it is better to assume too little than too much. |
| 9 |
Let's say that there is a library that you want to keep around without a |
| 10 |
slot and one of the same name at slot 0 (hypothetically, because I can't |
| 11 |
come up with an actual instance of this). Each and every time you try |
| 12 |
to clean your packages, it annoyingly uninstalls the unslotted package. |
| 13 |
On the other hand, you notice that one package didn't get uninstalled |
| 14 |
that you think should have been. You go ahead and uninstall the package |
| 15 |
and never hear from it again. That's why the current model is safer, at |
| 16 |
the cost of a single minor annoyance as opposed to repeated annoyance. |
| 17 |
|
| 18 |
But I do agree that all packages should be slotted and, if I'm not |
| 19 |
mistaken, they'll all have slots eventually. It's just a matter of |
| 20 |
waiting it out. |
| 21 |
|
| 22 |
-- |
| 23 |
/* |
| 24 |
* Bob Phan <bob@××××××××××××××××.net,rphan@××××.com> |
| 25 |
* Computational Chemistry Informatics |
| 26 |
* Neurogen Corporation |
| 27 |
* (203)488-8201 x4645 |
| 28 |
* |
| 29 |
* To understand recursion, you must first understand recursion. |
| 30 |
* http://www.endlessrecursion.net/ |
| 31 |
*/ |
Archives