1 |
Alec Warner wrote: |
2 |
> Jakub Moc wrote: |
3 |
>> Ciaran McCreesh napsal(a): |
4 |
>>> On Sun, 22 Oct 2006 02:18:52 -0700 David Shakaryan <omp@g.o> |
5 |
>>> wrote: |
6 |
>>> | I have created a small script to go through entries in package.mask |
7 |
>>> | and list those which are masking non-existent packages or versions. I |
8 |
>>> | then used this list to clean up package.mask. I tried to only remove |
9 |
>>> | versions that were removed and have a newer version in place, along |
10 |
>>> | with packages that were removed, but I accidentally /might/ have |
11 |
>>> | removed other entries, although I doubt it. I kept masks for future |
12 |
>>> | versions in place, as the maintainer of the package might have wanted |
13 |
>>> | to mask it ahead of time. |
14 |
>>> |
15 |
>>> So what happens when users have an old, masked package installed that's |
16 |
>>> no longer masked thanks to this change? |
17 |
>> |
18 |
>> Err, exactly nothing? If they didn't unmerge it, they'll continue to |
19 |
>> have it installed as they did before? |
20 |
>> |
21 |
>> |
22 |
>> |
23 |
> |
24 |
> For things like security packages; it is troublesome. |
25 |
> |
26 |
> 1.x has a sec vuln but 2.x fixes it; upstream isn't willing to backport |
27 |
> and both stay in the tree. So we mask 1.x for sec reasons. |
28 |
|
29 |
It seems like you didn't understand exactly what I did. The masks I |
30 |
removed are *ONLY* those which are masking a package or version that is |
31 |
no longer in the tree. |
32 |
|
33 |
- If 2.x and 1.x are both in the tree, and one of them is masked, I |
34 |
didn't touch the mask. |
35 |
- If only 2.x is in the tree, and 1.x is masked, then I removed the mask |
36 |
as it is quite useless. |
37 |
- If only 1.x is in the tree, and 2.x is masked, I let the mask stay in |
38 |
case the developer masked it ahead of time. |
39 |
|
40 |
Hope that clears it up. |
41 |
|
42 |
-- |
43 |
David Shakaryan |
44 |
GnuPG Public Key: 0x4B8FE14B |