1 |
On Wed, Jul 4, 2018 at 12:20 PM gevisz <gevisz@×××××.com> wrote: |
2 |
> |
3 |
> 2018-07-03 16:04 GMT+03:00 Rich Freeman <rich0@g.o>: |
4 |
> |
5 |
> > emerge --sync, with git at least, overwrites /usr/portage in place and |
6 |
> > so it will leave it in a bad state if verification fails. |
7 |
> |
8 |
> It sounds really aweful. |
9 |
> I did not know this as I always used only emerge-webrsync. |
10 |
> |
11 |
|
12 |
Yes. I do have scripts set to abort if it fails and I should notice. |
13 |
But, this isn't really viewed as acceptable by anybody - there just |
14 |
hasn't been a more complete solution devised. |
15 |
|
16 |
Part of the issue is that I think portage just has git do a pull |
17 |
in-place which just overwrites everything, and then portage does the |
18 |
signature check after that. |
19 |
|
20 |
I wonder if we can have portage instead do a fetch, then do the |
21 |
verification of HEAD, and then if it passes do a checkout. That will |
22 |
still leave you with invalid data in the git history, but it won't |
23 |
actually be checked out, so at least emerge won't be seeing it. |
24 |
|
25 |
-- |
26 |
Rich |