1 |
On Thu, 1 Oct 2009 16:49:43 +0200 |
2 |
Ulrich Mueller <ulm@g.o> wrote: |
3 |
> > Why go with an inferior solution? Why not go with a solution that |
4 |
> > requires the package manager to fix broken mtimes? |
5 |
> |
6 |
> Because it would be non-zero implementation cost for Portage, so |
7 |
> probably out of question for EAPI 3. |
8 |
|
9 |
It's cheap, and it's doing it the right way. If we were to design the |
10 |
feature up-front rather than going with whatever Portage does, we'd go |
11 |
with mtime fixing. |
12 |
|
13 |
> And it's not at all clear if the solution is inferior. Since half a |
14 |
> year, nobody cared to answer the question of comment 25 of mentioned |
15 |
> bug. |
16 |
|
17 |
Because comment 25 is entirely missing the point. The objection is not |
18 |
to preservation. The objection is to pure preservation with no handling |
19 |
for dodgy mtimes. |
20 |
|
21 |
> > Also, what are the rules regarding this and things like stripping |
22 |
> > and other fixes and changes that the package manager performs upon |
23 |
> > files before merging them? |
24 |
> |
25 |
> This is outside the scope of this proposal, and (at least for now) I'm |
26 |
> not going to work anything out. |
27 |
|
28 |
It's not. It's a necessary part of the proposal. You need to define the |
29 |
behaviour here, since if you don't, we're back to ebuilds relying upon |
30 |
undefined behaviour. |
31 |
|
32 |
What you're effectively saying by ignoring this is "mtimes must be |
33 |
preserved, except when they're not". That isn't good enough, since it |
34 |
would be entirely legal for a package manager to not do any |
35 |
preservation at all then. Alternatively, you're saying "mtimes must |
36 |
always be preserved", in which case Portage isn't compliant. This isn't |
37 |
something you can just ignore. |
38 |
|
39 |
-- |
40 |
Ciaran McCreesh |