1 |
>>>>> On Wed, 4 Nov 2009, Ciaran McCreesh wrote: |
2 |
|
3 |
>> Part of the problem (what you call "insufficient clarity") is that |
4 |
>> the proposal's original intention was to cover only the merge |
5 |
>> process, i.e. what takes place after pkg_preinst. Whereas you want |
6 |
>> to extend it to include everything that is taking place after |
7 |
>> src_install (for Portage, prepstrip and whatnot). |
8 |
>> |
9 |
>> If you limit it to the final merge process from D to ROOT, then the |
10 |
>> answer is easy, namely mtimes of all regular files must be |
11 |
>> preserved. |
12 |
|
13 |
> What I want is for the proposal to be sufficiently specific that it |
14 |
> covers exactly what the package manager can and cannot do, and what |
15 |
> ebuilds can and cannot rely upon happening. If you require mtime |
16 |
> preservation between pkg_preinst and the merge to /, the package |
17 |
> manager can just screw things up (by implementing reasonable |
18 |
> features) elsewhere. It is by no means clear to me that merely |
19 |
> requiring mtime preservation from after pkg_preinst to before |
20 |
> pkg_postinst, and allowing arbitrary mtime tinkering elsewhere, is |
21 |
> what is desired. |
22 |
|
23 |
Can you try to find a suitable wording? Otherwise, it's not clear to |
24 |
me how the council could resolve the issue during the next meeting. |
25 |
|
26 |
(And as my suggested wording [1] caused some unfortunate discussion, |
27 |
I don't feel like I should come up with a new one myself.) |
28 |
|
29 |
> As an example for the above, is it legal for a package manager to |
30 |
> rewrite any mtime that is before the start of the build process if |
31 |
> it does it after src_install but before pkg_preinst? |
32 |
|
33 |
So you really want this? ;-) My personal opinion is that it wouldn't |
34 |
break anything and we could therefore declare it as an allowed QA |
35 |
measure. And if it takes place before pkg_preinst then the ebuild |
36 |
could override it in special cases. |
37 |
|
38 |
But please be aware that the council (October meeting) has voted |
39 |
against this sort of mtime fixup. |
40 |
|
41 |
Ulrich |
42 |
|
43 |
[1] <http://bugs.gentoo.org/264130#c39> and following comments |