1 |
On Sun, Sep 29, 2013 at 12:24 PM, Tanstaafl <tanstaafl@×××××××××××.org> wrote: |
2 |
> On 2013-09-29 10:57 AM, Bruce Hill <daddy@×××××××××××××××××××××.com> wrote: |
3 |
>> |
4 |
>> On Sun, Sep 29, 2013 at 10:20:49AM -0400, Tanstaafl wrote: |
5 |
>>> |
6 |
>>> On 2013-09-28 8:30 AM, Bruce Hill <daddy@×××××××××××××××××××××.com> |
7 |
>>> wrote: |
8 |
>>>> |
9 |
>>>> This does not mean that on November 1 your system will not be able to |
10 |
>>>> boot. |
11 |
>>>> Its simply means that beginning November 1, Gentoo devs are not required |
12 |
>>>> to |
13 |
>>>> jump through hoops to make apps work on systems with /usr separate from |
14 |
>>>> /. |
15 |
>>>> |
16 |
>>>> Now, what are you going to do? That's the question. |
17 |
>>> |
18 |
>>> |
19 |
>>> This won't necessarily be the end of the worl, if, and ONLY if any and |
20 |
>>> all ebuild mainteainers are REQUIRED to provide very large and scary |
21 |
>>> warnings if they change something that will cause any systems with a |
22 |
>>> separate /usr and NO initramfs to fail to boot. |
23 |
>> |
24 |
>> |
25 |
>> The news item *IS* the warning. |
26 |
> |
27 |
> |
28 |
> Oh for fucks sake... BULLSHIT. |
29 |
> |
30 |
> If an ebuild maintainer changes something that will BREAK BOOTING on systems |
31 |
> that violate the 'no separate /usr without an initramfs' rule, what in the |
32 |
> FUCK is the problem with requiring them to WARN PEOPLE? |
33 |
|
34 |
The news item allows developers to assume that /usr is available from |
35 |
early boot. Therefore, they *could* be breaking *some* setups, and |
36 |
they will not even realize it. That is the beauty of having /usr |
37 |
available from early boot: it frees developers from thinking in all |
38 |
kind of different setups and combinations (it is on LVM? it uses raid? |
39 |
what level? it's on NFS? do I need a special filesystem?), so they can |
40 |
work in bringing more awesomeness into Gentoo. |
41 |
|
42 |
They cannot put a warning if they don't know something will break |
43 |
*some* setups. And the whole point of this is that they don't have to |
44 |
consider every single possible combination of setups; the point is not |
45 |
to force you to have an initramfs. |
46 |
|
47 |
The point is to guarantee early /usr availability. |
48 |
|
49 |
Regards. |
50 |
-- |
51 |
Canek Peláez Valdés |
52 |
Posgrado en Ciencia e Ingeniería de la Computación |
53 |
Universidad Nacional Autónoma de México |