1 |
On Mon, Nov 19, 2012 at 09:39:59AM -0500, Rich Freeman wrote: |
2 |
> On Sun, Nov 18, 2012 at 10:38 PM, Joshua Kinard <kumba@g.o> wrote: |
3 |
> > Correct me if wrong, but didn't the issue start with udev wanting to put the |
4 |
> > PCI ID database/file into /usr/share from /etc? |
5 |
> |
6 |
> Well, I can't vouch for what the first issue that arose was, but I do |
7 |
> recall discussion that bluetooth keyboards also required libraries in |
8 |
> /usr. |
9 |
> |
10 |
> Right now it takes less common situations to trigger problems with a |
11 |
> separate /usr, but the general trend is not favorable. It isn't just |
12 |
> software changes either - if I told somebody 15 years ago that USB |
13 |
> keyboards were going to be problematic but standard PS2 keyboards |
14 |
> would always work just fine, they'd say there was nothing to worry |
15 |
> about since PS2 keyboards would always be around. For all we know in |
16 |
> 5 years you won't be able to buy a USB keyboard. |
17 |
> |
18 |
> So, udev has been pushing things along to some extent, but the problem |
19 |
> is definitely bigger than udev. |
20 |
|
21 |
I'm glad someone else on this list finally realizes that udev did not break |
22 |
separate /usr on its own. I've been trying to explain this to people |
23 |
here for ages. |
24 |
|
25 |
It isn't just programs that use libraries in /usr/lib that are broken. |
26 |
Any program in early boot that tries to access data files in /usr/share |
27 |
before /usr is mounted is broken, so for example, locales do not |
28 |
work before /usr is mounted. |
29 |
|
30 |
William |