1 |
On Thu, Dec 27, 2012 at 1:22 PM, Mark David Dumlao <madumlao@×××××.com>wrote: |
2 |
|
3 |
> On Fri, Dec 28, 2012 at 12:28 AM, Kevin Chadwick <ma1l1ists@××××××××.uk> |
4 |
> wrote: |
5 |
> > |
6 |
> > Again you don't break the spec unless you have to and you don't change |
7 |
> > the spec unless it is an improvement or you have no choice. Non of |
8 |
> > which is the case. Just like you do not mould a mail RFC to a |
9 |
> > widely used technically inferior hotmail implementation. |
10 |
> |
11 |
> The spec - or implementation - of / and /usr separation is broken and |
12 |
> has been for quite a while now. Nobody here's even bothered answering |
13 |
> how the modern Gentoo distro / sysad would survive /usr being out of |
14 |
> sync with /, for instance, |
15 |
|
16 |
|
17 |
If the basics are kept in /, with prod-additionals kept in /usr, then you |
18 |
should be able to boot to "basics", and restore /usr. |
19 |
|
20 |
|
21 |
> or the fact that some udev programs tend to |
22 |
> be located in /usr, |
23 |
|
24 |
|
25 |
That's either a bug with those programs, or a need for architectural |
26 |
improvements within udev. Both plausible answers. |
27 |
|
28 |
|
29 |
|
30 |
> or even just a solid detailed specification on the |
31 |
> precise criteria for inclusion into /. |
32 |
|
33 |
|
34 |
For anyone arguing that / and /usr should be separate, the answer to this |
35 |
is "that ought to be common sense." |
36 |
|
37 |
Since I'm not someone who knows all there is to know about the software and |
38 |
interactions thereof, the most I can say is: |
39 |
|
40 |
* / ought to contain all binaries, libraries and static data necessary for |
41 |
booting beyond the point where / is mounted, and any machine-specific |
42 |
binaries, libraries and static data. |
43 |
* /usr ought to contain all binaries, libraries and static data not |
44 |
necessary for its own mount. |
45 |
|
46 |
|
47 |
> Even the FHS is mum on all the |
48 |
> extra crap we randomly decide between / and /usr to land in. |
49 |
|
50 |
|
51 |
So fix it. FHS was a document written to say "we have a standard" that |
52 |
happened to map almost cleanly to all the implementations of the day. Kinda |
53 |
like how SQL mapped "almost cleanly" to the existing RDBMSs that existed |
54 |
when it was introduced. Such is how standards documents are born. |
55 |
|
56 |
|
57 |
> You'd |
58 |
> think, for instance, something as clear cut as filesystem manipulation |
59 |
> tools, e.g., xfs_admin, would belong in /sbin rather than /usr/sbin. |
60 |
> But no it's not. Or - for crying out loud, at least a text editor that |
61 |
> isn't ed. |
62 |
> |
63 |
|
64 |
I'd say that warrants bug reports against those programs. Also, isn't |
65 |
busybox under /? I think it has nano built-in. |
66 |
|
67 |
|
68 |
> |
69 |
> Again, the broken state of the / and /usr split is a different thing |
70 |
> from the usefulness state of your own already installed distro. |
71 |
> |
72 |
> TLDR: The spec is broken. |
73 |
> |
74 |
|
75 |
It's not that the spec is broken. It's that the spec doesn't lay out every |
76 |
single detail imaginable, and as a consequence, people assuming that the |
77 |
spec should be able to do their thinking for them assume the spec is broken |
78 |
when it's silent on a given query. |
79 |
|
80 |
-- |
81 |
:wq |