| 1 |
On 09/29/2013 11:13 PM, William Hubbs wrote: |
| 2 |
> On Sun, Sep 29, 2013 at 09:21:01PM -0500, Daniel Campbell wrote: |
| 3 |
>>> /usr/lib/udev..... |
| 4 |
>>> /usr/lib/systemd..... |
| 5 |
>>> |
| 6 |
>>> were both placed in /usr despite objections from a number of folks. |
| 7 |
>>> |
| 8 |
>>> So claims that udev and systemd are not responsible are not true. |
| 9 |
> |
| 10 |
> Udev is installed in / in gentoo. I am a co-maintainer of udev and that |
| 11 |
> was fixed quite some time back, it is the Gentoo systemd team that installs |
| 12 |
> their version of udev in /usr. |
| 13 |
> |
| 14 |
> Installing udev or eudev, however, doesn't really solve the issue |
| 15 |
> though, because it is possible to run arbitrary programs from within |
| 16 |
> udev rules. |
| 17 |
> |
| 18 |
> Another unrelated concern is if you install a program in / that needs to |
| 19 |
> access something in /usr/share, this will be broken by not having /usr |
| 20 |
> mounted. This means that, for example, the locale logic of most software |
| 21 |
> can't work without /usr since it accesses files in /usr/share/locale. |
| 22 |
> |
| 23 |
> William |
| 24 |
> |
| 25 |
All that is required is that the programs and libraries necessary to |
| 26 |
locate and mount root and then to find and mount other filesystems be in |
| 27 |
root. This was a fundamental piece of the design of UNIX and was |
| 28 |
inherited by many UNIX derived systems. While debugging System IV |
| 29 |
systematizing, we had meetings with the folks from Murray Hill and they |
| 30 |
insisted that this had to be maintained. (circa 1982) |
| 31 |
|
| 32 |
The actual details of what has broken are not as important as the fact |
| 33 |
that the breakage happened and complaints about the breakage have be |
| 34 |
dismissed with disrespect and disparaging remarks about clinging to |
| 35 |
ancient history. That has been done at least twice in this set of threads. |
| 36 |
|
| 37 |
Certainly, Linux is an evolving and growing system, but there seems to |
| 38 |
be no natural selection process to cull the things that don't work. |
| 39 |
Change is NOT the problem, without change there can be no progress. But |
| 40 |
to change something for no good reason, simply to change it, is not healthy. |
| 41 |
|
| 42 |
-- |
| 43 |
G.Wolfe Woodbury |
| 44 |
redwolfe@×××××.com |
Archives