| 1 |
On Sun, Aug 18, 2013 at 11:48 PM, Rick "Zero_Chaos" Farina |
| 2 |
<zerochaos@g.o> wrote: |
| 3 |
> On 08/18/2013 11:10 PM, Rich Freeman wrote: |
| 4 |
>> On Sun, Aug 18, 2013 at 10:48 PM, William Hubbs <williamh@g.o> wrote: |
| 5 |
>>> Basically the status quo includes specific changes that were made in |
| 6 |
>>> our packaging to allow this sort of working separate /usr without |
| 7 |
>>> initramfs configuration to continue limping along, but we need to undo |
| 8 |
>>> them. |
| 9 |
>> |
| 10 |
>> Why do we have to undo them? |
| 11 |
>> |
| 12 |
>> I can understand that implementing them was a waste of time, but that |
| 13 |
>> time is already wasted. Does it take much effort to maintain the |
| 14 |
>> fixes? For all I know they do - but could you articulate this? |
| 15 |
>> |
| 16 |
>> I fully support that we should stop doing any more work to keep a |
| 17 |
>> separate /usr working without an initramfs. My question is why do we |
| 18 |
>> need to start undoing the work that has already been done? |
| 19 |
> |
| 20 |
> Because, the work that has already been done makes no bloody sense. |
| 21 |
> bzip2 is in / but not xz or lbzip2. |
| 22 |
|
| 23 |
As I said, I fully agree that the work that was already done makes no |
| 24 |
sense. The question is whether undoing that work makes any sense. If |
| 25 |
a city takes tax dollars to build a huge stadium I think that makes no |
| 26 |
sense. On the other hand once a stadium is built tearing it down |
| 27 |
makes no sense either - what is done is done. |
| 28 |
|
| 29 |
> |
| 30 |
> The cruelest prank of all of this bullspit, systemd is installed in / |
| 31 |
> despite many upstreams hardcoding THE CORRECT LOCATION of /usr. Yeah, |
| 32 |
> that's right, we are moving things to / and breaking systems and then |
| 33 |
> upstream just laughs at us when we report bugs. |
| 34 |
|
| 35 |
I said that I think anything not stable a year ago should be free to |
| 36 |
do whatever, which would include systemd. I don't think anybody who |
| 37 |
has been vocal on the issue really objects to moving systemd to /usr |
| 38 |
(and if you're going to move it you might as well do it before all the |
| 39 |
gnome users switch). |
| 40 |
|
| 41 |
> |
| 42 |
> It is not possible to keep systems running like this, and it is HARMING |
| 43 |
> us to even try. Please, stop moving things from /usr to /, and please |
| 44 |
> move things back where upstream expects them. Honestly I'm considering |
| 45 |
> a /usr merge on my system just to stop all this stupidity from breaking |
| 46 |
> my system. |
| 47 |
|
| 48 |
Can you give specific examples of how the current state of affairs is |
| 49 |
causing problems (especially for something other than systemd)? That |
| 50 |
is really what I'm getting at. I am fully capable of believing that |
| 51 |
moving stuff to / could be causing issues - I just haven't seen a |
| 52 |
single example on the lists in these discussions pointing one out. |
| 53 |
|
| 54 |
If you could point out some specific examples I think that this would |
| 55 |
really help your case. I'm pretty sympathetic to your arguments but |
| 56 |
even I'm on the fence about moving things back (as opposed to merely |
| 57 |
stopping the continued flow of stuff to /), and I suspect many on the |
| 58 |
Council will be far more skeptical. Some concrete examples of |
| 59 |
problems would probably go a long way to convincing everybody. |
| 60 |
|
| 61 |
If this is only about systemd feel free to point out some examples |
| 62 |
anyway so that we can at least get a vote to clarify whether systemd |
| 63 |
can move apart from the rest. |
| 64 |
|
| 65 |
Rich |
Archives