1 |
On Fri, Nov 09, 2012 at 07:21:25PM +0100, Fabian Groffen wrote: |
2 |
> On 09-11-2012 09:32:47 -0600, William Hubbs wrote: |
3 |
> > > Testing it. Testing the /usr move broadly. |
4 |
> > > You want gen_usr_ldscript on non-bootable systems (eg prefix, mingw) to |
5 |
> > > be (mostly) a noop too which you can't distinguish from only CHOST. |
6 |
> > |
7 |
> > That has been done already by toolchain a while back. Take a look at the |
8 |
> > code in toolchain-funcs.eclass. There are very few platforms where this |
9 |
> > function does anything at all these days. It would just be a matter of |
10 |
> > removing linux from the platforms it supports. |
11 |
> |
12 |
> They tested it? Are you referring to the platforms from bug #417451? |
13 |
> (excluding the Prefix platforms, because in the Prefix tree |
14 |
> gen_usr_ldscript just works, because it breaks existing installs to |
15 |
> disable it) |
16 |
|
17 |
Look at the first case statement in gen_usr_ldscript and tell me if you |
18 |
disagree with what I'm saying. The function always executes on darwin. |
19 |
On linux and any bsds, it executes for non-prefix setups. On any other |
20 |
platform, prefix or not, it does nothing. |
21 |
|
22 |
So, if I'm reading the code correctly, on prefix, more than likely it |
23 |
isn't doing anything. |
24 |
|
25 |
> > > IMHO this needs a discussion on -dev before going through the council. |
26 |
> > |
27 |
> > I would send the patch to do this to -dev anyway, so at that time I |
28 |
> > guess we could have folks apply it and rebuild their systems and see |
29 |
> > what breaks. |
30 |
> > |
31 |
> > Since applying the patch itself will not force any rebuilds, it should |
32 |
> > just end up being a natural migration; as things are rebuilt the |
33 |
> > libraries will move from /lib to /usr/lib with no harm being done. |
34 |
> |
35 |
> I thought so too, until I got dangling symlinks for older zlib for some |
36 |
> reason (preserve-libs? ebuild?) which wasn't really a funny experience. |
37 |
|
38 |
What symlinks did you have? |
39 |
|
40 |
> So, before this is done, we must test it in all forms, not rely on theory. |
41 |
|
42 |
I am not disagreeing with you. |
43 |
|
44 |
The only way to test is, after the time window for migrating to |
45 |
initramfs or sep-usr busybox expires, have people start migrating and |
46 |
see how it goes. |
47 |
|
48 |
> Alexis' profile suggestion nicely allows existing installs to remain as |
49 |
> is, new installs to be without /usr-split, and people to move over when |
50 |
> they deem that a good idea. |
51 |
|
52 |
If we do this with a variable, I suggest using it temporarily, but |
53 |
ultimately dropping it and having everyone switch over after things are tested. |
54 |
|
55 |
Again, I think 6 months or a year are too long for this testing window. |
56 |
|
57 |
William |