1 |
On 5/31/19 9:13 AM, Peter Humphrey wrote: |
2 |
> On Thursday, 30 May 2019 16:32:01 BST »Q« wrote: |
3 |
> |
4 |
>> The 17.1 profiles are soon to be marked stable, so I went ahead and |
5 |
>> migrated a little over a week ago, following the draft news item Michał |
6 |
>> Górny recently posted to -dev. FWIW, the migration seemed to go |
7 |
>> smoothly and I haven't noticed anything breaking except |
8 |
>> app-office/kmymoney won't build, apparently because the ebuild expects |
9 |
>> something to be in /lib which isn't there any more. But one dev said |
10 |
>> it builds fine on his 17.1 test system, so I dunno. I filed a bug, |
11 |
>> <https://bugs.gentoo.org/686750>. |
12 |
> Encouraged by this, I tried the migration this morning. I followed the enews |
13 |
> item that Mick quoted - I even printed it to keep myself straight. |
14 |
> |
15 |
> Before the start: |
16 |
> $ eselect profile show |
17 |
> Current /etc/portage/make.profile symlink: |
18 |
> default/linux/amd64/17.0/desktop/plasma |
19 |
> |
20 |
> I went through the migration, step by step, but after step 7, "unsymlink-lib |
21 |
> --finish", I found 110 files still in /lib and /usr/lib (list attached), of |
22 |
> the >3000 before the migration attempt, and /usr/local/lib still a symlink. |
23 |
> This can't be right, can it? |
24 |
|
25 |
It's perfectly OK for there to still be stuff under /lib /usr/lib |
26 |
/usr/local/lib. These are supposed to be things which are arch |
27 |
independent, such as config stuff and scripts. However, I'm pretty sure |
28 |
none of them should still be a symlink. What's the link pointing to? |
29 |
I'd be tempted to revert the unsymlink. Did you check the output of |
30 |
"unsymlink-lib --analyze" first? |