1 |
On Sunday 28 June 2009 05:47:23 Volker Armin Hemmann wrote: |
2 |
> On Sonntag 28 Juni 2009, Alex Schuster wrote: |
3 |
> > Volker Armin Hemmann writes: |
4 |
> > > On Sonntag 28 Juni 2009, Alex Schuster wrote: |
5 |
> > >> Or keep 4.3 as default, I don't think you could run into problems. |
6 |
> > > |
7 |
> > > he will over time. If you switch default compiler emerge -s world has |
8 |
> > > to be done. |
9 |
> > |
10 |
> > According to Alan McKinnon's (and my own experience), this is not |
11 |
> > necessary, unless there are ABI changes. But there were none between 4.1 |
12 |
> > and 4.3. |
13 |
> > |
14 |
> > http://www.mail-archive.com/gentoo-user@l.g.o/msg83724.html |
15 |
> > |
16 |
> > Wonko |
17 |
> |
18 |
> you don't have to compile between 4.2.0 and 4.2.1 - sure. |
19 |
> |
20 |
> But with 4.2 to 4.3 I only got a stable system after compiling everything |
21 |
> with the same compiler. So whatever Alan says - I know how borked my box |
22 |
> was with half of the libs compiled by one compiler and the rest by the |
23 |
> other. |
24 |
|
25 |
That's interesting. I run ~amd64 here and update almost daily - so I got |
26 |
practically every gcc version that hit the tree since 3.3 at some stage. And I |
27 |
never had the problem you describe. |
28 |
|
29 |
It's likely that you have a set of libs that indeed *are* sensitive to |
30 |
different gcc versions, and I'm not using those libs (so I don't get the |
31 |
problems). |
32 |
|
33 |
I wonder if it would be worth the effort to investigate this further and |
34 |
isolate problem packages. It seems that only Gentoo (and derivs) have this |
35 |
problem - binary distros rebuild everything from scratch with each new |
36 |
releases (equivalent to emerge -e world) so they tend to never run into these |
37 |
issues. |
38 |
|
39 |
-- |
40 |
alan dot mckinnon at gmail dot com |