1 |
On Sat, 26 Sep 2015 15:10:48 +0200, lee wrote: |
2 |
|
3 |
> It seems that everyone has the problem that some versions of some |
4 |
> packages don't go together with some versions of other packages the |
5 |
> 'some versions of some packages' depend on. |
6 |
|
7 |
That's just life, and 99% of th time it either doesn't matter or is |
8 |
handled by slots. A and B depend on C. A new version of C comes out but A |
9 |
can't work with it, so portage doesn't update it. B is still happy because |
10 |
it always worked with the older C, portage just tells you why it hasn't |
11 |
updated C. |
12 |
|
13 |
> Then emerge comes along and points this out as an extremely serious |
14 |
> problem while all it takes to solve this problem is someone convincing |
15 |
> the person observing what emerge does that the apparently serious |
16 |
> problems aren't relevant at all. |
17 |
|
18 |
It didn't say it was serious, although the overuse of exclamation marks |
19 |
could be seen as implying that (I have an automatic exclamation mark |
20 |
filter, so I don't really notice them). |
21 |
|
22 |
> So who is at fault here? The user taking emerges warnings seriously |
23 |
> because they don't want to break their system, or emerge by making |
24 |
> irrelevant warnings appear as being so serious problems that the |
25 |
> unsuspecting user gets so confused and scared of breaking their system |
26 |
> that they start to ask questions on mailing lists? |
27 |
|
28 |
The problem is that portage does not clearly distinguish between |
29 |
information, warnings and error messages. The simplest way of looking at |
30 |
it is "does this stop the emerge proceeding". In your original case, |
31 |
that was not the case. The emerge did stop, but because of the thing |
32 |
with hdf5 and the threads USE flag. Once you had cleared that, the |
33 |
emerge would most likely have proceeded despite the messages. |
34 |
|
35 |
> > quickpkg is really quick. |
36 |
> > Then, to reinstall from that: emerge -vak1 dev-libs/boost |
37 |
> |
38 |
> Oh, it's the whole updating thing. Besides a chance that I'll have to |
39 |
> fix something, it also brings in a new kernel to make and to install. |
40 |
> That takes time. |
41 |
|
42 |
Only if you do it. Unless your existing kernel has stopped working, why |
43 |
the rush to build a new one? |
44 |
|
45 |
> > The more freedom with the package manager, the more conflicts you |
46 |
> > might encounter. |
47 |
> |
48 |
> That doesn't mean that the package manager should be unable to provide |
49 |
> the user with a number of possible solutions and let them pick one. |
50 |
|
51 |
It did, it told you to add one USE flag or remove another. |
52 |
|
53 |
> Particularly, it doesn't mean that the package manager should give the |
54 |
> impression that things might go horribly wrong when some action is |
55 |
> performed unless they actually will. |
56 |
|
57 |
No, it shouldn't. But it is already well established that portage's |
58 |
output can be opaque from a user's perspective. That's a well trodden |
59 |
path that is not worth revisiting unless you can help with a solution. |
60 |
|
61 |
> >> Where and how do the above messages give me choices? They are |
62 |
> >> telling me that boost doesn't work with itself, |
63 |
|
64 |
No they aren't. They are saying that boost will not be upgraded, they are |
65 |
not saying that anything will not work. I've been seeing almost identical |
66 |
messages about ocaml for months now, things still work with the version I |
67 |
had before the messages began. |
68 |
|
69 |
> > There is, several in fact. |
70 |
> > One is called "Backups" |
71 |
> |
72 |
> You seriously expect a backup just to be able to undo an emerge --sync? |
73 |
|
74 |
Absolutely. All sync does is update the contents of a directory, if you |
75 |
backed up that directory you could restore it. |
76 |
|
77 |
> Ok, then make it as easy to boot from ZFS as it is to boot from ext4. |
78 |
> |
79 |
> On a side note, how difficult or easy, and how advisable, is booting |
80 |
> from btrfs, particularly for a xen PV guest which might have the kernel |
81 |
> residing on the host? (I might prefer that over using lvm.) |
82 |
|
83 |
I don't know about Xen, but on real hardware it's as simple as ext4 with |
84 |
a single drive, and transparently handled by dracut if you use RAID. |
85 |
|
86 |
> > The other one is portage snapshots. |
87 |
> |
88 |
> That sounds like something I should learn about. |
89 |
|
90 |
See above re backups, it's just a tarball of the portage tree. |
91 |
|
92 |
|
93 |
-- |
94 |
Neil Bothwick |
95 |
|
96 |
But there, everything has its drawbacks, as the man said when his |
97 |
mother-in-law died, and they came down upon him for the funeral expenses. |
98 |
-- Jerome K. Jerome |