1 |
On Thu, Jan 24, 2019 at 2:33 PM Jack <ostroffjh@×××××××××××××××××.net> |
2 |
wrote: |
3 |
|
4 |
> On 2019.01.24 15:17, Michael Jones wrote: |
5 |
> [snip....] |
6 |
> > A lot of people run emerge without checking the preliminary console |
7 |
> > output. |
8 |
> But most (I hope) of them understand the possible problems, and are |
9 |
> willing to "pick up the pieces" if something goes wrong. Most of them |
10 |
> do not blame the Gentoo developers in such a case. |
11 |
> |
12 |
|
13 |
Of course. |
14 |
|
15 |
|
16 |
> > I don't think it's fair to imply that doing that is somehow wrong. |
17 |
> Some/many of us might disagree. If you don't check the output to see |
18 |
> exactly what emerge is going to do, it may well not do what you |
19 |
> want/expect. Is it "wrong" to drive your car with a blindfold on. |
20 |
> |
21 |
|
22 |
You're absolutely right that if you don't check the output, it might be |
23 |
doing something different than what you expect. |
24 |
|
25 |
I don't really see how the car analogy is applicable though, so I can't |
26 |
agree with that. |
27 |
|
28 |
|
29 |
> > The vast majority of Linux distributions support unattended upgrades |
30 |
> > without a second thought. |
31 |
> The vast majority of Linux distributions are binary distributions, |
32 |
> where a package either works or not, and if it doesn't, it is very |
33 |
> likely a packaging error. The vast majority of Linux distributions do |
34 |
> not install from source, with lots of ways of configuring things to the |
35 |
> user's liking, many of which possible configurations will not work. I |
36 |
> would say that binary distributions are much more deterministic in |
37 |
> their behavior. You really can't run Gentoo safely for very long |
38 |
> without paying close attention to what you are doing - with both |
39 |
> installs and upgrades. |
40 |
> |
41 |
> |
42 |
Well,lets break this down into smaller pieces: |
43 |
|
44 |
> The vast majority of Linux distributions are binary distributions, |
45 |
where a package either works or not, and if it doesn't, it is very |
46 |
likely a packaging error. |
47 |
|
48 |
Sure, a lot of the time problems with binary distros are packaging errors. |
49 |
Bugs can happen in a lot of ways. Sometimes the package is just broken |
50 |
*shrug*. |
51 |
|
52 |
> The vast majority of Linux distributions do not install from source, with |
53 |
lots of ways of configuring things to the user's liking, many of which |
54 |
possible configurations will not work. |
55 |
|
56 |
Sure, that's true. |
57 |
|
58 |
While I agree that the number of user facing toggles increase the |
59 |
likelihood of something going wrong, I don't agree that this is inherently |
60 |
an excuse. Obviously it's an explanation as to how something going wrong |
61 |
wasn't able to be addressed ahead of time, but it's not automatically a |
62 |
fair excuse / justification for delivering an ebuild to the portage tree |
63 |
that has problems. |
64 |
|
65 |
I'm not arguing that the Gentoo developers are somehow beholden to users, |
66 |
nor am I saying that it's wise to assume everything will work correctly. |
67 |
I'm simply saying that it's not fair to say that not checking the output of |
68 |
emerge is inherently wrong. If it was inherently wrong, then emerge should |
69 |
not allow for a package to be emerged until the user has reviewed it's |
70 |
stated plan. |
71 |
|
72 |
> You really can't run Gentoo safely for very long without paying close |
73 |
attention to what you are doing - with both installs and upgrades. |
74 |
|
75 |
I don't know that this is true. I've had the same set of use flags |
76 |
configured on all of my Gentoo computers for 5-ish years now, and I've only |
77 |
very rarely messed with them. For the most part, running "eix-sync && |
78 |
emerge --update --newuse --deep @world" once a week has allowed me to have |
79 |
unattended upgrades for many months at a time, only needing to adjust one |
80 |
or two things every several months. Needing to tweak something is my |
81 |
exception, not my rule. |
82 |
|
83 |
My reason for replying initially was that I didn't think it was fair to |
84 |
make light of users who don't expect to *need* to scrutinize the output of |
85 |
emerge every single time they run it. Those people exist (hi, nice to meet |
86 |
you), and it's not fair to say they're wrong or somehow making a grave |
87 |
error in judgement. |
88 |
|
89 |
It's entirely fair to say that they are treading on thin ice, and that if |
90 |
they choose to do this they should understand the risks, but it's not fair |
91 |
to say they're automatically wrong to use the tool in a way that the tool |
92 |
allows itself to be used. |
93 |
|
94 |
Either way, we don't need to turn this into a long and in depth discussion, |
95 |
so I probably won't reply to the list again unless you have any specific |
96 |
questions or concerns for me. |
97 |
|
98 |
Happy compiling :-) |