1 |
Hi Grant, |
2 |
|
3 |
On 3/3/06, Grant Goodyear <g2boojum@g.o> wrote: |
4 |
> Yep. Having a USE flag enabled turns out not to be a guarantee. That |
5 |
> said, package builds do become deterministic, so (for example) if one |
6 |
> needs to know if msmtp was built with openssl or gnutls it is easy |
7 |
> enough to pull the logic from the msmtp ebuild. |
8 |
|
9 |
If a build process has errors, it should stop. That's still |
10 |
deterministic. I'd respectfully argue that it's far more |
11 |
deterministic than having each single package implement separate |
12 |
policy for handling conflicting USE flags. |
13 |
|
14 |
Policy belongs with the user, not in Portage. It's exactly the same |
15 |
as the kernel pushing policy into userspace. |
16 |
|
17 |
I believe that it's bad engineering to force (using your example) |
18 |
packages that depend on msmtp to have to copy the logic from the msmtp |
19 |
ebuild to understand how to correctly depend on that package. It |
20 |
violates the principle of Don't Repeat Yourself. |
21 |
|
22 |
If we're going to do this, then at least we should be implementing a |
23 |
consistent standard across all ebuilds. F.ex, when SSL and TLS |
24 |
conflict, we should have a standard saying that all ebuilds will |
25 |
consistenly favour one over the other. That's much more deterministic |
26 |
than having some ebuilds prefer SSL, and others prefer TLS (for |
27 |
example). |
28 |
|
29 |
> I'm sure that there is |
30 |
> a more elegant solution, but I'm not convinced that having the user |
31 |
> randomly throw USE flags at a package until some combination works is |
32 |
> necessarily it. I could be wrong, however. *Shrug* |
33 |
|
34 |
Why would the user be randomly throwing USE flags at a package? With |
35 |
the PHP ebuilds and the confutils eclass, we've already proved that |
36 |
you can tell the user exactly why the ebuild has bailed, and what they |
37 |
can do to fix it. |
38 |
|
39 |
> With an apology for singling you out (when yours is certainly not the |
40 |
> only, or even the most appropriate, example), that sort of emotional |
41 |
> response is how these threads begin to degenerate. There appears to be |
42 |
> an implicit assumption there that your view is clearly correct, and any |
43 |
> other is embarrassingly silly. Instead, I suggest that perhaps people |
44 |
> on both (all?) sides of the issue are rational, intelligent people who |
45 |
> simply differ on what happens to be the greatest evil. |
46 |
|
47 |
I accept the point, but, well, I *am* embarrassed. I guess I'm just |
48 |
willing to admit it when others would rather not be open and honest |
49 |
about how this makes them feel. I think it's fair to raise that as |
50 |
part of the debate. I don't think we talk enough about how we feel |
51 |
about what's happening to Gentoo these days. I think it's reasonable |
52 |
that issues like this are delt with both on an emotional intelligence |
53 |
level as well as an intellectual one. |
54 |
|
55 |
> I would argue that the user tends to get unexpected results in either |
56 |
> case, it's just a matter of whether the build crashes, or the resulting |
57 |
> package is somewhat unexpected. Given that fact, I'm arguing that |
58 |
> having a potentially-lengthy emerge crash out is the bigger evil. |
59 |
|
60 |
I can understand how that matters to first-time installs, or users |
61 |
running old hardware. My environment and concern are servers, where |
62 |
dual-Xeon or dual-Opterons are the norm. The time it takes to install |
63 |
a package is trivial against the time it takes to diagnose why it |
64 |
isn't working the way you would expect. |
65 |
|
66 |
Best regards, |
67 |
Stu |
68 |
|
69 |
-- |
70 |
gentoo-dev@g.o mailing list |