1 |
Stuart Herbert wrote: |
2 |
> I agree. Adopting a policy like this is a low quality solution for |
3 |
> servers. I've no opinion on how this affects desktops, but packages |
4 |
> for servers need to be precise. A policy that says "if two USE |
5 |
> flags deliver the same benefits, but conflict, pick one" is fine. But |
6 |
> saying "flip a coin" ... how on earth is that "quality"? |
7 |
|
8 |
See my previous post. |
9 |
|
10 |
> And how the heck is it going to work w/ USE-based defaults? This |
11 |
> creates a situation where package (b) cannot trust that a feature is |
12 |
> enabled in package (a), even if package (a) was built with the |
13 |
> required USE flags. |
14 |
|
15 |
Yep. Having a USE flag enabled turns out not to be a guarantee. That |
16 |
said, package builds do become deterministic, so (for example) if one |
17 |
needs to know if msmtp was built with openssl or gnutls it is easy |
18 |
enough to pull the logic from the msmtp ebuild. I'm sure that there is |
19 |
a more elegant solution, but I'm not convinced that having the user |
20 |
randomly throw USE flags at a package until some combination works is |
21 |
necessarily it. I could be wrong, however. *Shrug* |
22 |
|
23 |
> I'll go as far as saying that right now I'm embarrased that it looks |
24 |
> like this is going to become a Gentoo policy in its current form. |
25 |
|
26 |
With an apology for singling you out (when yours is certainly not the |
27 |
only, or even the most appropriate, example), that sort of emotional |
28 |
response is how these threads begin to degenerate. There appears to be |
29 |
an implicit assumption there that your view is clearly correct, and any |
30 |
other is embarrassingly silly. Instead, I suggest that perhaps people |
31 |
on both (all?) sides of the issue are rational, intelligent people who |
32 |
simply differ on what happens to be the greatest evil. |
33 |
|
34 |
> You're absolutely *not* creating a better user experience. You're |
35 |
> brushing a problem under the carpet ... and making it the users' |
36 |
> problem when they wonder why the built a package with a USE flag and |
37 |
> the package still doesn't work as they expect. |
38 |
|
39 |
I would argue that the user tends to get unexpected results in either |
40 |
case, it's just a matter of whether the build crashes, or the resulting |
41 |
package is somewhat unexpected. Given that fact, I'm arguing that |
42 |
having a potentially-lengthy emerge crash out is the bigger evil. |
43 |
|
44 |
> Until Portage supports resolving conflicting USE flags when the |
45 |
> deptree is built, the practical thing to do is for ebuilds w/ |
46 |
> conflicting USE flags to bail. |
47 |
|
48 |
I, quite respectfully, disagree. |
49 |
|
50 |
-g2boojum- |