1 |
On Thu, Jan 24, 2013 at 9:22 PM, Michael Orlitzky <michael@××××××××.com> wrote: |
2 |
> Using package.env is preferable, since it basically exists in lieu of |
3 |
> prefixing every environment variable with $PN. But I don't particularly |
4 |
> care about the details. I was just curious if there are real cases where |
5 |
> the config check would do harm. |
6 |
|
7 |
A few have come up. They mostly involve with building on systems |
8 |
where the package won't be deployed, in things like chroots, during |
9 |
things like upgrading stage3s before installing the kernel, and so on. |
10 |
The USB_SUSPEND case came up as well - things like that shouldn't be |
11 |
fatal. |
12 |
|
13 |
Does package.env support having more than one environment file per |
14 |
package? If somebody is already using package.env to tweak CFLAGs and |
15 |
such then also using it for this flag means basically doubling the |
16 |
number of configurations (ie having a debug-flags and |
17 |
debug-flags-ikwiad file, and ditto for any other tweaks). |
18 |
|
19 |
> |
20 |
> If there's no downside (i.e. no one will notice, except the people whose |
21 |
> machines would be broken), then the whole debate is stupid. |
22 |
> |
23 |
|
24 |
I don't think that we'd have a debate if there wasn't a sense that |
25 |
people would notice. The whole "your use case isn't the same as mine |
26 |
thus it doesn't matter" thing gets old. I don't expect devs to do a |
27 |
lot of work to maintain corner cases, but in this case the desire is |
28 |
to not make users patch ebuilds if they don't want to have them die |
29 |
when nothing is wrong, and accomplishing this doesn't really cost |
30 |
anybody time. |
31 |
|
32 |
> |
33 |
> Better to spare the innocents, and for the people who set |
34 |
> I_KNOW_WHAT_I_AM_DOING=y in make.conf, we can create |
35 |
> RESOLVED:I_THOUGHT_YOU_KNEW. |
36 |
> |
37 |
|
38 |
Remember, it isn't the skipped check that broke the system - it was |
39 |
the package change that required the check in the first place. If the |
40 |
package just silently fails (even due to an override) then that is a |
41 |
bad design. It should at least give as many warnings as it does |
42 |
presently (but non-fatally), and it should still be preceded with a |
43 |
news item. |
44 |
|
45 |
Otherwise we're just finding creative ways to drive away users. Sure, |
46 |
we can call them stupid on their way out the door, but while I can't |
47 |
speak for anybody else, I'm mainly here because I'd like to do some |
48 |
good, and I wouldn't mind it if I found Gentoo useful along the way. |
49 |
|
50 |
I think what we need is a combination of: |
51 |
1. Only truly critical issues trigger fatal errors. This can't be |
52 |
used for default config checks. |
53 |
2. The fatal errors are overridable. |
54 |
3. If overriden the errors are still output. |
55 |
4. The error output includes instructions on how to override, and the |
56 |
instructions are package-specific (ie don't stick i_know_what_im_doing |
57 |
in make.conf). |
58 |
5. The change introducing the issue is preceded with a news item. |
59 |
|
60 |
Honestly I think #5 is more important than having the check in the |
61 |
first place. Also, most of these can be handled at the eclass level - |
62 |
the package should explain the issue, but all the |
63 |
instructions/behavior can be generic (that is, generic instructions on |
64 |
how to do the override in a package-specific way, such as through |
65 |
package.env or whatever). |
66 |
|
67 |
Rich |