| 1 |
On Fri, Aug 19, 2016 at 12:58 AM, Michał Górny <mgorny@g.o> wrote: |
| 2 |
> On Thu, 18 Aug 2016 15:21:16 +0200 |
| 3 |
> Alexis Ballier <aballier@g.o> wrote: |
| 4 |
> |
| 5 |
>> On Thu, 18 Aug 2016 08:13:14 -0400 |
| 6 |
>> Rich Freeman <rich0@g.o> wrote: |
| 7 |
>> |
| 8 |
>> > If you just check your packages occassionally to make sure they build |
| 9 |
>> > with gold it completely achieves the goal, and it will actually result |
| 10 |
>> > in fewer bugs using the non-gold linker as well. |
| 11 |
>> |
| 12 |
>> That's what a tinderbox is for. The only QA problem I see here is that |
| 13 |
>> QA doesn't automate that kind of checks anymore since Diego left. Maybe |
| 14 |
>> QA should ask Toralf to run a ld.gold tinderbox and avoid asking people |
| 15 |
>> to randomly test random packages ? |
| 16 |
> |
| 17 |
> Yes, tinderboxing makes a lot of sense if the bugs are afterwards |
| 18 |
> ignored by package maintainers. Or in the best case, the maintainer |
| 19 |
> tells reporter (Toralf) to file the bug upstream. |
| 20 |
> |
| 21 |
|
| 22 |
TBH, these are really two different problems. |
| 23 |
|
| 24 |
1. I think raising awareness of underlinking is good. |
| 25 |
|
| 26 |
2. I think encouraging developers to test their own packages with the |
| 27 |
gold linker is good, because it helps accomplish #1, and increases |
| 28 |
their awareness in general. |
| 29 |
|
| 30 |
3. I think that having a tinderbox systematically testing using the |
| 31 |
gold linker is also good. |
| 32 |
|
| 33 |
4. I think that hitting devs with a cluebat when they ignore valid |
| 34 |
bugs is good. |
| 35 |
|
| 36 |
The flip side of this is that we're not necessarily better off if |
| 37 |
maintainers just abandon packages because they have terrible build |
| 38 |
systems. At some point you need to work with them. However, if |
| 39 |
they're not willing to at least stick in a slot operator dependency |
| 40 |
when asked to, then sure we should have a talk with them. (A slot op |
| 41 |
dep will of course help by triggering rebuilds, but it doesn't |
| 42 |
actually directly fix the underlinking issue, which would require |
| 43 |
fixing the build system.) |
| 44 |
|
| 45 |
I think the big thing is acknowledging that packages that are missing |
| 46 |
dependencies or which are underlinked are defective. Sure, it would |
| 47 |
be nice if somebody else came along and helped find our mistakes. |
| 48 |
However, that in itself doesn't excuse us from having made them in the |
| 49 |
first place. And it certainly doesn't excuse giving people a hard |
| 50 |
time when they politely point them out. |
| 51 |
|
| 52 |
-- |
| 53 |
Rich |
Archives