1 |
On 10/26/2015 01:52 PM, Rich Freeman wrote: |
2 |
> On Mon, Oct 26, 2015 at 3:47 AM, Kristian Fiskerstrand <k_f@g.o> wrote: |
3 |
>> I haven't read through the references yet, but is it also stated that |
4 |
>> eapply_user needs to be applied at least once in addition to being |
5 |
>> idempotent? |
6 |
>> |
7 |
> |
8 |
> It says it must be applied once. The reference in this case is just a |
9 |
> diff to one line, so you should probably just read it. :) |
10 |
> |
11 |
> If anybody has additional pros/cons to the idempotent proposal that |
12 |
> haven't already been raised I'm all ears. I was against the change |
13 |
> but I'm willing to go along with it based on the arguments so far. |
14 |
> |
15 |
|
16 |
It's still not clear to me how this is supposed to work at all. Are we |
17 |
expected to run eautoreconf unconditionally just in case a user will |
18 |
throw in a patch that alters the build system? I definitely won't do |
19 |
that, because autoreconf is the source for a lot of problems and build |
20 |
failures. |