1 |
On 04/08/10 15:29, Ciaran McCreesh wrote: |
2 |
> On Thu, 8 Apr 2010 05:02:25 -0700 |
3 |
> Brian Harring <ferringb@×××××.com> wrote: |
4 |
>> 4) if there are questions re: use cycle breaking or other bits, feel |
5 |
>> free to ask prior please- council meeting times unfortunately right |
6 |
>> now intersect badly with my paying work so it's hard to be online to |
7 |
>> answer questions during the meeting (that said per the norm I'll try). |
8 |
> |
9 |
> Please detail your cycle breaking algorithm that works in such a way |
10 |
> that it's guaranteed not to toggle flags that will break a system, but |
11 |
> that does not require any changes to be made to ebuilds etc telling the |
12 |
> package manager what those flags are. |
13 |
> |
14 |
That would violate the Entscheidungsproblem. |
15 |
|
16 |
But why would you make the cycle breaking depend on an oracle? Once we |
17 |
have the method in place we can find the two special cases and think of |
18 |
ways to fix them. Abandoning the idea as a whole just because there may |
19 |
be a corner case that isn't as easy appears quite silly to me. |
20 |
|
21 |
Brian's proposal is the only one I've seen that is deterministic and |
22 |
sane, so I think we should figure out if we can improve it instead of |
23 |
giving up at the first bump in the road. |