1 |
On Tue, 8 Mar 2016 22:40:34 +0100 |
2 |
Ulrich Mueller <ulm@g.o> wrote: |
3 |
|
4 |
> >>>>> On Tue, 8 Mar 2016, Brian Dolbec wrote: |
5 |
> |
6 |
> > As discussed a few years ago. We agreed to trim the package moves |
7 |
> > updates to maintain 4 years history active in the tree. It currently |
8 |
> > still goes back to include 2010. There was a recent edit to a 2010 |
9 |
> > file that triggered a full re-processing of those files. |
10 |
> |
11 |
> That's not correct. I had noticed reprocessing of all files already |
12 |
> around March 1st, definitely before the edit of the 1Q-2010 file. |
13 |
|
14 |
There has been at least one time recently that a change in the rsync |
15 |
tree generation scripts caused an massive download of files and |
16 |
re-processing of the updates. I believe that was the incident you were |
17 |
referring to. I saw chat (#gentoo-dev) a few days ago from dwfreed. He |
18 |
was testing out more changes to those scripts. So, it is possible that |
19 |
a script update once again caused the re-processing. I'm inquiring if |
20 |
that was indeed the reason. |
21 |
|
22 |
> |
23 |
> > Last night for me it was triggered again. I don't know the cause, |
24 |
> > but I really don't care enough to investigate. But it prompted bug |
25 |
> > 576752. |
26 |
> |
27 |
> > That said, being 2016, I am going to trim 2010 and 2011 files from |
28 |
> > the active tree in 48 hours unless anyone shows a valid and |
29 |
> > justifiable reason to keep them. |
30 |
> |
31 |
> That would (maybe) cure the symptoms, but don't you think that the |
32 |
> real bug should be found? I suggest that you wait with removal of the |
33 |
> files until then. |
34 |
> |
35 |
> Ulrich |
36 |
|
37 |
No, it won't cure the symptoms, just, shorten the number processed. |
38 |
Also it will eliminate a possible need to edit those old files like had |
39 |
been done recently. |
40 |
|
41 |
-- |
42 |
Brian Dolbec <dolsen> |