1 |
On Wed, Nov 06, 2013 at 08:58:36PM -0800, "Paweł Hajdan, Jr." wrote: |
2 |
> On 11/6/13 10:43 AM, Andreas K. Huettel wrote: |
3 |
> >> On Wed, Nov 06, 2013 at 09:18:43AM +0100, Andreas K. Huettel wrote: |
4 |
> >>> 8. Revival of archives.gentoo.org |
5 |
> >> |
6 |
> >> There isn't any reference for this one, who's spear-heading it; as infra |
7 |
> >> I'd like to know. We do have the raw emails for the archive, what's |
8 |
> >> broken is strictly the web interface. |
9 |
> > |
10 |
> > I am, and the only available reference is the announcement itself, see the |
11 |
> > paragraph at the top about missing references :). |
12 |
> > |
13 |
> > My only intention for this topic so far is to ask my council colleagues for |
14 |
> > their opinion on a general statement like "archives.gentoo.org functionality |
15 |
> > was useful, and it would be nice to have our own online, definitive archive of |
16 |
> > the more important mailing lists back working again at some point in the |
17 |
> > future". |
18 |
> Curious: does anyone have any doubts about usefulness of archives.g.o? |
19 |
> |
20 |
> Then, do you know what actually broke? It's really surprising to me |
21 |
> since it seems that hardly anything changes there. |
22 |
We used a custom template in mhonarc to generate a variant of |
23 |
Guide/ProjectXML from the emails. Along with that template, there was |
24 |
some custom code to ensure we generated consistent IDs even if the prior |
25 |
mhonarc listing was damaged; so that a link to an email once it was |
26 |
posted would always be consistent. |
27 |
|
28 |
The parts above we still have... what we don't have, is some fixes that |
29 |
allowed mhonarc to scale to the crazy number of emails we were putting |
30 |
into the archive. It's default was generating an entire new index and |
31 |
each message again, every time it was run. We hacked incremental support |
32 |
onto there, that was lost. |
33 |
|
34 |
Along with the rest of Gentoo that is moving away from Guide/ProjectXML; |
35 |
archives need to move to the future, but it needs to be something |
36 |
scalable. The archive of raw email itself exceeds 20GiB in size; more |
37 |
than 75% of which should probably be public (the remaining fraction is |
38 |
stuff like core/trustees/council etc). |
39 |
|
40 |
-- |
41 |
Robin Hugh Johnson |
42 |
Gentoo Linux: Developer, Trustee & Infrastructure Lead |
43 |
E-Mail : robbat2@g.o |
44 |
GnuPG FP : 11ACBA4F 4778E3F6 E4EDF38E B27B944E 34884E85 |