1 |
On Fri, Jan 18, 2013 at 2:13 PM, Benedikt Böhm <hollow@g.o> wrote: |
2 |
|
3 |
> On Fri, Jan 18, 2013 at 9:41 AM, Michael Weber <xmw@g.o> wrote: |
4 |
> |
5 |
>> On 01/18/2013 09:28 AM, Benedikt Böhm wrote: |
6 |
>> > but - and that's quite important i guess - i only use my own clone of |
7 |
>> > the portage tree which i sync from time to time and i also keep |
8 |
>> > different versions stable, etc. |
9 |
>> HEAVY USER! But you have full control. |
10 |
>> Do you have any sophisticated mechanism to detect tree breakage (i.e. us |
11 |
>> f*** up), like Samuli replying -commit to -dev or irc activity? |
12 |
>> |
13 |
>> Or do you simply delay commit? (re-schedule on weekends/nights) |
14 |
>> |
15 |
> |
16 |
> i manually sync with the gentoo-x86 repo from time to time (except |
17 |
> security issues, which i sync as soon as they are fixed) |
18 |
> |
19 |
> i have no automated way to detect any tree breakage, but when i sync the |
20 |
> complete tree i do extensive manual testing on a handfull of machines |
21 |
> (either staging machines, or ones that are not really important) |
22 |
> |
23 |
> but the main reason i even have a clone is to sync updates in batches. |
24 |
> it's really a hassle to keep servers in sync when changes to gentoo-x86 |
25 |
> happen any other minute. i also need to adapt my chef cookbooks for some |
26 |
> updates, so i do it all in one batch (sync, test, adapt, test, deploy) |
27 |
> |
28 |
|
29 |
i forgot to add one main difference here: i only have ~1000 packages in my |
30 |
repo, which makes it a lot faster for metadata, rsync, eix and all that ... |
31 |
the repo is only used for server deployments. no desktop, no games, only |
32 |
very few X11 ebuilds for headless testing etc. |