1 |
On Wednesday 17 August 2005 15:14, Grobian wrote: |
2 |
> The whole point was that I like avoiding storing data double |
3 |
> (redundant), if that can be done easily. |
4 |
It has its own pro and cons as quite everything. You can avoid storing data |
5 |
double but makes more difficult to access it by the priority end users (devs; |
6 |
while users use changelogs, too, devs are the ones who really need to access |
7 |
them). |
8 |
|
9 |
> As for the forum example: it wouldn't be a foreign key if there wasn't a |
10 |
> left outer join to look up the respective value for the column. And so |
11 |
> that left outer join is here to generate the Changelog to be "backwards |
12 |
> compatible" |
13 |
The problem is where this is going to be generated. If this is going to be |
14 |
generated in the staging box before going in rsync, every dev that doesn't |
15 |
use rsync should use cvs log command to get the data, and this is quite |
16 |
unpractical (and requires bandwidth and adds load to the cvs server). |
17 |
|
18 |
So we should balance redundancy and load... I still think that a little |
19 |
redundancy of the changelogs saves us from having to add load to the staging |
20 |
or cvs box... |
21 |
|
22 |
-- |
23 |
Diego "Flameeyes" Pettenò |
24 |
Gentoo Developer - http://dev.gentoo.org/~flameeyes/ |
25 |
(Gentoo/FreeBSD, Video, Gentoo/AMD64, Sound, PAM) |