1 |
On Sun, Oct 30, 2005 at 02:47:51PM +0900, Jason Stubbs wrote: |
2 |
> > post sync has a limited shelf life; it's usable now for the following |
3 |
> > 1) rolling gensync functionality in |
4 |
> > 2) rolling glsa notices on sync in |
5 |
> > 3) rolling package watch lists of updates in |
6 |
> > 4) a method to trigger transfer of an rsync temp dir on a rsync proxy |
7 |
> > for a lan to the actual directory rsyncd distributes from. |
8 |
> |
9 |
> 1, 2 and 3 are slated for inclusion. What will happen with 4 and other similar |
10 |
> site-specific hooks? |
11 |
|
12 |
My intention for post sync is to ignore it for 3.0 code. |
13 |
|
14 |
Reasoning pretty much comes down to the fact it's a general hook; we |
15 |
have no way of knowing if they're triggering cache generation, |
16 |
arbitrary repo syncing, package scans, etc. |
17 |
|
18 |
That's not to say that hooking code in won't be possible, just that |
19 |
post sync in it's current form I view as strictly a 2.x thing; so... |
20 |
for #4, how I'd probably implement it in 3.0 would be a treeset, |
21 |
|
22 |
rsync repo, |
23 |
rsyncd share point repo |
24 |
|
25 |
with the sync calls chained; basically a slaved repo. |
26 |
|
27 |
Intend to do similar tricks with slaving the cache for ospreyon a |
28 |
sidenote. |
29 |
|
30 |
Either way, configuration will allow something to be done, although |
31 |
again, post sync won't be the route, will have to specify the wrappers |
32 |
per repo (and potentially do some trickery for #4 in your config). |
33 |
~harring |