On Wednesday, October 13, 2004, 6:34:07 PM, Jeff Chan wrote:
On Wednesday, October 13, 2004, 5:43:45 PM, Raymond Dijkxhoorn wrote:
It seems a timing issue. On the rsync server i noticed the multi list is a little outdated. Most likely the generation scripts are stalled? Jeff, can you look into this?
There are a couple issues. One is that I'm slightly updating the script that creates multi to better detect when the individual lists update. (That's where that shell script question came from.) The other is that I added a lock file to prevent that script from running when the sc data was being generated. I added that due to a condition where the sc data was empty when multi needed it. I'm adjusting these things now and they seem to be working better, but of course I'll keep checking them.
All this will be corrected when I rewrite things in a cleaner, simpler and more integrated way, which is on my list of things to do. Second versions are usually better than the first, due to seeing how things could be done better.
OK, I've watched the slightly revised code run through a few cycles, and it seems to be working correctly.
The flaw before was that if a domain appeared on a new list (say sc) but was already on a list (say ws and ob), then multi would not get updated immediately to reflect that it was in the new list (e.g., sc) also. The reason why is that the change detection was only looking at domains and not the lists they came from. Now things to note when any list gets changed, not just changes to the domains. Now updates to multi are happening appropriately more often.
All this will be replaced with something better in the next version, but until then updates to multi should happen more quickly and as originally intended.
Jeff C. -- "If it appears in hams, then don't list it."