| Pages: [1] :: one page |
| Author |
Thread Statistics | Show CCP posts - 0 post(s) |
Captain Thunk
Explode. Now. Please.
   |
Posted - 2008.10.28 14:38:00 -
[1]
After reading about the enormous fight in E-O over the last two days, it's occured to me that there is a possible problem with CCPs remarkable API system.
For the big alliances it is now conceivable that the 100 killmail limit given by the API XML return may not actually cover big fights of the future. Because of the way cronjobs work, it's exceptionally difficult to have the scripts run ever 60 minutes due to the fact there is a caching of 1hr - the script misses by a few seconds. With fights now on an unprecedented scale, thanks largely to CCPs successful efforts in reducing lag, it's quite conceivable that a member corp of an alliance can generate well over 100 killmails within an hour.
I'm hoping to cover this CCP could look at the possibility of extending the 100 mail limit and perhaps changing the caching to something like 55 minutes to allow parsing to take place on the hour with success.
Captain Thunk
|
Jethro Jechonias
Ki Tech Industries
|
Posted - 2008.10.28 14:50:00 -
[2]
What might be more practical is to implement a system similar to the wallet paging.
Return 100 kills per page and allow you to page back for the kills made within the last 24 hours.
Longer would of course be nice, if the server is up to the task.
|
Captain Thunk
Explode. Now. Please.
|
Posted - 2008.10.28 20:11:00 -
[3]
I believe that mechanism already exists mate. If the 100 kills requires more than one page then you need to do the beforeID thing to get the second or more pages. Extending it to 24hrs regardless of number of mails would be very good though. |
Captain Thunk
Captain Morgan Society
|
Posted - 2008.11.10 18:39:00 -
[4]
Noticed that since 9th November the KillLog XML has been returning 200 kills for a corp up from the previous 100.
Thanks very much CCP 
Captain Thunk
|
sumtool
Amarr
Sniggerdly
Pandemic Legion
|
Posted - 2008.11.12 04:23:00 -
[5]
You can already do the 'beforeKillID' thing to get more results. In your cron job, you want to do something like this:
1: get last 100 kills and process
2: check your 'last kill processed' ID and if the first kill from list #1 is higher than that, you need to
set the 'beforeKillID' and go back to step 1 until you're in sync
3: store the highest kill ID you've processed so far.
Done :D
|
Captain Thunk
Captain Morgan Society
|
Posted - 2008.11.12 08:04:00 -
[6]
Edited by: Captain Thunk on 12/11/2008 08:05:12
 Originally by: sumtool
You can already do the 'beforeKillID' thing to get more results. In your cron job, you want to do something like this:
1: get last 100 kills and process
2: check your 'last kill processed' ID and if the first kill from list #1 is higher than that, you need to
set the 'beforeKillID' and go back to step 1 until you're in sync
3: store the highest kill ID you've processed so far.
Done :D
The script already does that 
It fetches as many pages of XML as CCP will allow it to, retrieving the beforeKillID until it returns that it's exhausted. CCP hard cap it at 200, previously 100, mails
(If ever CCP set it to unlimited walkback then all EDK killboards using it are going to retrieve every mail available whenever the script fires)
Captain Thunk
|
| |
|
| Pages: [1] :: one page |
| First page | Previous page | Next page | Last page |