| 08:32.38 | *** join/#riftuidev Imhothar (~Imhothar@ppp-93-104-170-200.dynamic.mnet-online.de) |
| 10:10.08 | *** join/#riftuidev Imhothar1 (~Adium@ppp-93-104-170-200.dynamic.mnet-online.de) |
| 13:40.55 | ZorbaTHut | hrm, it should've made it, but something might have screwed up :/ will have to check it on monday |
| 13:41.15 | ZorbaTHut | ah, if you got it to work, then |
| 13:49.40 | Baanano | yep, it didn't work on the pts but works on live |
| 13:50.42 | Baanano | something i've noticed is that using it doesn't decrease the auctionanalyze queue count |
| 13:51.53 | Baanano | or maybe it is being replenished too fast to notice, but it always reports 3650 |
| 13:53.03 | *** join/#riftuidev borbra (~quassel@ti0033a400-6088.bb.online.no) |
| 13:54.01 | Baanano | can you confirm if that's the expected behavior, Zorba? |
| 13:58.23 | Baanano | I'm worried about it being too good to be true: the different itemtype count on the AH is ~6000 at any given time, and I'd want to have at least a week of data from them -> if the auctionanalyze gets replenished that fast it shouldn't be a problem, but if it gets drastically reduced in the future I'd need a different design approach |
| 14:00.05 | ZorbaTHut | it should give you one year of data accessing per second. this will be nice and speedy if you're querying a week at a time, and a whole lot slower if you're trying to pick up tons of historical data |
| 14:00.32 | ZorbaTHut | I can't promise it will stay that way; this was an estimate I came to with the person who did the development of the feature, and the conclusion was "let's do this and see what happens" |
| 14:00.51 | ZorbaTHut | (note that every query is considered to be at least ten days to account for request overhead) |
| 14:01.38 | ZorbaTHut | however, the general pattern of "it is throttled roughly based on the amount of data you request" is going to stay the same, so . . . encouraged to request only what you need :V |
| 14:05.24 | Baanano | so about 2 minutes to collect 1 week of data for each listed item |
| 14:05.31 | Baanano | sounds reasonable |
| 14:05.44 | ZorbaTHut | yep, that sounds pretty sensible |
| 14:08.35 | Baanano | about the ten days min: for users playing daily, only the last day of data would be required, not nice having to pay 10x days for what should be the average case :P |
| 14:08.55 | ZorbaTHut | true, but it's not for you, it's for our servers. queries have overhead. :) |
| 14:09.16 | ZorbaTHut | if we removed the cap, we'd have to reduce the total allowed queries by a lot, just to avoid people slamming us with 365 one-day queries per second |
| 14:09.58 | ZorbaTHut | I also considered adding ten "days" to each query to account for overhead but decided this method would be a bit simpler overall - it's the same technique we use in the storage system, so there's precedent |
| 14:11.09 | *** join/#riftuidev Fiia (579e85a6@gateway/web/freenode/ip.87.158.133.166) |
| 14:25.58 | *** join/#riftuidev ZorbaTHut (tocohl@maximilian.pavlovian.net) |
| 14:25.59 | *** mode/#riftuidev [+o ZorbaTHut] by ChanServ |
| 14:48.39 | Baanano | oh, that wasn't a serious complaint, the current system seems reasonable :) |
| 15:01.59 | ZorbaTHut | excellent ^^ |
| 17:03.21 | *** join/#riftuidev Noshei (~quassel@c-24-9-171-99.hsd1.co.comcast.net) |
| 18:09.57 | *** join/#riftuidev borbra (~quassel@grm-128-39-146-62.studby.uia.no) |
| 23:45.26 | seebs | "Average rift zone's population mostly trying to climb steep cliffs after falling off" factoid actually statistical error. Scarlet Georg, which has a really large cliff and a winding path up it, is an outlier adn should not have been counted. |