These are chat archives for arenanet/api-cdi

12th
Apr 2016
David Reeß
@queicherius
Apr 12 2016 17:09
@Eearslya Just added an email, still can't edit. Doesn't matter too much tho, I'll check next time I find something I wanna change. :D
darthmaim
@darthmaim
Apr 12 2016 17:11

Only registered accounts can become autoconfirmed. There are two requirements:

  • The account must be 24 hours old.
  • The user must save 10 edits to any page.

To satisfy the second requirement, many users simply edit their userpage multiple times.

David Reeß
@queicherius
Apr 12 2016 17:18
@darthmaim good to know, thanks! I think I have that many edits, but not logged in. :D
okay team I have an issue that I'm not sure how to fix
/v2/wvw/matches returns kills/deaths right
the match sums include obsidian sanctum kills
(which makes the sums not match up with the maps)
not sure if I should exclude OS kills or expose them explicitly
Nabrok
@Nabrok
Apr 12 2016 18:25
I'd say make them explicit
especially if they count for points per kill
I'm not actually sure if they do; asking around and checking the code
if they don't then I'm going to cull them
Nabrok
@Nabrok
Apr 12 2016 18:26
How many kills do we usually get in OS during a week?
live stats are saying like 2k/team for current data
wait no those are shard ids sec
match 2-5 has about 20 deaths/team
Nabrok
@Nabrok
Apr 12 2016 18:27
Does that include fall deaths?
yes
there's about 10 kills/team
Nabrok
@Nabrok
Apr 12 2016 18:28
yeah, not many people there these days
Happened to do it on Sunday for the guild mission, didn't see anybody.
OS is better for GvGs than guild halls
and I think there's are actually still some GvG guilds
Nabrok
@Nabrok
Apr 12 2016 18:31
Probably a small intersection on the venn diagram of "people who like jumping puzzles" and "people who like pvp"
I'm just going to push up the naive fix for now which is to filter out OS kills/deaths
because like ... I'm not sure if those count for PPK?
I'll find some way to expose those explicitly -- it would make the most sense if OS was an object on the match.maps array but that feels kind of dangerous
want to talk with some application developers before I think about that
doing a live deploy, things might break
changelog:
  • Fix an error in /v1/maps.
  • Clean up the error message in /v2/commerce/exchange when the value provided is too low.
  • Fix kill/death sums to not include OS in /v2/wvw/matches.
pretty sure I didn't merge in anything that has dependencies on stuff that's not released (aka all the backend changes that are waiting on the next patch) but i'm not really a good programmer so hold onto your butts
windwarrior
@windwarrior
Apr 12 2016 20:10
famous.last.words
it'll probably be fine
darthmaim
@darthmaim
Apr 12 2016 20:11
define fine
windwarrior
@windwarrior
Apr 12 2016 20:11
thats saying that there is a 0-100% chance of breakage
darthmaim
@darthmaim
Apr 12 2016 20:11
:fire:
there is a 1-99% chance of breakage
(nothing is ever certain)
darthmaim
@darthmaim
Apr 12 2016 20:12
are you certain that nothing is ever certain?
:O
nothing excluding this statement is ever certain
windwarrior
@windwarrior
Apr 12 2016 20:12
parse error
which endpoint? my test suite is still running
windwarrior
@windwarrior
Apr 12 2016 20:13
no that stattement
sorry
oh okay
lye @lye purchases new pants
windwarrior
@windwarrior
Apr 12 2016 20:14
all my apps are still working
:)
the test suite didn't kick back anything, so looks like it's good
windwarrior
@windwarrior
Apr 12 2016 20:17
the paper of the algorithm I am implementing: the result is in range -Inf..0, my code says.. 4 :'(
darthmaim
@darthmaim
Apr 12 2016 20:17
just 4 off
could be worse
yeah just add result -= 4 and everything should be fine
windwarrior
@windwarrior
Apr 12 2016 20:18
well might as well substract Infinity then
i'm not familiar enough with ieee754 to know if that produces what you'd expect it to produce
windwarrior
@windwarrior
Apr 12 2016 20:21
I will just assume the published paper is wrong
I cannot be right
darthmaim
@darthmaim
Apr 12 2016 20:22
if the tests are failing, adjust the tests
smiley
@codemasher
Apr 12 2016 20:59
this chat room is pure comedy gold :D
windwarrior
@windwarrior
Apr 12 2016 22:31
turns out, I was an idiot after all
:D
note to self: natural language processing is stuff for really really really clever people
hmm, I can sort github issues by :thumbsup:
that's pretty helpful actually
Pat Cavit
@tivac
Apr 12 2016 23:58
yeah that's a super-cool feature