00:33:46 Unstable branch on cbro.berotato.org updated to: 0.29-a0-90-gc37658a0fe (34) 00:48:16 <03wheals> Yermak: yes it is, same as not getting dragon armours 00:54:23 Monster database of master branch on crawl.develz.org updated to: 0.29-a0-90-gc37658a0fe 01:34:47 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-4567-gb362bb5a12 03:12:33 =nemelex HODe 03:12:34 Defined keyword: nchoice => DDAM|DjCA|BaAK|TeEn|FoIE|PaAr|GhDe|HODe 04:24:14 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-4567-gb362bb5a12 06:03:39 <09gammafunk> > Invalid brand on item scarf of see invisible, annulling. from objstat, wonder what that is 06:46:55 <06advil> is sinv supposed to be ok on a scarf? 06:47:09 <06advil> is_armour_brand_ok says no it looks like 07:04:57 <09gammafunk> no, sinv definitely is not a type of scarf, 07:05:10 <09gammafunk> I'm just not sure why this instance is happening and triggering a debug/error message 07:07:04 <09gammafunk> happening early in items() so it's probably going to be hard to track down without a seed 07:07:18 <09gammafunk> was worried it was a vault placement of some kind, but I can't find any instance 07:17:52 <09gammafunk> another instance: > Invalid brand on item orb of see invisible, annulling. 07:18:18 <09gammafunk> wonder if it would be good to print the seed for errors if we're in objstat/mapstat mode 07:31:29 -!- allbery_b is now known as geekosaur 07:50:19 <06advil> or just run on seed 1 like tests do (or at least, they do iirc) 07:57:06 That might make the results for more than one iterator less than exciting 07:57:06 *iteration 09:59:43 <06advil> @ebering summary of the problem as I understand it: total score on individual player pages is showing a different value than total score on the overall player table. For high scores, they're off by as much as 10k 1. the former is calculated in python code in the player page template from the individual scores 2. the latter is pulled from the db, where the total score is updated via a SQL query that's constructed from individual scores 09:59:44 3. the code looks similar for these, but it is in a completely different language so it's a bit hard to tell without more understanding of what's intended than I have 09:59:59 <06advil> I'll dig up the relevant links, one sec 10:02:53 <06advil> * relevant commits: https://github.com/crawl/dcss_tourney/commit/50e592a1fd5144cd5a338fa9bfe5b1de06cb9e7b and https://github.com/crawl/dcss_tourney/commit/9d83257750276d0dedfa5485f5c566e91c94b2bf * python template code: https://github.com/crawl/dcss_tourney/blob/0.28-tourney/templates/player-individual-categories.mako#L39 (sum) and https://github.com/crawl/dcss_tourney/blob/0.28-tourney/templates/player.mako#L28 (individual category 10:02:54 calc) * SQL query construction: three functions starting from https://github.com/crawl/dcss_tourney/blob/0.28-tourney/query.py#L1692 10:04:45 <06advil> @gammafunk may have looked at this a bit further 10:04:45 <12ebering> ok, the overall score column in the DB is what's incorrect 10:04:45 <12ebering> the individual category page is correct 10:05:14 <12ebering> ah 10:05:20 <12ebering> score_term doesn't do the directionality 10:05:30 <12ebering> cool I got a few mins with my computer will push a fix 10:05:32 <09gammafunk> I was wondering if it were something like that 10:05:35 <09gammafunk> nice thx 10:05:44 <06advil> ah great thanks 10:05:47 <09gammafunk> that update with no where is spooky 10:06:10 <09gammafunk> just reading how all that sql is assembled as a giant string concat is a bit spooky 10:06:41 <09gammafunk> yeah I should have realized, cuz I can recall the directionality being present in the template calculation 10:07:13 Unstable branch on crawl.akrasiac.org updated to: 0.29-a0-90-gc37658a (34) 10:11:36 <12ebering> pushed a fixing commit but haven't run local test yet 10:11:36 <12ebering> will start up the docker image but it might not finish before I must go again 10:14:47 <09gammafunk> from my understanding, this recalculates score whenever there is a change of points? or is it unconditionally recalculated with every overview update (e.g. if I restart taildb, it will be recalculated) 10:15:25 <12ebering> it is unconditionally recalculated ever overview update 10:15:29 <09gammafunk> cool 10:15:34 <09gammafunk> so should be fine to test it with live db 10:15:52 <12ebering> assuming I didn't just push a syntax error 😉 10:16:29 <09gammafunk> well if you did, that'd just generate an error in taildb/loaddb 10:18:27 <09gammafunk> I can wait for your docker, but let me know if it can't complete and you have to go 10:18:27 <12ebering> which I did 10:18:36 <12ebering> well that means my local db/log are up to date now 10:18:40 <12ebering> so should be fast 10:21:17 <12ebering> missing nemelex data tho 10:22:04 <12ebering> ok, that fixes it 10:22:04 is it possible for users to bonk sequell into picking the next nchoice? 10:22:08 <09gammafunk> cool 10:22:11 <09gammafunk> !nchoice 10:22:11 <04Cerebot> Time for a new nchoice! It will appear shortly on the tournament website (if it hasn't yet). Type "=nemelex XXXX" to update !nchoice with the new combo, where XXXX should be replaced by the new combo. 10:22:17 <09gammafunk> it's not sequell, but t scripts 10:22:23 <09gammafunk> check t page spicy 10:22:42 thanks! 10:23:01 <09gammafunk> taildb restarted, so we'll see in a sec 10:23:25 =nemelex DjSu 10:23:27 Defined keyword: nchoice => DDAM|DjCA|BaAK|TeEn|FoIE|PaAr|GhDe|HODe|DjSu 10:25:13 spicyCebolla: If you see it's not updated in sequell, you can run that command in one of the valid channels to update, just might want to check to kw to ensure that you didn't miss one (sometimes late at night, multiple may be missed if they're won quickly) 10:25:55 <09gammafunk> doh! 10:26:03 <09gammafunk> > _mysql_exceptions.ProgrammingError: (1064, "You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'AS DECIMAL(7,0))' at line 2") 10:26:30 <09gammafunk> guess @ebering had his revenge about syntax error, hopefully I can fix if he's not around 10:26:48 <12ebering> I pushed a fix to the fix 10:27:06 <09gammafunk> ok, retrying 10:28:28 <06advil> Maurice says, "Very interesting." You continue putting on your amulet of magic regeneration. x2 Maurice hits you with a +1 dagger of venom! Maurice steals your amulet of magic regeneration! Maurice puts on an amulet of magic regeneration. You are poisoned. -> ASSERT(item.base_type == OBJ_ARMOUR) in 'item-prop.cc' at line 1318 failed. 10:28:37 <06advil> pretty good 10:28:41 <09gammafunk> nice! 10:28:55 <09gammafunk> it's going through player pages now 10:29:05 <09gammafunk> seems we have too many players 10:30:12 <09gammafunk> what the heck 10:32:05 <09gammafunk> yermak's overview score disagrees with player page total score by 3 points? 10:32:05 <06advil> haha now that sounds like a rounding issue 10:32:05 <09gammafunk> https://cdn.discordapp.com/attachments/747522859361894521/939936196933599292/unknown.png 10:32:05 <09gammafunk> yeah, heh 10:32:05 <06advil> I wonder what this player was even doing (swapped amulets from regen during this maurice fight), does MR affect maurice? 10:32:05 <09gammafunk> that rounding didn't seem quite equivalent 10:32:05 <09gammafunk> https://cdn.discordapp.com/attachments/747522859361894521/939936348402499704/unknown.png 10:32:05 <09gammafunk> @ebering it's much better, but looks like not quite perfect, suspect rounding approach differences 10:32:22 <12ebering> yes, rounding errors will still be sad because python vs mysql rounding 10:32:49 <09gammafunk> and I suppose pulling that player page score from the db isn't feasible? 10:32:57 <09gammafunk> in terms of performance, or something 10:33:33 <09gammafunk> I'm sure the python code can be made to agree with a bit more work, at least 10:36:47 <03wheals> huh, i could have sworn we fixed the bug with Maurice stealing items being equipped 10:36:51 <12ebering> Yeah I don't remember why it was done that way 10:42:11 <06advil> maybe it was fixed for EquipOffDelay but not EquipOnDelay? 10:42:41 <06advil> since there was code handling this for the off case already, I just copied it and it seemed to work 10:43:09 <06advil> (I was able to replicate it, so it was definitely not fixed) 10:44:10 03advil02 07* 0.29-a0-91-gdebf466: fix: don't crash on theft during EquipOnDelay 10(4 minutes ago, 2 files, 7+ 0-) 13https://github.com/crawl/crawl/commit/debf4667c2db 10:45:47 <06advil> !git f6fec0c0d40bf 10:45:47 <06advil> rip notchei 10:45:52 <04Cerebot> wheals * 0.26-a0-1030-gf6fec0c: Don't let Maurice steal stuff being worn (1 year, 2 months ago, 2 files, 34+ 3-) https://github.com/crawl/crawl/commit/f6fec0c0d40b 10:46:04 <06advil> @wheals well, your commit probably fixes some cases, but during an EquipOnDelay it isn't yet being used 10:46:26 <06advil> maybe 10:46:28 <09gammafunk> @ebering ok I think I see why, it was probably for convenience since that data wasn't loaded into a dict that gets passed to the template, so I can modify query to add it along with the other data 10:47:06 <09gammafunk> and then just use that in the template instead of trying to calculate it again 10:47:12 <06advil> nevermind, I don't see why that commit didn't work for this case, but it didn't 10:48:01 <03wheals> the worst kind of bugfix, sadly 10:48:09 <06advil> I wonder if the pointer comparison is really valid for all those cases; the crashlog you linked is definitely removal 10:48:55 <06advil> if you want to try to replicate it you could undo my commit, and then drop everything but an amulet, make Maurice & get him next to you, then wear the amulet 10:48:57 <09gammafunk> oh, hrm, do we actually already have it 10:49:29 <06advil> well, you can still replicate him stealing it by doing that 10:50:00 <09gammafunk> yeah we do have it already, so it's even easier 10:50:01 <06advil> without undoing the commit 10:52:42 ah ok, thanks gammafunk 10:52:58 i figured i couldn't do the thing, but glad i checked. next time i'll know how to do the thing 10:59:25 Unstable branch on crawl.kelbi.org updated to: 0.29-a0-91-gdebf4667c2 (34) 11:17:24 <09gammafunk> @ebering ok, pushed a fix to use the db value for the play page as well, and it seems to work ok 11:22:44 <12ebering> rad 13:24:48 Is it intentional that confused monsters will (from what I can tell) only hit themselves when they are adjacent to another monster? 13:25:07 confused monsters /can/ only hit themselves* 13:31:18 <10PleasingFungus> wasn’t aware they could hit themselves 13:31:24 <10PleasingFungus> the player can’t 13:31:42 player gets "You're too confused to move!", monsters hit themselves 13:31:49 They can! But only when next to another monster, which is the part that I don't get. 13:31:54 been true as far back as I can remember 13:32:16 If they're "messing up" the attack, it might as well happen while they're next to the player. 13:32:46 wonder if something changed such that they can't attack aligned monsters any more? 13:32:50 (Disclaimer: I only investigated this because I was peeved from Yet Another Hydra Death) 16:24:32 Unstable branch on underhound.eu updated to: 0.29-a0-91-gdebf4667c2 (34) 23:34:49 Unstable branch on crawl.develz.org updated to: 0.29-a0-91-gdebf4667c2 (34) 23:57:20 Windows builds of master branch on crawl.develz.org updated to: 0.29-a0-91-gdebf4667c2