| 00:02.13 | Ammako | does anyone else just have wow settings annoyingly reset themselves seemingly randomly |
| 00:02.34 | Ammako | between in-game and at character select screen |
| 00:13.11 | *** join/#wowuidev ycarene (~jporterf@184-155-172-207.cpe.cableone.net) |
| 00:25.32 | *** join/#wowuidev purl (ibot@208.53.50.136) |
| 00:25.32 | *** topic/#wowuidev is BfA Author Keys! http://www.wowinterface.com/forums/showthread.php?t=56267 || 7.3.x ToC: 70300 || http://pastebin.com || Battletags: http://goo.gl/2bm7n || This channel is logged, via purl |
| 00:54.26 | VIad | is it advancedinterfaceoptions with enforce settings on startup enabled? |
| 00:54.58 | Semlar | not if they've updated it |
| 00:55.03 | Semlar | that feature was removed |
| 00:55.23 | Ammako | that could be it actually |
| 00:55.58 | VIad | yeah I noticed |
| 00:56.02 | VIad | in the update that is |
| 00:56.14 | VIad | but before the update it was resetting some cvars that made certain stuff a bit iffy |
| 00:56.30 | VIad | the idea is great but would have to keep some cvar list manually maintained, lame for you :( |
| 00:56.32 | Semlar | what it was doing was making sure your settings were exactly what they were before |
| 00:56.57 | Ammako | i got 1.2.3 when pre patch dropped but never checked for more updates afterwards |
| 00:57.09 | Semlar | so it helpfully reset the "have you watched this cinematic" cvar back to "no" |
| 00:58.01 | Ammako | There's one problem solved |
| 00:58.39 | Ammako | thanks for heads up |
| 01:11.29 | Ammako | neat https://pastebin.com/rsSrDJnt |
| 01:22.00 | Ammako | Also, for whatever reason, I get taint error popups when trying to copy/paste text I entered in chat editbox, but only when framestack is on |
| 01:59.09 | Funkeh` | TheDanW: what exactly is the factor that goes into GetSpellLink failing to return a link for a real spell? |
| 02:09.30 | *** join/#wowuidev Phixion (m@rtin.ms) |
| 02:38.14 | Semlar | <PROTECTED> |
| 02:40.32 | Funkeh` | lol, fucking moron |
| 03:09.23 | Funkeh` | I envision adding an OnUpdate check to make sure the event is kept registered |
| 03:12.36 | *** join/#wowuidev jaxdahl (~foobar@23-123-208-107.lightspeed.okcbok.sbcglobal.net) |
| 03:31.58 | *** join/#wowuidev booboo (683b79ca@gateway/web/freenode/ip.104.59.121.202) |
| 03:32.54 | booboo | is there some way to temporarily fix the 'Your Character Boost is Ready' dialog popping everytime I login? |
| 03:36.19 | booboo | I think TheDanW mentioned there was a hotfix for it in the pipeline which I assume hasn't gone out yet |
| 04:01.49 | jaxdahl | i'm playing around with using C_Club.RequestMoreMessagesBefore upon logging in, but the message.author.name field is blank at first - i wonder what i need to request or wait for |
| 04:07.05 | jaxdahl | what's that command that shows all the events that fire? |
| 04:07.36 | Funkeh` | <PROTECTED> |
| 04:09.42 | jaxdahl | thanks |
| 04:11.42 | jaxdahl | this blizzard API documentation stuff is pretty neat, started getting added around 7.2/7.3 right? |
| 04:15.45 | jaxdahl | if i understand this correctly, it adds the messages to the frame, then when it gets updated member info, it transforms the messages? |
| 04:15.51 | jaxdahl | in the community chat frame |
| 04:28.55 | foxlit | Andols: Is it possible to do a diff between 2 specific builds on townlong-yak? << it might be |
| 04:30.01 | foxlit | diffs are precomputed for specific build pairs; if you hit one of those, it should show just fine |
| 04:33.30 | foxlit | (the URL syntax is /framexml/newBuild/Example.toc/diff/oldBuild (both newBuild and oldBuild should resolve tags, so you could use e.g 7.0). There are no links to alternative diffs) |
| 04:33.40 | jaxdahl | is the search there currently broken? |
| 04:35.24 | jaxdahl | it's probably that i do not understand the scope of the search. |
| 04:36.02 | foxlit | jaxdahl: maybe, but probably not in the way you imagine it to be |
| 04:37.13 | foxlit | (it's not a full-text search over framexml, but rather a sort of "go to definition" thing) |
| 04:46.55 | jaxdahl | yeah it doesn't seem to pick up definitions in globalstrings |
| 04:47.15 | jaxdahl | works fine for functions it seems |
| 04:51.39 | foxlit | yep, it mostly ignores non-function global assignments |
| 04:52.18 | foxlit | (there are a few exceptions involving some special tables and mixins) |
| 05:01.09 | *** join/#wowuidev bluspacecow (~bluspacec@2404:4402:1349:7e00:dc2b:896a:a1a0:75f9) |
| 05:01.09 | *** mode/#wowuidev [+v bluspacecow] by ChanServ |
| 06:36.44 | *** join/#wowuidev Tentacled (~sinus@p4FE07948.dip0.t-ipconnect.de) |
| 07:02.13 | *** join/#wowuidev VIad (~Vlad@cm-84.213.146.58.getinternet.no) |
| 07:10.01 | *** join/#wowuidev Andols2 (~yaaic@m83-185-247-56.cust.tele2.se) |
| 07:31.40 | *** join/#wowuidev stolenlegacy (~stolenleg@unaffiliated/stolenlegacy) |
| 07:58.56 | *** join/#wowuidev modernist (~mke@2a00:23c5:3481:ed00:9ac:bfb0:3ea6:a23c) |
| 08:03.29 | *** join/#wowuidev Ssateneth (~ident@unaffiliated/ssateneth) |
| 08:10.27 | *** join/#wowuidev Miiru (~Miiru@2a02:8109:ec0:5558:5092:178d:4331:13f2) |
| 08:20.52 | *** join/#wowuidev bluspacecow (~bluspacec@2404:4402:1349:7e00:9135:5915:c26c:a425) |
| 08:20.52 | *** mode/#wowuidev [+v bluspacecow] by ChanServ |
| 08:30.42 | bluspacecow | https://us.battle.net/forums/en/wow/topic/20765996631?page=1#post-8 any suggestions ? |
| 08:33.05 | AcidWeb | Guy who approved guild panel changes in communitiesui was never in guild or did not understand what he was doing. |
| 08:33.52 | AcidWeb | In what universe decreasing permission granularity is a good idea? |
| 08:35.37 | Torhal | AcidWeb: The thing that threw me for a loop was the fact that suddenly I'm not seeing guild chat. |
| 08:35.46 | Torhal | After I opened the panel and saw it all there, I WTFd. |
| 08:39.18 | bluspacecow | *fluffs Torhal's hair* |
| 08:45.59 | AcidWeb | Entire Community implementation feels terribly unpolished. And... it is nothing special. But Blizzard advertises it like it is second coming. |
| 08:46.48 | jaxdahl | felt like they are trying to copy slack/discord style chat |
| 08:49.36 | Simca_ | communities are nice, BUT yes are you completely right that the lack of permission granularity sucks a lot. we like having everyone able to change the guild note in our small guild, now we can't. |
| 08:49.45 | Simca_ | that's a huge step backwards |
| 08:50.13 | Simca_ | also, after you've added the communities you want to your chat window |
| 08:50.28 | Simca_ | the screen that is pulled up when you push J is still chat screens |
| 08:50.31 | Miiru | I do personally not understand the decision to include the guild functions into the community pane |
| 08:50.33 | Simca_ | that is pretty pointless |
| 08:50.42 | blucleancow | one step forward |
| 08:50.45 | blucleancow | 2 steps back |
| 08:50.59 | Simca_ | when I push J, I want to come up to the guild roster screen. my guild chat and communities chat are in my chat window now |
| 08:51.09 | Simca_ | i don't need a second chat window on a keybind |
| 08:51.34 | blucleancow | you stick you right foot in |
| 08:51.37 | blucleancow | your left foot out |
| 08:52.36 | blucleancow | Think I saw a thread on UI and Macros Forum (US) on a way of setting it back. some complicated macro |
| 08:52.53 | Simca_ | the guild tab in general should never default to the chat window mode |
| 08:53.00 | AcidWeb | If we talking about pointless changes - you can minimalize community window but it is still not movable by default. |
| 08:53.42 | Simca_ | i know they want consistency between guild and community, but having community opening to chat makes sense because having it in the chat window is optional. having guild chat in chat is standard, so you don't want to see the guild chat window when opening the guild tab, you want to see guild news, or the roster |
| 08:54.33 | Simca_ | like i see the train of logical thought they took to reach these decisions, but if the UI were less logical and more functional, it'd be wildly better for everyone involved |
| 08:54.55 | AcidWeb | "Invite member" button there is also disabled and should work as /ginvite |
| 08:55.01 | Simca_ | they went for standardization over usability and that doesn't make sense in a customer-oriented experience |
| 08:55.29 | bluworkcow | There's Cvar you change that changes which tab comes up when you push the guild/communities show key binding |
| 08:55.32 | bluworkcow | IIRC |
| 08:55.44 | Simca_ | really?! |
| 08:55.48 | nevcairiel | there are plenty reasons to show the chat, for one it shows backlog, which your chat window does not |
| 08:55.59 | nevcairiel | i read that quite often after login |
| 08:56.25 | nevcairiel | also if you change the tab, it remember that for this session at least |
| 08:56.50 | nevcairiel | so i login, read backlog, switch to roster to see who is t here, and it stays like that =p |
| 08:57.06 | Simca_ | only if you don't switch to a community view and back because it doesn't remember there |
| 08:57.44 | Simca_ | but do you think that if we measured the number of times the panel was opened that you would find that viewing guild chat history beat out looking at the roster? |
| 08:57.58 | Simca_ | because i find that unlikely, and it validates my point that they didn't reasonably consider usability |
| 08:58.20 | bluworkcow | guildRosterView GuildInfo |
| 08:58.26 | bluworkcow | takes you straight to the guild info page |
| 08:58.37 | bluworkcow | <PROTECTED> |
| 08:58.44 | bluworkcow | takes you to the old guild roster UI |
| 08:59.20 | Simca_ | oh, i didn't even know that existed. that's neat. |
| 08:59.22 | nevcairiel | also personally, i hate inconsistent UIs, because it gets really hard to predict what it'll do |
| 08:59.40 | bluworkcow | of course that's the console mode way of setting it |
| 08:59.48 | bluworkcow | I guess a /console guildRosterView GuildInfo |
| 08:59.51 | bluworkcow | may work |
| 08:59.58 | nevcairiel | so if i press this button, it'll do this, but if i press this button, which is basically the same thing, it'll do something e lse? wut? |
| 09:00.09 | bluworkcow | whats this red button do ? |
| 09:02.07 | bluworkcow | actually woops that won't work. forgot I switched to guild info and it remembers the last used pane |
| 09:02.23 | Simca_ | yeah it didn't work |
| 09:02.49 | nevcairiel | it definitely uses the guildRosterView cvar tho |
| 09:02.57 | bluworkcow | :( |
| 09:03.18 | bluworkcow | i tried GuildInfoTab as that what fstack says the button is |
| 09:03.36 | nevcairiel | possibly options in GUILD_ROSTER_COLUMNS |
| 09:03.59 | nevcairiel | for some reason |
| 09:04.42 | Simca_ | i mean usually it's using enums for cvars bluworkcow not frame names |
| 09:04.54 | bluworkcow | :( |
| 09:04.54 | Simca_ | are you sure it isn't just like guildRosterView 3 |
| 09:04.56 | Simca_ | or something |
| 09:05.00 | bluworkcow | i dunno |
| 09:05.10 | bluworkcow | when i first saw that cvar it had a string argument in it |
| 09:05.17 | Simca_ | oh ok |
| 09:05.34 | Simca_ | nevcairiel: that probably remembers last sorting order or something |
| 09:05.42 | bluworkcow | you could track down where it's stored then clear it's value and use console mode to see what it's initially set to |
| 09:05.57 | *** join/#wowuidev Ssateneth (~ident@unaffiliated/ssateneth) |
| 09:06.39 | bluworkcow | https://us.battle.net/forums/en/wow/topic/20766186544#post-16 |
| 09:07.13 | bluworkcow | kk work tiems |
| 09:11.43 | Simca_ | i guess in the meantime i'll use that |
| 09:11.50 | *** join/#wowuidev ycarene (~jporterf@184-155-172-207.cpe.cableone.net) |
| 09:23.54 | Andols2 | Anyone knows where I should start looking if I want to redirect /G to /1 or whatever number that community is at? |
| 09:24.12 | nevcairiel | do you not have a guild? |
| 09:25.21 | Andols2 | Strange situation, the guild is currently on 3 realms |
| 09:26.05 | Andols2 | An not playing mythic so does not matter in that way |
| 09:29.31 | nevcairiel | the actual sending is done in ChatEdit_SendText |
| 09:29.33 | nevcairiel | might start there |
| 09:29.47 | nevcairiel | where it sends to is controlled by the "chatType" attribute |
| 09:31.18 | Andols2 | Ok :) Then I got a place to start :) |
| 10:15.23 | *** join/#wowuidev ycarene (~jporterf@184-155-172-207.cpe.cableone.net) |
| 10:37.29 | *** join/#wowuidev Sintacks (~Syntax@75-109-22-174.swbycmtk01.res.dyn.suddenlink.net) |
| 10:40.53 | *** join/#wowuidev TheDanW (~TheDanW@198.74.38.57) |
| 10:40.53 | *** mode/#wowuidev [+o TheDanW] by ChanServ |
| 10:49.58 | *** join/#wowuidev jinxybob (~jinxybob@host86-180-70-151.range86-180.btcentralplus.com) |
| 11:15.21 | AcidWeb | I presume you still can't override loading screen textures? |
| 11:41.12 | *** join/#wowuidev Shishire_ (~emperorsh@c-24-62-210-13.hsd1.nh.comcast.net) |
| 11:46.00 | *** join/#wowuidev Shishire (~emperorsh@pdpc/supporter/student/shishire) |
| 11:46.25 | Miiru | meh I am bugged cant access the island stuff |
| 12:01.05 | *** join/#wowuidev Ressy (Ressy@WoWUIDev/WoWAce/ARL/Ressy) |
| 12:27.37 | VIad | it would be amazing if the system allowed in the guild UI to select a community or communications and make /g /o redirect to it automatically just to make cross realm guilds a bit easier. |
| 12:46.27 | *** join/#wowuidev stolenlegacy (~stolenleg@unaffiliated/stolenlegacy) |
| 13:56.14 | *** join/#wowuidev Shadnix_ (dagger@p4FF9F22B.dip0.t-ipconnect.de) |
| 14:02.57 | *** join/#wowuidev modernist (~mke@2a00:23c5:3481:ed00:9ac:bfb0:3ea6:a23c) |
| 14:03.36 | *** join/#wowuidev modernist (~mke@2a00:23c5:3481:ed00:9ac:bfb0:3ea6:a23c) |
| 14:04.00 | *** join/#wowuidev modernist (~mke@2a00:23c5:3481:ed00:9ac:bfb0:3ea6:a23c) |
| 14:04.55 | *** join/#wowuidev Miiru (~Miiru@2a02:8109:ec0:5558:5092:178d:4331:13f2) |
| 14:05.28 | *** join/#wowuidev Semlar (~semlar@72.189.128.149) |
| 14:57.53 | *** join/#wowuidev modernist (~mke@2a00:23c5:3481:ed00:9ac:bfb0:3ea6:a23c) |
| 15:10.06 | Semlar | is 12 a good load average for a server |
| 15:10.21 | nevcairiel | it could be if its designed for that |
| 15:10.52 | nevcairiel | give it 32 cores and it laughs at 12 |
| 15:11.17 | Semlar | <PROTECTED> |
| 15:11.48 | Semlar | this isn't from cpu usage and i'm not actually sure how to even identify what's causing it |
| 15:12.16 | Semlar | the system has 4 cores |
| 15:12.17 | nevcairiel | linux load depicts the number of processes currently waiting for a resource, that could be waiting to be scheduled to run on the CPU, or waiting for IO |
| 15:12.45 | nevcairiel | well, currently running or waiting for a resource, to be precise |
| 15:12.46 | Semlar | it can only be disk or network IO right? |
| 15:13.09 | nevcairiel | they could also be waiting for the CPU, but you seem to rule that out |
| 15:14.19 | Semlar | <PROTECTED> |
| 15:14.49 | AcidWeb | check iowait of system |
| 15:15.23 | Semlar | <PROTECTED> |
| 15:15.45 | nevcairiel | the blocked processes would seem to be doing that |
| 15:15.50 | Semlar | there is a spike in cpu at the start, but after that the load average continues to go through the roof |
| 15:16.18 | Semlar | this actually caused the server to stop serving pages until it went back down |
| 15:16.37 | nevcairiel | thsoe processes were likely waiting on IO |
| 15:16.41 | nevcairiel | what that exactly means, who knows |
| 15:17.22 | Semlar | i'm not able to track disk io with this for whatever reason |
| 15:17.56 | Iriel | Semlar: How did memory look? |
| 15:17.57 | AcidWeb | is this system encrypt many things? |
| 15:18.17 | Semlar | the system writes to a database a lot |
| 15:18.34 | Semlar | a lot of small inserts |
| 15:18.43 | Semlar | but this only happens like once a week |
| 15:18.55 | nevcairiel | local database or on another server? |
| 15:18.57 | Semlar | local |
| 15:19.26 | Semlar | <PROTECTED> |
| 15:19.37 | Semlar | pretty stable |
| 15:19.56 | Iriel | I would expect your problem is I/O blocking |
| 15:20.12 | AcidWeb | ^ |
| 15:20.20 | Semlar | <PROTECTED> |
| 15:20.30 | Iriel | For example, is your database configured to allow any buffering or does it sync after every write? |
| 15:21.04 | Semlar | it's pretty much whatever postgres 9.5 is configured to do by default |
| 15:22.18 | AcidWeb | it is running on default config? |
| 15:22.34 | Semlar | i didn't change much |
| 15:22.56 | Semlar | <PROTECTED> |
| 15:23.14 | nevcairiel | the default postgres config is usually pretty bad, because its extremely conservative on resource usage, and most people would probably want to increase its allowed memory use |
| 15:23.22 | AcidWeb | yep |
| 15:23.55 | nevcairiel | from the graphs, it looks like some process was blocking the database for a few minutes, and all other access was put on hold, slowly piling up |
| 15:23.56 | Semlar | and then people can search the database later |
| 15:24.25 | nevcairiel | certain maintenance tasks can do that, but y ou would probably know if one of those ran, right? :) |
| 15:24.32 | Iriel | "the process that was bliocking the database" is very likely to BE the database. |
| 15:25.21 | Semlar | i wasn't performing any maintenance task on the server, but i'm also on a vps container that could theoretically be affected by other users |
| 15:25.34 | Iriel | High volume write activiy into an RDBMS is generally a bit of a challenge. |
| 15:25.56 | Semlar | i am running individual inserts, which i know is probably a horribly inefficient way to do this |
| 15:26.19 | *** join/#wowuidev Breen (~Breen@80.175.117.105) |
| 15:26.29 | Iriel | Are they in individual transactions (or autocommitted)? |
| 15:26.32 | Semlar | i don't know that much about optimizing a database though |
| 15:26.56 | Semlar | they're separate transactions |
| 15:27.17 | *** join/#wowuidev modernist (~mke@2a00:23c5:3481:ed00:9ac:bfb0:3ea6:a23c) |
| 15:27.25 | Iriel | Yeah, so each transaction is going to have to write itself to disk (and block until that is synced) |
| 15:28.06 | Iriel | How critical is this data? Do you have the option of running a db instanced tuned for high performance but low recoverability? |
| 15:28.14 | Semlar | it's not that important |
| 15:28.41 | Semlar | it wouldn't be a problem if it lost some of the data, i'd rather not lose all of it |
| 15:29.48 | Iriel | https://www.postgresql.org/docs/9.5/static/non-durability.html |
| 15:30.16 | Semlar | i think i did make these tables unlogged |
| 15:30.19 | Iriel | if you perform (good) backups, you can trade off short term durability and performance |
| 15:31.06 | Iriel | Hm, if they're unlogged then you've already given up recoverability |
| 15:31.06 | Semlar | should i be doing something like storing a number of messages into a queue and then writing those in batches |
| 15:31.28 | Iriel | It's hard to answer that in the abstract |
| 15:32.40 | Semlar | it's weird to me that this only happens occasionally, but i'm not sure what triggers it |
| 15:32.49 | Iriel | Though that general pattern works in some scenarios fairly well (queue in memory, commit in batches) |
| 15:33.14 | Iriel | Do you have any graphs of your system's request activity? |
| 15:33.15 | Semlar | is it just a snowball effect from something taking longer than usual and then it gets much worse as it tries to recover |
| 15:33.27 | Iriel | Usually contention isssues do start to escalate once they satrt |
| 15:34.12 | Iriel | One thing you can do to control for it is more aggressively throttle the number of requests you're willing to handle, create backpressure against the source. |
| 15:34.28 | Iriel | But for systems that are serving requests for humans they don't necessarily want to wait. |
| 15:35.13 | Iriel | Another possibility could be that there's an I/O intensive system job running that's causing the interactive ones to back up. |
| 15:35.49 | Iriel | (I've seen issues in the past with backups and replication jobs running during peak hours on systems and causing latency problems) |
| 15:36.08 | Semlar | i don't think this was from traffic to the server, i think it was from the script that's collecting data |
| 15:37.05 | Iriel | Another soure of issues can be having the database on the same volume as: A) it's backup B) the output file for something interacting heavily with the database c) Some other highly active process. |
| 15:38.24 | Semlar | there's a script that writes like a 50mb json file to the same disk every few minutes |
| 15:38.46 | Semlar | if this is a disk io issue that's probably not helping |
| 15:38.58 | Iriel | it's probably not. |
| 15:39.08 | Iriel | do you bzip2 (or similar) that file WHILE WRITING? |
| 15:39.15 | Iriel | If not, you could think about that, it'll help a bit |
| 15:39.24 | Iriel | (AT the cost of CPU, but you seem to have no shortage of that) |
| 15:39.31 | Semlar | i don't think so, just gzip while serving |
| 15:39.44 | Iriel | json tends to compress well. |
| 15:40.01 | Iriel | Still 50MB isn't that big. |
| 15:40.20 | Semlar | it's on an SSD too |
| 15:40.25 | Iriel | (I deal with 1GB+ json files at work, they're a pain in the ass) |
| 15:40.52 | Semlar | i bet |
| 15:41.48 | Iriel | The SSD certainly helps with seek time problems from disparate workloads, but doesn't necessarily help on pure throughput. |
| 15:42.22 | Semlar | so i can try adjusting some the postgres settings to speed things up |
| 15:42.40 | Semlar | and maybe commit inserts in larger batches |
| 15:43.18 | Iriel | If you can batch your inserts that'd probably help without sacrificing other things (other than data in the batch that hasn't been inserted yet) |
| 15:43.25 | Semlar | i have 8gb of ram and it generally isn't using that much of it |
| 15:43.27 | Iriel | But untilimately you want to measure your IO |
| 15:43.42 | Iriel | (Disk IO that is) |
| 15:44.10 | *** join/#wowuidev ScytheBlade1 (~Death@about/pxe/ScytheBlade1) |
| 15:44.29 | Iriel | Can you recreate this on demand by runnig those scripts? |
| 15:44.37 | Iriel | Does it always happen at the same time of the day? |
| 15:44.49 | Semlar | no, i don't know how to reproduce this |
| 15:45.04 | Semlar | it has happened a few times about a week apart |
| 15:45.14 | Iriel | Maybe see if you can set up some iostat logging |
| 15:45.18 | Semlar | but not at exactly the same times |
| 15:46.05 | Iriel | brb |
| 15:46.24 | *** join/#wowuidev Iriel (~Iriel@173-8-135-209-SFBA.hfc.comcastbusiness.net) |
| 15:46.24 | *** mode/#wowuidev [+o Iriel] by ChanServ |
| 15:47.01 | Semlar | i'm using grafana to generate those system graphs, but it can't get disk usage because the file it normally reads for it isn't being used by my service provider |
| 15:47.25 | Iriel | Can you just log iostat into a file? |
| 15:47.30 | Iriel | Every few seconds? |
| 15:47.56 | Semlar | probably |
| 15:48.08 | Iriel | or vmstat |
| 15:48.21 | Iriel | though I think iostat should be more useful here |
| 15:48.51 | Semlar | i have iotop too |
| 15:49.28 | Iriel | something like iostat -dtx 10 into a logfile |
| 15:50.37 | Iriel | iotop would be useful if you're there during the issue |
| 15:53.46 | Thunder_Child | You can log with iotop but it's slightly hacky. |
| 15:54.10 | Semlar | my load average is currently 2.7 |
| 15:56.25 | Semlar | <PROTECTED> |
| 16:04.25 | Iriel | So very little wait, more writes than reads, but the writes are smaller |
| 16:04.54 | Semlar | i am also currently downloading a backup of my database |
| 16:05.36 | Iriel | Hopefully if you can keep that logged to a file or something you can catch the next event |
| 16:05.55 | Semlar | yeah i'll just pipe the output to a file |
| 16:06.04 | Semlar | maybe not the most elegant solution |
| 16:07.02 | Iriel | It gets the job done |
| 16:07.15 | Iriel | What's that device backed by? |
| 16:07.19 | Semlar | this has been driving me nuts, because the load average does go up, but the cpu usage is generally very low on my server |
| 16:07.33 | *** join/#wowuidev Miiru (~Miiru@2a02:8109:ec0:5558:1c2d:5546:71b1:4190) |
| 16:07.37 | Iriel | That's because the resource you're blocked by isn't CPU |
| 16:07.58 | Iriel | So it's normal when you're blocked by something ELSE that CPU drops because the processes can't do anything |
| 16:10.20 | Semlar | i am not performing any kind of scheduled backups because i'm a terrible server owner |
| 16:10.33 | Semlar | but this data isn't that important |
| 16:11.36 | Thunder_Child | s/server owner/DBA/ |
| 16:12.41 | Iriel | Is this running in some kind of container where your DB is writing via a container image and not directly to a disk volume? |
| 16:13.49 | Semlar | as far as i know it's writing directly to the disk |
| 16:13.59 | Mr_Rabies | so stupid that WoD raids don't trigger the legacy loot option |
| 16:14.13 | nevcairiel | just level one more level already |
| 16:14.29 | Mr_Rabies | if you can solo it at that expansion level, you deserve all that loot |
| 16:14.50 | Semlar | i'm running inside of a virtuozzo container from my host |
| 16:14.57 | Iriel | Semlar: i was reacting to the device name 'ploop' is a container format, |
| 16:15.05 | Iriel | which maps to file(s) on an underlying filesystem |
| 16:15.10 | Mr_Rabies | i actually really like doing raids only one expansion old, because the bosses don't keel over in one hit |
| 16:15.15 | Mr_Rabies | there's actually a bit of a battle there |
| 16:15.17 | Semlar | ah, i have no idea how it works behind the scenes |
| 16:16.46 | Semlar | <PROTECTED> |
| 16:17.22 | nevcairiel | I'm not a fan of Virtuozzo stuff |
| 16:17.28 | Mr_Rabies | oh wow either the scaling or the removal of artifact stuff is apparent here |
| 16:17.32 | Mr_Rabies | i can no longer solo the butcher lol |
| 16:17.33 | nevcairiel | might as well have been a process blocking IO outside of your container |
| 16:18.59 | Semlar | i think i would be paying significantly more for a host that wouldn't possibly be affected by noisy neighbors |
| 16:19.30 | Iriel | Typically when running a production DB in a container, the DB is using a distinct data volume |
| 16:20.00 | Chipzz | Semlar: I agree with most of what Iriel said, but I want to add that the writes are not necessarily the problem |
| 16:20.27 | Chipzz | another thing that could be going on is a large SQL query running on the same table you're trying to insert in |
| 16:20.38 | Chipzz | (at the same time you're doing those inserts) |
| 16:20.59 | Chipzz | that query could also block the inserts |
| 16:21.14 | Chipzz | or the other way around, the inserts are blocking a large query |
| 16:21.20 | Iriel | Chipzz: I was picking on the writes because there are a lot of them, they are small, but they're going to be trying to fsync like anything which tends to get in the way, but I compeltely agree it _could_ be any other blocking "thing". |
| 16:21.38 | Chipzz | Iriel: didn't say what you were saying was wrong :) |
| 16:21.46 | Chipzz | the writes are a bad thing, agreed |
| 16:21.55 | Chipzz | but with an SSD it shouldn't be an issue |
| 16:21.56 | Iriel | Chipzz: I know, I was agreeing with your assessment 8-) |
| 16:21.58 | Semlar | this database isn't being accessed for reads outside of my direct control |
| 16:22.26 | Iriel | Really the next step has to be _measure_ the next event, or _reproduce_ the event (then measure that) |
| 16:22.29 | Semlar | it's being used as more of a backup for the data |
| 16:22.48 | Iriel | Why not just write to log files? |
| 16:22.48 | Semlar | maybe not a particularly reliable backup |
| 16:22.49 | Chipzz | also |
| 16:22.57 | Chipzz | pg requires vacuuming your dbs every now and then |
| 16:23.00 | Chipzz | have you done that? |
| 16:24.15 | Chipzz | https://www.postgresql.org/docs/9.5/static/app-vacuumdb.html |
| 16:24.24 | Semlar | i don't know if it's automatically vacuuming the db |
| 16:24.26 | Chipzz | vacuumdb -a |
| 16:24.59 | Chipzz | Semlar: how many rows are in the table(s), and do you have indexes on them? |
| 16:25.49 | Semlar | there is a time and text index on the tables |
| 16:25.55 | Chipzz | if you have a lot of indexes on a table, doing a lot of separate inserts is a bad thing because the indexes will have to be updated for each row |
| 16:27.18 | Semlar | 12 million rows |
| 16:27.48 | Chipzz | yeah, that might not be very good :) |
| 16:27.50 | Semlar | and 14 million rows |
| 16:27.59 | Semlar | in two different tables |
| 16:28.18 | Chipzz | https://techblog.covermymeds.com/databases/command-equivalents-in-postgres-coming-from-mysql/ -> "select * from pg_stat_activity;" |
| 16:28.41 | Chipzz | might be a good debugging tool if you can run it at the time the problem happens |
| 16:29.24 | Chipzz | if the data is for historical purposes, you should consider partitioning your tables |
| 16:30.28 | Chipzz | https://www.postgresql.org/docs/9.1/static/ddl-partitioning.html not sure how easy that is |
| 16:30.44 | Iriel | Chipzz: If this is due to a poorly indexed/designed/managed DB, it's a bit puzzlling why it'd only exhibit the problem intermittently, and NOT aligned with any particular batch job. Usually those kinds of isues are very predictable (aligned with eiither high request volume or some specific high-impact query). |
| 16:31.20 | Semlar | the volume of data being inserted should be relatively stable |
| 16:31.41 | Semlar | but high |
| 16:31.49 | Iriel | (Oracle's magical query optimizer notwithstanding - but we're talking postgresql here) |
| 16:32.16 | Chipzz | Iriel: updating a 12m index several times/sec is gonna be slow regardless :) |
| 16:32.48 | *** join/#wowuidev stolenlegacy (~stolenleg@unaffiliated/stolenlegacy) |
| 16:32.52 | Iriel | Chipzz: Agree, but the key word there is _regardless_, that'd be slow all the time, not just during an event. |
| 16:32.56 | Chipzz | Iriel: the 12mil index needs to updated separately for each insert statement, I suspect doing it in batch would be way more efficient |
| 16:33.14 | Chipzz | Iriel: but the inserts only happen intermittently IIRC? |
| 16:33.36 | *** join/#wowuidev stolenlegacy_ (~stolenleg@unaffiliated/stolenlegacy) |
| 16:33.47 | Iriel | Sadly I had to restart my IRC client, but I thought the inserts were mostly constant |
| 16:34.01 | Semlar | they're messages coming in from a number of chats |
| 16:34.55 | Chipzz | Semlar: TBH it sounds to me like for this usecase you want a nosql db, not pg, but just my 2c |
| 16:35.08 | Iriel | If the inserts are bursty and the bursts co-incide with the event, then 100% the first thing to do should be optimize that insert flow (batching, appropriate indexing, possibly partitioning (though you don't have enough volume or devices for me to think that's all that helpful) |
| 16:35.20 | Semlar | i was using mongo at one point with the same problems |
| 16:35.21 | Fisker | hey Iriel |
| 16:35.29 | Fisker | Getting the Pixel 3 (XL)? |
| 16:35.41 | Iriel | Fisker: only if my pixel 2 dies |
| 16:35.52 | Fisker | I see |
| 16:36.01 | Semlar | assuming it's even an issue with the database |
| 16:36.04 | Chipzz | what makes all this a bit pointless is that you said it was a VM on a shared machine... |
| 16:36.16 | Iriel | Fisker: I have too many expensive hobbies to indulge every new device 8-) |
| 16:36.18 | Chipzz | your performance is dependant on external factors anyway |
| 16:36.27 | Fisker | me too Iriel |
| 16:36.42 | Chipzz | so, meh... |
| 16:36.46 | Fisker | Destruction Derby really curbed my spending budget |
| 16:37.15 | Semlar | it's unclear, should my load average be going up if this is being caused by interference from another user |
| 16:37.44 | Chipzz | depends on how the host is configured, but another VM could starve your IO in theory, yes |
| 16:38.22 | Semlar | is there something i can do to identify whether this is being caused by external factors |
| 16:38.24 | Chipzz | Iriel: he also mentioned a full text index... that's gonna be heavy to update |
| 16:38.41 | Fisker | Looking for Pixel 3 to be released officially in Denmark this time, then work is buying me one \o/ |
| 16:39.20 | Iriel | Chipzz: yeah, agreed |
| 16:40.41 | Semlar | realistically i don't need frequent access to anything older than about a week |
| 16:40.41 | Iriel | If this really is bursty input, with heavy indexing, then the earlier thought about batching incoming messages and then processing them in large chunks is going to help overall capacity (at the cost of latency and possible consistency) |
| 16:41.54 | Iriel | The good news is, if this really is due to the overhead of inserts and indexing, you should be able to make a load test and reproduce the problem. |
| 16:42.02 | Semlar | partitioning should let me archive older messages right? |
| 16:42.11 | Iriel | Spin up another container, simulate the peak load. |
| 16:42.36 | Iriel | The challenge with partitioning is going to be your text indexes. |
| 16:43.06 | Iriel | (Though I dont know how postgres applies partitioning to indexes) |
| 16:44.49 | Semlar | alright, let's say i want data from the last week to be searchable, and older data to be kept if i need it for something, but not accessed by normal site traffic |
| 16:46.20 | Semlar | i can limit the working set of data to just the last 7 days |
| 16:47.28 | Semlar | should i partition one table or set it up differently |
| 16:48.36 | Iriel | Usually I see that approached by having 'normal' and 'archive' tables |
| 16:48.44 | Iriel | and jobs that move data during off-peak times. |
| 16:48.52 | Iriel | There's substantial complexity in trying to solve that problem with partitions |
| 16:50.36 | Iriel | Something like elastic search is designed for the kind of problem you're trying to solve |
| 16:51.09 | Semlar | so i have one table with all the live data on it that just periodically gets trimmed down at the cut-off date and older data is moved somewhere else? |
| 16:52.04 | Iriel | Semlar: That's the "simple mode" approach using an RDBMS, yes -- small inserts into live table, bulk transfers during quiet times from live to archive. |
| 16:52.27 | Semlar | the date isn't even particularly important, if it's more efficient to make the cut-off based on file sizes that works too |
| 16:52.47 | Semlar | and just move entire files |
| 16:53.31 | Iriel | At a certain point the complexity of managing your data (rather than letting a DB take care of it) becomes significant. |
| 16:54.12 | Iriel | One of the hard parts of system architecture is knowing where to put your complexity/time investment |
| 16:54.34 | Iriel | How many day's worth of data is in your current system? |
| 16:57.53 | Semlar | about a month and 3 weeks on one table, i think the other one is about 4 months of more specific data |
| 16:59.30 | Semlar | they are both essentially archives for my own use at the moment, but i want to switch the way it's currently set up to use the database more |
| 17:00.07 | Semlar | so the website will have direct access to the last ~week of data and people will be able to search it |
| 17:01.11 | Semlar | right now i have a fairly hacky setup where the live data is being written to a json file that acts as a cache that people are loading |
| 17:01.42 | Semlar | while it's also being written to the database, but not accessed directly from it |
| 17:03.57 | Semlar | the script stores the entire working set of data in ram at all times, it trims it to the desired length and periodically writes it to a json file every few minutes |
| 17:04.41 | Semlar | while writing a full log of it to the database for longer storage |
| 17:04.50 | Iriel | Well, if all of our assumptions about botttlenecks and behavior are correct, a better DB-oriented design would be: Use a queue/stream on incoming events - insert live events in chunks from queue; Periodically move archive records out of live tables. |
| 17:05.10 | Mr_Rabies | me reading this conversation https://www.youtube.com/watch?v=Ccoj5lhLmSQ |
| 17:06.19 | Iriel | Alternatively, ditch the DB completely, keep the data in files, and then use some kind or pure index (SOLR, etc) which you can rebuild as needed from the files. |
| 17:07.16 | Iriel | Ideally, before doing anything, be able to simulate your actual workload so you can measure and be sure you understand where the actual problems are. |
| 17:08.18 | Iriel | Without measurement the best we can do is "suggestions for things what might be a good approach", the probability we can come up with the optimal approach is nearly zero |
| 17:08.57 | Semlar | alright, so i should write a script that can stress test the server and see how it behaves |
| 17:09.20 | Semlar | that's probably a good idea |
| 17:16.50 | Iriel | That's how it's done in the "Real world" |
| 17:18.15 | nebula | fuck it, we'll do it live |
| 17:18.40 | Iriel | That's ALSO how it's done in the real world! |
| 17:18.58 | Semlar | pfft, i'm sure my visitors won't mind |
| 17:19.09 | Semlar | all 6 of them |
| 17:20.12 | Iriel | To be honest, the biggest advantage of scripting the stress test is that you dont have to wait around for people to break it |
| 17:21.40 | Semlar | i can probably just pull x number of messages out of the database and run them back through the script as if it were live data |
| 17:23.46 | Semlar | conveniently the database itself is not being accessed directly by visitors so it doesn't really matter if i add a bunch of fake data |
| 17:24.16 | *** join/#wowuidev ls- (~ls@49.48.51.101) |
| 17:25.15 | ls- | Simca_, foxlit, do you remember how to disable/hide <Deferred XML Node object blah blah blah> error? O_o |
| 17:25.21 | ls- | I forgot >_> |
| 17:25.54 | Simca_ | it's new in 8.0, i don't think there is an easy way to do it. nevcairiel had something that filtered on error type. |
| 17:26.05 | nevcairiel | fix the problem instead of hiding it! :) |
| 17:26.07 | Simca_ | the correct way to handle it is to stop declaring the same object repeatedly |
| 17:26.34 | nebula | just install details, problem solved |
| 17:26.35 | Simca_ | but if it isn't your addon (cough, Details!) then yeah there's not a lot of options |
| 17:26.51 | Semlar | <PROTECTED> |
| 17:27.17 | Simca_ | sigh |
| 17:27.25 | ls- | it's not for me, it's actually for this dude: http://www.wowinterface.com/forums/showthread.php?p=328847 |
| 17:27.39 | Simca_ | i can't believe he is doing that instead of not using XML in his libraries |
| 17:28.10 | Semlar | look, sometimes you have to do things quickly because you don't have time to do things correctly |
| 17:28.33 | ls- | sadly, we can't remove xml stuff from ouf >_> |
| 17:28.44 | nevcairiel | stop making it a embedded library |
| 17:28.46 | nevcairiel | :) |
| 17:28.59 | nevcairiel | have people install it as the oUF addon |
| 17:29.14 | ls- | that's what I told that dude to do :D |
| 17:29.23 | Semlar | i would personally just replace the error handler with one that conveniently filters out my addon |
| 17:29.31 | ls- | but people running multiple instances of ouf at the same time are rare, imho... |
| 17:29.34 | Semlar | thus solving the problem once and for all |
| 17:29.37 | ls- | we don't really expect anyone to do that |
| 17:30.23 | Simca_ | i mean it sucks if the addon in question is using a ridiculous amount of xml, like declaring every single frame in it |
| 17:30.39 | nevcairiel | an addon has no reason to re-declare the same thing |
| 17:30.39 | Simca_ | because the rewrite in question is going to be very extensive |
| 17:30.42 | nevcairiel | only libraries might do that |
| 17:30.46 | Semlar | transmog roulette is written entirely in one xml file |
| 17:30.56 | Simca_ | right, usually addons that double as libraries, nev |
| 17:31.18 | Simca_ | as a pure library wouldn't need xml unless written by semlar |
| 17:31.33 | Semlar | embrace it |
| 17:31.40 | Semlar | xml is the answer to all of your problems |
| 17:31.49 | Simca_ | except this one |
| 17:31.55 | Semlar | just keep throwing more xml at it |
| 17:32.04 | Semlar | it'll be fine |
| 17:32.51 | Semlar | it's always a good sign when it goes into scientific notation trying to output the number of rows in my table |
| 17:33.50 | Semlar | <PROTECTED> |
| 17:34.18 | ls- | Simca_: we'd gladly remove that last remaining xml-template from ouf, but, unfortunately, we can't do so... xml is the only way >_> |
| 17:45.02 | p3lim | Semlar: the lines between points gets really weird in argus if you're not coming from the same zone (WorldFlightMap) |
| 17:45.24 | Semlar | it's not actually supposed to draw lines in argus but i didn't turn it off |
| 17:46.01 | Semlar | the 3 zones don't have coordinates relative to each other that make any sense |
| 17:46.32 | Simca_ | ls-: ah one of those xml attributes they never provided a lua version of, i guess. should bug TheDanW about it |
| 17:47.42 | *** join/#wowuidev Ammako (~Ammako@unaffiliated/ammako) |
| 17:48.04 | nevcairiel | Simca_: its not about attributes, but some things when dealing with secure raid frames, you need a template for |
| 17:48.11 | Semlar | i don't know if foglight's going to get updated |
| 17:48.20 | Simca_ | oic |
| 17:48.37 | nevcairiel | i mean they could try to m ake that available in Lua |
| 17:48.43 | nevcairiel | but it would be some more engineering |
| 17:48.51 | Ammako | having issues with it Semlar? |
| 17:49.06 | nevcairiel | <PROTECTED> |
| 17:49.09 | nevcairiel | it wasnt that bad |
| 17:49.12 | Semlar | i have no idea how i'm supposed to draw textures behind the map with the current version of things |
| 17:49.26 | p3lim | become your own map provider? |
| 17:49.59 | nebula | the first sign of the apocalypse |
| 17:50.11 | Ammako | the mapocalypse |
| 17:51.12 | p3lim | I think MapCanvasDetailLayer.lua is the default implementation |
| 17:58.10 | nevcairiel | its not |
| 17:59.45 | Semlar | where do i go to get a list of all the mixin functions like "AddDataProvider" |
| 18:00.03 | Semlar | i don't know how this system works |
| 18:00.05 | jaxdahl | regex to the rescue |
| 18:00.11 | jaxdahl | problems++ |
| 18:02.53 | nevcairiel | those are probably in MapCanvas |
| 18:03.42 | Semlar | oh, it looks like AddDataProvider is just a lua function for MapCanvas |
| 18:03.48 | Semlar | i thought it was a built-in thing |
| 18:03.59 | nevcairiel | there really is no built-in stuff anymore |
| 18:04.21 | nevcairiel | most of the fancy new stuff is in Lua |
| 18:04.28 | nevcairiel | or documented in the API docs with a C_ prefix |
| 18:06.51 | [TOM_RUS] | why is there no timestamp in comminities chat? |
| 18:07.47 | ls- | nevcairiel: it's also about attributes, theoretically, we could use `initialConfigFunction` and shove all :SetAttributes into it, but, iirc, you can set "_*" attributes from there... |
| 18:08.22 | ls- | out of these 4 attributes: https://github.com/oUF-wow/oUF/blob/master/oUF.xml#L42-L49, you can actually set only 'refreshUnitChange' |
| 18:09.27 | ls- | as for the rest... 188x FrameXML\RestrictedExecution.lua:431: Call failed: FrameXML\RestrictedFrames.lua:470: Invalid attribute name :p |
| 18:10.05 | ls- | ugh, I meant, you can't set "_*" attributes... |
| 18:11.44 | nebula | yea, being able to support clique from lua would pretty much mean you'd never need to use xml |
| 18:12.24 | nevcairiel | ls-: yes i know |
| 18:12.31 | nevcairiel | the _ attributes was the reason i created my template |
| 18:12.31 | Constie | Sometimes I feel a little silly using kgPanels to do all sorts of random small things like other people use WeakAuras, because the config makes more sense to me. https://i.imgur.com/1VvKBPL.jpg |
| 18:13.05 | ls- | nevcairiel: I was just replying to your "its not about attributes" :> |
| 18:13.35 | nevcairiel | the point is that there is Lua api to set those |
| 18:13.46 | nevcairiel | just the protection mechanic doesnt let you |
| 18:13.50 | nevcairiel | so its not as easy as adding new api |
| 18:13.59 | ls- | ofc |
| 18:40.25 | *** join/#wowuidev Andols (~yaaic@193.10.192.28) |
| 18:43.15 | VIad | I do notice calling C_Map.GetPlayerMapPosition on myself does create a table every call. there was a talk regarding how to minimize performance impact on this. what was the outcome of that talk? |
| 18:43.42 | nevcairiel | there is no way to change how the function works =p |
| 18:44.58 | VIad | considering to only call it on specific map update events like player stops to move otherwise perhaps once a second to not overdo it :P |
| 18:45.10 | nevcairiel | throttling of course works |
| 18:45.44 | VIad | I guess the issue would be less noticable if C_Map.GetPlayerMapPosition could be fed the table you want it to re-use |
| 18:45.44 | Semlar | i'm getting a message in the chat about there being too many errors in my UI, but nothing in bugsack |
| 18:48.05 | nebula | yay for hiding errors |
| 18:50.03 | nevcairiel | there is a bug in the game that can cause debuglocals() to error |
| 18:50.14 | nevcairiel | i told TheDanW about it, but nothing yet |
| 18:50.28 | *** join/#wowuidev Andols2 (~yaaic@m83-185-247-56.cust.tele2.se) |
| 18:50.30 | nevcairiel | so that results in an error loop |
| 18:50.43 | nevcairiel | buggrabber could work around it by pcall'ing that function |
| 18:50.46 | nevcairiel | but alas, it does not |
| 18:53.57 | ls- | throwing shade, eh? :P |
| 19:02.03 | Semlar | this sure is making it harder to debug errors |
| 19:02.42 | Semlar | does this happen whenever you throw an error inside of a hook? |
| 19:04.00 | nevcairiel | i've particularly seen it happen when erroring in map things |
| 19:04.02 | nevcairiel | for some reason |
| 19:04.09 | Semlar | yeah it's not because of the hook |
| 19:04.32 | Semlar | it looks like some of the earlier code in the function, which runs successfully, it making it happen |
| 19:06.38 | Semlar | it's your math function that i copied |
| 19:07.50 | Semlar | i'm borrowing some of your code from fogclear to see why this isn't working |
| 19:08.05 | Semlar | local key = exploredTextureInfo.textureWidth * 2^36 + exploredTextureInfo.textureHeight * 2^24 + exploredTextureInfo.offsetX * 2^12 + exploredTextureInfo.offsetY this is breaking it |
| 19:09.17 | Semlar | or at least it stopped breaking the errors when i commented out this block |
| 19:09.39 | nevcairiel | i had a hunch that the large integers might be related |
| 19:09.41 | Semlar | let me see if i can narrow it down |
| 19:09.43 | nevcairiel | but its still a bug in the game |
| 19:11.07 | Semlar | okay i take it back, it's this line "pin.layerIndex = pin:GetMap():GetCanvasContainer():GetCurrentLayerIndex()" |
| 19:11.35 | Semlar | i had it commented out with the rest of the stuff |
| 19:14.17 | Semlar | if i just have this https://i.imgur.com/fmzuY5w.png it throws this in-game https://i.imgur.com/CDfttwu.png |
| 19:14.33 | Semlar | some combination of this is causing it to happen |
| 19:15.11 | nebula | i recommend you remove the breakthings() call |
| 19:15.33 | Semlar | i need to throw an error! |
| 19:16.09 | Semlar | maybe it's triggering an infinite loop with RefreshOverlays |
| 19:16.42 | nevcairiel | its definitely a problem in the error handler |
| 19:16.46 | nevcairiel | i debugged it before |
| 19:16.50 | Semlar | i'm not sure how, if i comment out local data = errata[artID] it throws an error like normal |
| 19:16.57 | Semlar | which doesn't make any sense |
| 19:17.27 | Semlar | maybe it's related to your keys |
| 19:17.44 | nevcairiel | comment out the debuglocals() stuff in BugGrabber.lua and it stops |
| 19:18.06 | nevcairiel | line 366 |
| 19:18.11 | nevcairiel | that call causes it |
| 19:18.28 | Semlar | interesting |
| 19:18.45 | nevcairiel | something in the error context causes the debug function to error |
| 19:19.07 | Semlar | hopefully it's trying to output this entire table |
| 19:23.49 | Funkeh` | if you're convinced pcall would resolve it why comment it out instread of adding it and seeing what the output is..? |
| 19:24.17 | nevcairiel | because its easier to tell someone to comment stuff then to tell them to re-write the logic to pcall it :D |
| 19:24.28 | nevcairiel | because pcall adds a second argument etc |
| 19:24.32 | nevcairiel | return value, i mean |
| 19:25.16 | Funkeh` | well i meant for yourself |
| 19:26.07 | nevcairiel | i did that |
| 19:26.15 | nevcairiel | its some weird error internal to the function with next |
| 19:29.13 | *** join/#wowuidev jinxybob (~jinxybob@host86-180-70-151.range86-180.btcentralplus.com) |
| 19:43.48 | *** join/#wowuidev ls- (~ls@183.88.0.177) |
| 20:26.16 | *** join/#wowuidev Andols (~yaaic@193.10.192.28) |
| 20:28.39 | *** join/#wowuidev Andols2 (~yaaic@m83-185-247-56.cust.tele2.se) |
| 20:30.28 | *** join/#wowuidev Andols3 (~yaaic@m83-185-247-56.cust.tele2.se) |
| 20:38.14 | jaxdahl | the CLUB_STREAMS_LOADED event fires, then i have a list of clubs+streams, but i do not know which ones it is going to subscribe to until the CLUB_STREAM_SUBSCRIBED event fires. but then I do not know exactly when it is done subscribing to them |
| 20:45.58 | *** join/#wowuidev Tentacled (~sinus@p4FE07948.dip0.t-ipconnect.de) |
| 20:49.06 | *** join/#wowuidev Andols (~yaaic@78-70-44-4-no170.tbcn.telia.com) |
| 20:51.05 | Semlar | foxlit: is there a quick way to have luacasc download and open a specific db2 file from the cdn |
| 20:51.53 | nevcairiel | cdn access is always inherently slow-ish |
| 20:52.00 | nevcairiel | but you can set a cache path so repeated calls are ok |
| 20:52.06 | Semlar | i'm not sure what casc.open("http://eu.patch.battle.net:1119/wow/#eu") does |
| 20:52.23 | nevcairiel | dont you have wow installed anyway |
| 20:52.27 | nevcairiel | just point it at that |
| 20:52.28 | Semlar | i want to run this on a server |
| 20:52.33 | nevcairiel | i see |
| 20:52.53 | Semlar | preferrably without installing a 60gb game |
| 20:53.28 | nevcairiel | i used it like this in the past, with a cache path, which helps future lookups https://pastebin.com/jXMBwQy0 |
| 20:53.37 | nevcairiel | (adjust url as required) |
| 20:58.39 | Semlar | it looks like it has to download about 200mb of index files before it can look for something? |
| 20:58.48 | nebula | yes |
| 20:58.48 | nevcairiel | that sounds about normal |
| 20:58.58 | Semlar | that took 3 minutes to run |
| 20:58.59 | nebula | it has to get the encoding table |
| 20:59.09 | nevcairiel | not sure how low that gets when the cache is hot |
| 21:00.28 | Semlar | what does it do after a hotfix? |
| 21:00.39 | nevcairiel | probably doesnt account for hotfix data |
| 21:00.49 | nevcairiel | since thats seperate |
| 21:02.39 | nevcairiel | if the server version actually changes, it'll probably download a new index |
| 21:04.27 | *** join/#wowuidev modernist (~mke@2a00:23c5:3481:ed00:9ac:bfb0:3ea6:a23c) |
| 21:08.47 | jaxdahl | when did dalaran mapID change from 1014 to 627? |
| 21:10.08 | Ammako | Didn't all (most?) mapIDs change with 8.0? |
| 21:10.19 | nevcairiel | all of them, yes |
| 21:11.40 | nevcairiel | the entire map system was changed, and all maps were re-assigned a new type of ID |
| 21:14.14 | jaxdahl | so it's C_Map.GetBestMapForUnit("player") to get the zone i'm in? |
| 21:14.21 | nevcairiel | yes |
| 21:22.19 | Semlar | assuming the map is what you're after |
| 21:23.38 | Mr_Rabies | i really hope someone sacrifices their soul to fix libtourist lol |
| 21:23.52 | nevcairiel | libtourist should just die, its so pointless |
| 21:23.59 | nevcairiel | and probably entirely in need of a rewrite |
| 21:24.04 | Semlar | what does it do |
| 21:24.39 | Semlar | isn't most of this built into the world map? |
| 21:25.36 | nebula | I am currently working on adapting LibTourist to these changes - please be patient... -the maintainer |
| 21:26.02 | Mr_Rabies | i don't have a dog in the fight, except i've got other addons that would work fine so long as the lib was updated :v |
| 21:26.59 | nevcairiel | you would think so, but probably not |
| 21:27.07 | Semlar | ah it's probably mostly for the boats and zeppelins and portal location |
| 21:27.16 | nevcairiel | those actually show on the default map now |
| 21:27.43 | Semlar | they would need to be tied to where they take you |
| 21:27.52 | Semlar | if this is being used to calculate the fastest route somewhere |
| 21:29.40 | nebula | they're "map links" so you do get both |
| 21:39.33 | *** join/#wowuidev Shooter (~pi@80.49.169.131.ipv4.supernova.orange.pl) |
| 21:39.42 | jaxdahl | channel topic is in need of a slight update |
| 21:43.56 | ThiefMaster | is there any util function that gets the (pet) spellbook id for a given spell name? trying to fix a weakaura that tries to count the amount of targets in pet melee range by checking `IsSpellInRange("Claw", "nameplate" .. i)` but apparently this API does not support pet spells when using a spell name, only when using a spellbook id |
| 21:47.53 | *** join/#wowuidev Ketho (~Ketho@2a02:a44e:ac87:1:7db6:c63c:f0f3:edaf) |
| 22:27.28 | Tentacled | For traveling around you could just use Zygor's addon... In its basic guide-less version it has a pretty nifty travel system. (Shameless plug.) |
| 22:29.27 | VIad | nevcairiel: would one be able to use UnitPosition("player") and another API to calculate the world map coordinates? |
| 22:29.39 | nevcairiel | of course |
| 22:30.33 | nevcairiel | with C_Map.GetMapPosFromWorldPos |
| 22:30.49 | VIad | I wonder if that is cheaper |
| 22:30.57 | VIad | might not be :P |
| 22:31.07 | nevcairiel | it not only returns a table, but also wants you to give it the input as a table =p |
| 22:31.18 | Tentacled | And that's a million dollar question, seeing as C_Map insists on using horribly wasteful vector tables. |
| 22:31.20 | VIad | yeah. the input table could be re-used but hmm |
| 22:31.32 | VIad | still returns a table so :/ |
| 22:31.52 | nevcairiel | you can use my map library, its table-free! |
| 22:31.55 | VIad | ! |
| 22:32.00 | VIad | that is true |
| 22:32.03 | nevcairiel | well, it doesnt make any for positions querys a nyway |
| 22:32.06 | nevcairiel | it contains some internally! |
| 22:32.11 | Tentacled | But is it s-table? ;) |
| 22:39.36 | VIad | nevcairiel: the UnitPosition ID in arg4 is 1220 any easy way to make it into the uiMapID internally? |
| 22:39.52 | nevcairiel | those dont map to each other like that |
| 22:39.56 | VIad | yeah figured |
| 22:40.06 | nevcairiel | 1220 is like the entire continent |
| 22:40.14 | nevcairiel | not a single zone |
| 22:40.25 | nevcairiel | those numbers refer to actual physical zones |
| 22:40.31 | nevcairiel | so a continent, or an instance |
| 22:40.33 | Alram | Can y'all use worldmapoverlay stuff in Lua? |
| 22:41.24 | nevcairiel | the api named above can actually determine which uiMapID is the most appropriate based on the instance id and coordinates |
| 22:41.32 | nevcairiel | so it uses the overlay stuff |
| 22:41.37 | nevcairiel | but we dont get direct data access |
| 22:41.43 | VIad | yeah |
| 22:41.47 | Alram | Fair enough |
| 22:41.58 | VIad | there goes the UnitPosition dream for easy and cheaper coordinates |
| 22:43.21 | nevcairiel | like I mentioned, if you're really overly concerned about getting coordinates in OnUpdate and wasting tables, you can use HereBeDragons, which has no memory waste for that task |
| 22:44.14 | VIad | more saddened by the fact there is no way to at least give it a table to populate and it wastes memory like that |
| 22:44.22 | VIad | everything else was amazing API wise :) |
| 22:44.27 | VIad | loads of improvements |
| 22:59.39 | nebula | doesn't really matter |
| 22:59.47 | nebula | the combatlog is complete shit and yet the game runs fine |
| 23:14.37 | VIad | cleu dont return tables though |
| 23:18.15 | Semlar | pre-optimization is the root of all evil |
| 23:18.17 | nebula | doesn't need to, in a raid it still creates a ton of garbage |
| 23:20.26 | jaxdahl | i have an addon that was related to Artifact Power items that I believe is no longer useful, not sure if there's some way to mark an addon as obsolete on curseforge, or if i should just delete it |
| 23:22.40 | Ammako | Might as well repurpose it for azerite |
| 23:22.58 | nebula | you don't use azerite items, though, you just gain it like exp |
| 23:23.11 | nevcairiel | yeah the items dont really exist |
| 23:23.15 | nevcairiel | they get auto-vacuumed |
| 23:23.19 | jaxdahl | https://www.curseforge.com/wow/addons/artifactpowersaver it was this |
| 23:23.22 | nebula | and i don't think you can change the status of your project anymore |
| 23:23.26 | Ammako | rip |
| 23:23.47 | Ammako | I was trying to look up how it works but couldn't find for certain |
| 23:24.58 | Torhal | jaxdahl: I can mark it as abandoned if you'd like. |
| 23:25.34 | nebula | yea, i saw the cp could still change to inactive/abandoned, but you don't have a project setting for it |
| 23:45.31 | jaxdahl | naming a new addon is hard |
| 23:48.26 | jaxdahl | 'Community Chat History' or 'Backtalk' |
| 23:57.45 | jaxdahl | what does the 'This Project is Experimental' checkmark do |
| 23:59.25 | nebula | not much anymore, it use to keep you out of the mod queue (and unpublished) until you uploaded a file/tagged a release |
