Minetest logo

IRC log for #minetest-hub, 2019-09-08

| Channels | #minetest-hub index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:04 NathanS21 joined #minetest-hub
01:29 rdococ hmm
01:30 rdococ I think that, if shaders are enabled at least, fog could vary in colo
01:30 rdococ r
03:46 sofar does anyone know how to modify the builtin sprint speed?
04:10 ANAND sofar: fast_move without flying?
04:11 sofar yes
04:11 sofar afaics the engine just applies a hardcoded modifier
04:11 sofar nothing to tell the client that it's modifier is a specific value
04:11 sofar same for sneak speak
04:11 ANAND There's https://github.com/minetest/minetest/blob/36bfc67574aaa54c3e7f814ec7c5336e9dbc1ac4/minetest.conf.example#L1370
04:12 ANAND The `movement_speed_fast` setting
04:12 ANAND AFK
04:12 sofar huh, maybe I was looking at the wrong code then
04:36 sofar that works, lol
05:42 ANAND :D
07:32 Krock joined #minetest-hub
07:44 ANAND Hi Krock :)
07:50 Krock hi ANAND
08:42 ShadowNinja joined #minetest-hub
10:01 Krock joined #minetest-hub
10:11 Fixer joined #minetest-hub
10:58 lisac joined #minetest-hub
13:13 rdococ hi
13:20 Krock hi
13:25 ANAND hi
13:25 jas_ !mod
13:25 MinetestBot jas_: Sailing and Tropical Islands [sailing] by cakenggt - https://forum.minetest.net/viewtopic.php?t=22159
13:33 Krock !mod
13:34 MinetestBot Krock: BigPixelZombie [bpz] by azekill_DIABLO - https://forum.minetest.net/viewtopic.php?t=14495
14:39 tomraceror joined #minetest-hub
16:02 VanessaE oof.  worldedit's  //move function needs some TLC.  it's slow compared to //set  -- a region about 1.1M nodes is taking many minutes to move down by 1m, yet a region of 54M nodes took only some tens of seconds to //set.
16:02 VanessaE (it's been chewing on that //move for over 10 mins now)
16:02 VanessaE and using a lot of memory, too..
16:04 sfan5 it intentionally does the moving in steps
16:04 sfan5 maybe it shouldn't
16:04 VanessaE wasn't there some kind of "buffer" that lets big commands run in the background?
16:05 sfan5 no idea
16:07 VanessaE it's grabbed about 2GB of RAM since the command started; does it load the area to move, into a table or something?
16:15 VanessaE who actively maintains WE these days anyway?
16:15 VanessaE (so that I can yell at someone more appropriate ;) )
16:17 sfan5 me
16:17 VanessaE oh ok.
16:18 VanessaE consider yourself yelled at then ;)
16:33 VanessaE oof.  //move uses remove_node() and set_node() and meta-handling
16:33 VanessaE no wonder it's so slow
16:59 VanessaE HybridDog has a fork, where he claims some operations are ridiculously-fast, I think one of which is //move [off] though I tend to avoid him as I don't trust him due to his level of paranoia, this may be one time it's okay to look
16:59 VanessaE shit
17:00 VanessaE missed a \n  :P
17:11 VanessaE sfan5: any chance you can do something to make //move faster?  it's been going for over an hour...
17:12 sfan5 working on it
17:12 sfan5 right now, in fact
17:12 VanessaE oh! ok.  thanks :)
17:17 aerozoic joined #minetest-hub
18:43 tumeninodes joined #minetest-hub
18:45 tumeninodes rubenwardy: having a slight issue with ContentDB updating a couple mods. Got a notification "Oops something went wrong" so waited and tried again. WOuld have used the mssg option but did not want to  add a third update instance. The two involved are https://content.minetest.net/packages/TumeniNodes/pkarcs/ and https://content.minetest.net/packages/TumeniNodes/angledwalls/ both show a status of "importing" since last night
18:46 rubenwardy redis broke
18:46 rubenwardy I hate technology
18:47 tumeninodes oh nos
18:47 sfan5 does remove_node delete metadata?
18:47 tumeninodes tech is a love/hate thing
18:47 sfan5 i guess not and it doesn't matter anyway
18:49 Krock it should hmm
18:51 Krock addNodeAndUpdate(p, MapNode(CONTENT_AIR), modified_blocks, true);
18:51 Krock > true. removes meta
18:51 Krock meta is removed or Map::removeNodeMetadata(): Block not found is logged
18:51 sfan5 i see
19:14 sfan5 minetest needs an api function to get 1) a monotonic time and 2) one with precision better than seconds
19:20 Krock minetest.get_us_time()`
19:21 sfan5 oh cool
19:21 sfan5 i thought that there was one but couldn't remember the name
19:21 Krock sounds like you're modding again?
19:22 sfan5 i'm making WorldEdit suck less
19:23 Krock til worldedit sucks
19:38 VanessaE I wouldn't say it sucks.  that's a bit extreme ..
19:42 tumeninodes rubenwardy: why doe the link to report issues on CDB link to your Twitter acct? This means that anyone who wishes to report an issue is required to have a Twitter acct? 0_o I only mention it as this may cause some to not bother issuing a report, with no desire to have a Twitter acct.
19:42 tumeninodes does
19:45 rubenwardy doesn't it link to rubenwardy.com/contact?
19:45 rubenwardy oh wait, you're referring to the site-wide error page
19:45 rubenwardy it's outdated
19:46 tumeninodes yes, the error page which popped up when I ran into the update issues
19:46 rubenwardy that appears on all sites on my machine
19:46 tumeninodes oh
19:46 rubenwardy I created that page years ago
19:46 rubenwardy >>> 'logfile /var/log/redis/redis-server.log' // Can't open the log file: Permission denied
19:46 rubenwardy any ideas? Permissions are correct
19:46 tumeninodes kick it
19:47 Krock chmod 777
19:47 tumeninodes or that
19:52 tumeninodes anyone here use hugo? (just asking as I'm thinking of creating a small website) https://gohugo.io/
19:54 Krock ^ Calinou please use this
19:55 Krock I don't know what it's capable doing but it surely fits
19:56 sfan5 oh great the new copy code fails randomly
19:56 Calinou I used it a bit for prototyping, but my current personal website uses Jekyll actually :P
19:56 sfan5 is it possible that two vmanips in very close locations conflict with each other when only one(!) of them is actually written back to the map?
19:56 Calinou Hugo is still pretty cool
19:56 Krock Calinou: use it for its name
19:56 Calinou :D
19:57 tumeninodes https://dzone.com/articles/static-site-generators-overview-gatsby-vs-hugo-vs
19:58 Calinou Gatsby has a different approach, with many advantages but also downsides
19:58 Calinou >Learning Golang can be difficult.
19:59 Calinou uh, you don't need to know Go to use Hugo
19:59 Calinou (also, it's Go, not Golang :P)
19:59 tumeninodes no xml support though
19:59 sfan5 oh i found the bug
19:59 Krock parallel vmanips should be possible
19:59 tumeninodes xml = still relevant
20:00 sfan5 Krock: they are, the bug was in my code
20:00 Calinou yeah, it's not supported yet, https://github.com/gohugoio/hugo/issues/4470
20:00 Calinou you could maybe convert XML to another format using a separate tool still
20:02 tumeninodes tough getting an accurate result doing a vs x 3 and no knowledge if the organization doing the vs may have a bias
20:03 tumeninodes another matter is apache 2.0 vs MIT
20:05 Calinou what do you want to license?
20:05 tumeninodes I personaly like apache
20:06 tumeninodes me personaly on an aspect which would matter? nothing
20:07 Calinou usually, you should pick the license that's used the most often in the community in question
20:07 Calinou Rust software is often licensed under both Apache 2 and MIT
20:08 tumeninodes anything I put on a website would be under it's own specific license as far as my own content
20:08 Calinou content is better licensed using a Creative Commons license, preferably CC0, CC BY or CC BY-SA :)
20:08 Calinou but you can license the code under MIT or Apache 2 just fine
20:10 tumeninodes apache might have been an option related to MT project? I am only saying as it can contain patents and (correct me if I'm wrong) maybe Celeron could have applied for a patent for MT engine?
20:12 tumeninodes just so it existed. Again, I am no legal expert, it is just a quick thought that passed into my thick skull
20:12 Calinou no, that's not what a patent grant is
20:12 tumeninodes well.... poop
20:13 Calinou https://opensource.com/article/18/2/how-make-sense-apache-2-patent-license
20:13 Calinou (found via a quick search, there's more comprehensive articles out there)
20:13 tumeninodes I was thinking it might have been a way around the trademark matter of the past
20:15 sfan5 VanessaE: done
20:15 VanessaE yay! :)
20:16 sfan5 there's a catch though: //move y -1 is as slow as before
20:16 rubenwardy joined #minetest-hub
20:16 VanessaE bummer.
20:16 tumeninodes Calinou thank you for that link... just started reading and bookmarked to finish later tonight. Very good stuff to know
20:16 sfan5 instead use //copy y -1 and manually remove the remaining row at the top OR move whatever thing high into the air and then back down where you want it
20:17 VanessaE oh I see, so that the source and destination regions don't overlap?
20:17 sfan5 yeah
20:17 sfan5 /copy is still slow when regions overlap AND you are moving towards +, //move is still slow when regions overlap
20:18 VanessaE noted.
20:19 VanessaE I guess because there's no way to do a VM read or write "in reverse"?
20:19 NathanS21 joined #minetest-hub
20:19 sfan5 there is I just didn't bother implementing it
20:21 VanessaE ok.
20:29 VanessaE holy crap, that's massively faster.
20:29 VanessaE (//stack)
20:30 VanessaE ~3 mins to do something that before was still going for an hour.
20:30 VanessaE nice :)
20:30 rubenwardy tumeninodes: I've added a status message to ContentDB so people don't get confused at least
20:31 rubenwardy no sure how to fix this
20:31 rubenwardy I've `su`'d into the redis user, and can write to the file fine
20:32 rubenwardy maybe redis doesn't like system links?
20:32 rubenwardy I have /var/log linked to /home/root-overflow/var/log   because OVH/Kimsufi is retarded
20:35 tumeninodes that will definitely be useful.
20:36 sfan5 why not let it log to /dev/null for now?
20:39 rubenwardy ah, thanks
20:39 rubenwardy that works
20:40 tumeninodes \o/
20:41 tumeninodes afk
22:19 VanessaE sfan5: bug in //allocate
22:19 VanessaE just do //1, //allocate <some small schem>.  the //1 position will move.
22:20 VanessaE repeat //allocate <file> and it'll keep moving.
22:22 sfan5 hmm
22:24 VanessaE and //load moves the coords as well
22:25 VanessaE https://i.imgur.com/e3By1Dr.png   I put pos1 right about where the crosshairs is in the lower right in the image.  //allocate created the selection box shown, //load put the structure where you see it
22:26 VanessaE s/is/are/
22:26 sfan5 it's not actually documented anywhere that allocate should "anchor" the schematic at pos1
22:27 VanessaE it always has in the past
22:27 VanessaE that's its purpose
22:27 sfan5 the fact that it does is because most schematics have (0,0,0) as lowest position
22:27 VanessaE it just was never clearly specified which corner of the structure it will be (SW I think)
22:28 VanessaE either way this is a bug, //allocate is not supposed to move pos1, and the schematic loaded outside of the box.  that has never happened before.
22:28 VanessaE (and I know the schem is good, I //save'd it a few hours ago)
22:29 sfan5 that is not supposed to happen
22:32 sfan5 looking at the code I don't see how it's possible that it saves such an "incorrect" schematic
22:33 VanessaE want a copy?
22:33 sfan5 sure
22:33 VanessaE https://daconcepts.com/vanessa/extra/thingy.we
22:36 VanessaE doesn't look like any coords in the file are at 0
22:36 VanessaE so //save must have b0rk3d it
22:37 VanessaE I was at commit f589c988 when I saved the fiule
22:37 VanessaE file*
22:37 VanessaE if that helpe
22:37 VanessaE if that helps
22:37 sfan5 hm yeah I can see the issue
22:37 sfan5 you saved a bunch of air, but since that air isn't saved //allocate detects a smaller area than the schematic was actually saved from
22:39 VanessaE yeah, it was just some floating structure, wanted to stash it before building that road (just in case I trashed something)
22:41 sfan5 more precisely "a bunch of air" means there are entire XY, XZ, YZ slices with nothing in them
22:43 VanessaE so maybe //save should add do-nothing "dummy" nodes at the four corners to keep //allocate et al. happy?
22:43 VanessaE eight corners*
22:45 sfan5 quite hacky, but I was thinking about doing that
22:47 VanessaE sometimes hacky works :)
22:57 sfan5 hmm no, if I add an air node there it would overwrite something when the schem is loaded
22:57 VanessaE not air
22:57 VanessaE a dummy node
22:57 VanessaE like ignore
22:57 VanessaE or something else that can't actually be placed.
22:58 sfan5 ignore works but throws a red warning at you on //load
22:59 sfan5 i'll just add a warning in chat for now
23:29 VanessaE "can't be fixed or worked around in a simple way."  dummy nodes :)  (insert on //save, explicitly ignore them on //load)
23:30 VanessaE (of course that would only work with a WE schematic, but then again I have no idea what happens if you try to place a MT schem that contains invalid nodes?
23:30 VanessaE )
23:50 * VanessaE notes Fixer logging into VE-S...
23:50 Fixer VanessaE: needed a test join just before going to sleep in a sec
23:50 VanessaE just making sure it's you :)
23:51 Fixer VanessaE: me, is that just me or fps is so much better?
23:51 Fixer and smoother
23:51 VanessaE *shrug*
23:51 VanessaE it IS a less complex spawn than the old place.
23:52 Fixer VanessaE: not just that, i've noticed it in my Just Test capture few minutes ago
23:52 Fixer VanessaE: i've noticed a lot of servers still on 0.4, wtf?
23:53 VanessaE no idea about that, aside from sheer inertia
23:53 VanessaE and "5.x, oh that's the unstable dev branch"
23:53 VanessaE and mobile clients not updating to the 5.x code bas
23:53 VanessaE base*
23:54 Fixer VanessaE: are you runnin latest git server?
23:55 VanessaE not *latest* latest, but it's recent 5.1.0-dev
23:55 VanessaE last updated a month or so ago, I suppose

| Channels | #minetest-hub index | Today | | Google Search | Plaintext