Time  Nick       Message
11:52 VanessaE   Can someone please investigate a potential connection between server sign-ons, namely when the media is being downloaded, and sharp drops in FPS for everyone who's already active on that server?
11:52 VanessaE   e.g. if user A signs on and starts playing, then user B signs on, user A's fps drops rapidly while user B is receiving media
11:55 VanessaE   oh that confirms it alright
11:55 VanessaE   user sign-ons cause a huge fps drop, even when they are not anywhere remotely neaby
11:56 VanessaE   nearby*
11:57 VanessaE   signing out and back in is sufficient to fix the fps loss for a while,
11:58 VanessaE   without walking around or doing anything to change how much or what kind of map data is loaded.
12:04 chlue      VanessaE: at how much fps are you playing? I never saw something but I only have 30fps
12:04 VanessaE   typically 30 to 35 fps but when someone signs on and starts playing, it drops to 15 or lower.
12:04 VanessaE   it reached 9 earlier before I restarted my client.
12:04 VanessaE   as a test, I'm removing stu's wield3d model
12:04 VanessaE   just in case that's the cause.
12:14 VanessaE   ok wield3d was not the cause, though it did make it a bit worse
12:19 VanessaE   ok that confirms it for ure
12:19 VanessaE   sure*
12:20 VanessaE   a player signing on causes everyone else's fps to drop out, sometimes requiring those clients to restart....which of course repeats the cycle for everyone else
12:21 VanessaE   I'm standing in one place, watching the performance graphs (f5) while players sign in/out
12:21 VanessaE   the graphs show a clear and unmistakable correlation.
12:31 chlue      VanessaE: which part of the graph? And you are not playing on the same system the server is running on? (Yes I have to ask this :-) )
12:49 zat        Devels... what is the procedure to get a feature I just incorporated to Minetest to be tried by someone?
12:49 zat        hopefully being able to commit it through git
12:54 chlue      Guess the typical way is to write about it in the forum and if its in a nice state make a pullrequest on github. But I am not a dev - I am only lurking here.
12:55 zat        l00rk
12:55 nore       the usual way is to make a pull on github, and then ask here to the devs
12:56 zat        (09:54:34 AM) Jordach: zat, create github account, fork minetest repo, git pull minetest repo, make changes, git push, pull request, ???, profit
12:56 zat        8D
12:57 chlue      the ??? part will definite contain patience, code changes, more patience,...
13:00 zat        well the code is very stable now
13:06 nore       what do you think of https://github.com/minetest/minetest/pull/688 ?
13:50 VanessaE   chlue: all of the graphs changed all at once, except for the network graph
13:50 VanessaE   it showed a brief spike but that was about it
13:50 VanessaE   I play on the same machine as the server runs on, but the fps dropouts are affecting *everyone* even when I am not playing.
13:51 VanessaE   the server runs as a separate process, it is not part of my client.
15:28 VanessaE   I am conducting a test on my survival world.
15:28 VanessaE   No mods whatsoever - not even default.  no entities whatsoever (/clearobjects followed by a restart)
15:29 VanessaE   no player models.  Nothing.  literally nothing but unknown nodes and air.
15:29 VanessaE   players are still seeing horrible fps loss.
15:29 VanessaE   what is left?  What have I missed>
15:34 celeron55_ what does the graph of received blocks show on each client?
15:34 VanessaE   the bottom graph?
15:34 VanessaE   it'
15:34 celeron55_ umm... top probably
15:34 VanessaE   "client_received_packet"... it's flowing quite fast at the moment
15:35 VanessaE   oh
15:35 celeron55_ at the point when a client connects
15:35 VanessaE   I'm not seeing any such graph.  closest is processed meshes, which is mostly idle, with spikes here and there.
15:35 VanessaE   oh, um
15:35 VanessaE   sec.
15:35 celeron55_ processed meshes is fine
15:35 celeron55_ that's what i mean anyway
15:36 VanessaE   no effect on processed meshes due to sign-ins
15:36 VanessaE   but I can't say I expect it since there's no game content right now :)
15:36 celeron55_ but still an effect on FPS?
15:36 VanessaE   lemme see what the users are saying
15:37 VanessaE   for me, farthest drawn is 119, 30 fps
15:38 VanessaE   (view range is not visible in F5 for me, because of the extra long default build/version string)
15:38 VanessaE   some users are reporting fps as low as 7 or 8
15:38 VanessaE   others around 50
15:39 celeron55_ ehm
15:40 celeron55_ so that is static fps?
15:40 VanessaE   nope.
15:40 VanessaE   it drops over time.
15:40 VanessaE   nore says he's getting 50 at a range of 240, but was getting 60 initially.
15:40 VanessaE   dan was getting 15 with an unknown range, now about 8 with a range of 10
15:41 celeron55_ ah now i get what is happening
15:41 nore       I now have 60 again... it depends on where I am
15:41 VanessaE   (well, "about 10" which probably means ~40)
15:41 VanessaE   celeron55_: opinions?
15:41 VanessaE   twoelk says 12 fps at a range of 95
15:42 VanessaE   he started at around 30 I think he said originally.
15:42 celeron55_ well here's a thing you can test: stand still until the client unloads all blocks, then turn around, stand still and see whether the fps behaves exactly like you would have just connected
15:42 VanessaE   sure
15:43 VanessaE   ok, relayed.
15:43 VanessaE   dan says now, 3 fps at 35m.
15:44 celeron55_ by default the client takes quite long until it unloads stuff though; the default is 10 minutes
15:44 VanessaE   celeron55_: related?  https://github.com/minetest/minetest/issues/184
15:45 celeron55_ can you take screenshots of all profiler pages right after connecting, and once the fps has gone as low as it gets?
15:45 VanessaE   sure
15:45 VanessaE   though for me, fps aren't affected as much at the momebnt
15:48 celeron55_ is it possible to point out what kind of computers get the largest impact? like how many cores, how fast cores and what kind of GPU?
15:49 nore       http://nore.mesecons.net/screenshot_2908334981.png
15:49 nore       http://nore.mesecons.net/screenshot_2908340834.png
15:50 nore       http://nore.mesecons.net/screenshot_2908344921.png
15:50 nore       ^ these are the screenshots
15:51 Dan_D      http://s7.postimg.org/ccxyuymu3/screenshot_2908357664.png, http://s7.postimg.org/qi3rwrvvf/screenshot_2908360926.png, http://s7.postimg.org/ydebbl5i3/screenshot_2908362858.png
15:51 Dan_D      The 3 F6 pages
15:51 Dan_D      Not sure of order
15:51 VanessaE   ok, and here come mine:
15:51 Calinou    only unknown nodes?
15:51 Dan_D      Yes
15:51 Dan_D      VanessaE removed EVERYTHING
15:51 VanessaE   http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f5.png  ~~  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f6_page1.png  ~~  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f6_page2.png  ~~  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f6_page3.png
15:51 Calinou    how can I contribute?
15:52 Calinou    :P
15:52 VanessaE   Calinou: to rule out any possibility of it being some weird mod bug.
15:52 Dan_D      The lag we mean
15:52 VanessaE   sign onto my survival server and take the screenshots as c55 mentioned.
15:52 nore       as for your information, I have an i7 processor at 2.3 GHz, and I have no FPS impact
15:52 VanessaE   Dan_D: not lag, unstable/low fps.
15:52 celeron55_ umm... so is this right after connecting?
15:52 Dan_D      FPS changes to 14 when turning to look at unloaded chunks
15:52 Dan_D      Yes it is
15:53 Dan_D      Are you the real Celeron55?
15:53 VanessaE   celeron55_: yes, mine are - within about 30 seconds (it was as fast as I could snap the screens)
15:53 Dan_D      Mine are within 10 seconds I would say
15:53 VanessaE   behind me, the map has not yet been loaded.
15:53 Dan_D      I looked at unloaded map chunks and got 14fps like I said
15:54 VanessaE   (max view is locked to 120 and max fps to 30 on my setup)
15:54 nore       mine are after about 2 minutes without moving
15:54 Dan_D      My max view is 35 nodes
15:54 Dan_D      and I get 1-2 fps with it
15:56 celeron55_ so let's get the same sample of data once it gets to maximum slowness
15:56 Dan_D      You mean with all mods?
15:56 Dan_D      And full view range
15:56 VanessaE   celeron55_: with or without all the usual mods in place?
15:56 celeron55_ i mean, when it gets as slow as it gets with that setup
15:56 VanessaE   ok, without any mods.
15:57 Dan_D      Without mods, full view range?
15:57 Dan_D      And give all 3 F6 pages?
15:57 celeron55_ i want to see the profiler output in two situations: 1) right after connecting, 2) after the client is as slow as it gets in the same place within that connection
15:57 Dan_D      Ok
15:57 VanessaE   ok.
15:57 Dan_D      I am still in one place after the last load of data
15:57 nore       for me, FPS have not changed since connection
15:57 Dan_D      I have full view ON
15:58 nore       so I don't know where the problem is
15:58 celeron55_ that should give an idea about what exactly changes most (and that is likely related to the FPS drop)
15:58 Dan_D      Actually, I have moved, but about 20 nodes
15:59 Dan_D      Away from the spawn
16:00 Dan_D      Full view on, I get: http://s15.postimg.org/4pykxnb8r/screenshot_2908983779.png, http://s15.postimg.org/3q8a1xw2z/screenshot_2908985873.png, http://s15.postimg.org/5grb3fdm3/screenshot_2908988525.png
16:00 celeron55_ the thing that matters in the second sample is that everything except connection time is equal to the first sample (to the extent that the same FPS would be expected), but similarly as VanessaE has explained, for some uknown reason it isn't
16:01 VanessaE   celeron55_: if you wanna see for yourself also, the test is at vanessae.mine.bz port 30001.
16:02 nore       and btw, I get errors while connecting to the server: 18:00:45: ERROR[main]: draw_hotbar(): mainlist == NULL
16:03 celeron55_ Dan_D: so, what seems to be unexpected in the second sample, in your opinion?
16:03 VanessaE   nore: that's probably the result of not having the default mod.
16:03 Dan_D      Such a huge fps drop
16:03 celeron55_ Dan_D: it looks like 10 times the amount of geometry is drawn
16:03 Dan_D      And the fact without  ful view, I omly get ~nodes viewv range
16:03 celeron55_ so that's quite a good reason for half the FPS 8)
16:04 Dan_D      ~10 nodes
16:04 celeron55_ ehm... with full view range, what did you expect?
16:04 Dan_D      With all the mods installed, every last one, 5fps was a good high number compared to average of 1-2fps
16:05 Dan_D      WIHOUT full iview range
16:05 VanessaE   aside:  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f5-2.png
16:05 celeron55_ can you take the second shot without full view range?
16:05 Dan_D      Yes
16:05 VanessaE   right after connecting, as soon as the graph was full I snapped the shot
16:05 Dan_D      I will wait a while
16:05 Dan_D      Then take the screenshot
16:06 celeron55_ VanessaE: looks like world loading
16:06 VanessaE   right, about as I expected.
16:07 VanessaE   shall I just stand in one place for a while or should I be walking around?
16:07 Dan_D      I am getting ~14 fps, normal view, wandering around
16:08 Dan_D      Hey, unknown blocks for signs still have an infotext
16:09 celeron55_ VanessaE: i don't know; it's really up to you to know how to get to the situation that you originally explained
16:09 VanessaE   yes, because that data is stored in the map.
16:09 Dan_D      As long as metadata is kept
16:09 VanessaE   celeron55_: well normally it is worst by running around, so yeah
16:10 Dan_D      I generally get a better fps when moving
16:10 Dan_D      When I stand still it slowly goesdown, over periods of 1-2 hours
16:11 celeron55_ probably something is considerably slow in discarding loaded blocks from being actually drawn
16:11 celeron55_ but i don't really know how to know what
16:11 celeron55_ and also i haven't yet seen data that would prove that something like it happens
16:11 Dan_D      WHAT THE HELL? THERE IS A HUUUUGE HOLE UNDER SPAWN!
16:11 celeron55_ i'm still waiting for such pair of profiler output that shows that such happens
16:13 Dan_D      I also get the bug where there are random triangular flashes of colour across the screen
16:13 celeron55_ Dan_D: what GPU?
16:14 VanessaE   celeron55_: I have a feeling it just isn't going to be as severe under these conditions as usual because hardly anyone is signing on/off - which seems to be the trigger
16:14 Dan_D      GPU?
16:14 Calinou    graphics processing unit
16:14 Dan_D      How do I found out? Onlinux Xubuntu?
16:14 celeron55_ VanessaE: well make the conditions *somehow*
16:14 Calinou    it is incorrect to ask "which GPU" by the way, you want to ask "which graphics card"
16:14 Calinou    lspci | grep VGA
16:14 celeron55_ VanessaE: i don't care what the conditions are :P
16:14 VanessaE   celeron55_: heh ok
16:14 VanessaE   I'll re-enable the usual set of mods then
16:14 celeron55_ or, well, i want to know what the conditions are, but i'm not the one who chooses what they are
16:15 Dan_D      00:02.0 VGA compatible controller: Intel Corporation N10 Family Integrated Graphics Controller
16:17 celeron55_ Dan_D: ok, so that's different from at least mine (950GM), which sometimes can do it too; much less likely to be a driver issue
16:17 Dan_D      Ok
16:17 VanessaE   the server has been returned to its normal state.
16:17 celeron55_ i don't expect anyone to be able to trace it out though
16:18 Dan_D      So, Are you The Real Celeron55?
16:18 celeron55_ personally i have only seen it on my laptop and not on any of the many other computers that i use
16:19 celeron55_ Dan_D: yes; with a small 'c'
16:19 Dan_D      Ok
16:19 celeron55_ oh, i apparently have some fallback nick
16:19 Dan_D      CeLeRoN55
16:19 celeron55  hi dAN_d
16:19 Dan_D      ThAt Is FiNe
16:21 Dan_D      9fps at spawn
16:21 celeron55  VanessaE: anyway once there is data to prove that it really happens and it's not just about world being loaded or people imagining things, then someone like hmmmm will have a completely irresistible urge to fix it 8-)
16:21 twoelk     as this seems to be time related somebody might record a video of the sign in sign out process
16:21 VanessaE   With the full standard game+mods used on my server, right after connecting:  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f6_page1_take2.png  ~~  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f6_page2_take2.png  ~~  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f6_page3_take2.png
16:21 VanessaE   I'll capture again after some time
16:22 celeron55  VanessaE: i suggest walking or flying around a lot within the unload timeout; then it could be worst
16:22 VanessaE   agreed
16:23 Dan_D      And I thought a copy of the lua pi came with minetest in /doc/lua_api.txt
16:23 Dan_D      But it seems not
16:24 celeron55  depends on who has packaged your copy :P
16:24 Dan_D      Linux nightly
16:25 Dan_D      See ya guys for now
16:27 VanessaE   celeron55: one thing that may interest you.  drawtime is supposed to directly correlate to fps, right?
16:27 VanessaE   the reciprocal of one being approximately equal to the other I mean
16:27 celeron55  fps = 1000/drawtime
16:27 VanessaE   right.
16:27 celeron55  also the profiler shows FPS
16:27 celeron55  umm actually
16:28 VanessaE   the interesting part:  it does not follow when the system starts to get slow.
16:28 celeron55  fps = 1000/(drawtime+sleeptime)
16:28 celeron55  also... actually that's wrong too
16:28 celeron55  i keep forgetting stuff 8)
16:28 VanessaE   for example, at a good time I might see ~35 or 40 fps with mid 20's draw time.  Another time, I might see mid 20's draw time, but half the fps.
16:29 celeron55  yes, i forgot that the client does other processing too
16:29 celeron55  so it maybe does something else in that case, and i have no idea what
16:29 VanessaE   it's ok, what matters is whether it should have a direct correlation e.g. a given value should always result in a given fps.
16:29 VanessaE   ok, fair enoug
16:32 celeron55  there's no perfect keeping track of what takes time in addition to drawing, but there might be enough to see what it might be doing
16:33 celeron55  (i want to be objective and wait for the actual data before guessing haphazardly and making everyone think it's what i guessed)
16:34 celeron55  (it would be totally useless)
16:36 VanessaE   ok, after running around a bit an waiting some, http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f6_page1_take3.png  ~~  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f6_page2_take3.png  ~~  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f6_page3_take3.png
16:36 VanessaE   obviously not as severe as usual but there's very little sign on/off activity right now
16:37 celeron55  so did you see the degradation when playing?
16:37 VanessaE   oh wait, here we go
16:38 VanessaE   lemme take another set, it's just dropped to ~11 fps
16:39 twoelk     how much influenze can the client have on fps?
16:39 twoelk     as in different builds different pc
16:40 VanessaE   http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f6_page1_take4.png  ~~  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f6_page2_take4.png.png  ~~  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/f6_page3_take4.png
16:40 VanessaE   that's just the result of standing at the spawn for too long.
16:41 VanessaE   as I sit here doing nothing at all, fps has continued to drop, now 7.4 fps
16:43 VanessaE   twoelk: at least some effect, sure.  different compiler profile, different CPU, different libraries...
16:44 VanessaE   looking around seems to be causing my fps to rise back up.  now 30.
16:45 celeron55  VanessaE: the accidental middleway sample is actually very useful 8)
16:45 VanessaE   but the view range is still only 29.
16:45 VanessaE   oh?
16:45 celeron55  it shows what can change without actually affecting FPS much
16:49 celeron55  nope... i can't find any sane reason from this
16:50 VanessaE   exactly.  the only correlation I've been able to find is that someone logging in causes everyone else's fps to drop.
16:51 celeron55  it's definitely doing "something"
16:51 celeron55  something other than drawing
16:51 twoelk     I am running the game on a PIII 2,5 GHz with some  1.7gb of RAM under WinXP now with usually under 10fps
16:51 VanessaE   right, the big question is what :)
16:53 celeron55  twoelk: that's generally too old to run minetest, and we are discussing a very particular performance problem
16:53 twoelk     at the office I run it on a IntelCore i7-2600k-64bit, with 4 cores, 3,4 Ghz and 8 gb Ram with usually over 20fps
16:53 twoelk     smae client on a portable usb hdd
16:53 celeron55  twoelk: that's offtopic
16:53 twoelk     *same
16:54 celeron55  VanessaE: do you know a past version of minetest that does not have this behavior?
16:54 twoelk     just saying my fps depends on the pc to some point
16:55 twoelk     so comparing is difficult
16:55 VanessaE   celeron55: I'm afraid I couldn't be sure.  #184 is the closest mark I could give you, but it's been prettu acute over the past several weeks
16:55 VanessaE   pretty*
16:55 VanessaE   (#184 is oer a year old though)
16:56 celeron55  the description in it is so convoluted that i'm too lazy to try to figure out what that series of actions is about :P
16:56 VanessaE   back then, I used to lock my view range at a certain distance and just let my fps auto-adjust.
16:57 VanessaE   the gist of that particular issue is that more loaded map = lower fps, even when the extra map data is not visible.
16:57 VanessaE   i.e. the extra data is out of range, or is behind you, etc.
16:58 celeron55  some of that is completely expected and not related to this
16:58 VanessaE   right
16:59 VanessaE   fps has dropped to 8.3 while the client was hidden behind my browser.  there are two players visible - one right in front of me and one off in the distance.
16:59 celeron55  what does the graphs show?
17:00 VanessaE   http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2010122013%20-%2001:06:02%20PM.png
17:01 VanessaE   the server is currently using about 12% CPU on its core, the client is maxed out (as expected)
17:01 VanessaE   just fyi, not that it's related ;)
17:01 celeron55  oh that really shows how much "other" takes time
17:01 VanessaE   yes
17:01 VanessaE   what's in "other"?
17:01 celeron55  everything but drawing :P
17:02 VanessaE   heh
17:03 celeron55  you can see the shape of it also, but the shape doesn't really tell much
17:03 VanessaE   *nod*
17:04 VanessaE   for scale, those mesh spikes are about 3 seconds apart.
17:04 celeron55  it's probably some oscillating water or something
17:04 celeron55  or some mod doing something every second
17:05 VanessaE   *shrug* might be a blinky plant somewhere, but that's like 1 mapblock, so probably no real concern there
17:05 celeron55  no, it's like 20 mapblocks as you can see
17:05 celeron55  8)
17:05 VanessaE   heh
17:05 VanessaE   ok fine, 20 then :P
17:05 VanessaE   didn't notice the value next to the graph :)
17:07 celeron55  do you run your client with verbose logging enabled?
17:07 celeron55  i really don't know whether something about this might end up in log, but it's probably worth a try
17:07 VanessaE   yes, it's at log level 4.
17:08 celeron55  or... actually, a simpler way could be to add like 5 more timers in the main loop to output to the profiler
17:08 VanessaE   fps has, btw, dropped to 4.4 :D
17:08 celeron55  let it drop more
17:08 VanessaE   ok
17:08 celeron55  does memory consumption do anything odd?
17:08 VanessaE   nope
17:09 VanessaE   3.7 GB with HDX 256px.  This is pretty normal for my box.
17:10 celeron55  oh by the way, does it consume 100% of one core?
17:10 VanessaE   indeed so
17:11 celeron55  okay, so it isn't waiting for i/o (had to clear that out)
17:11 VanessaE   http://pastebin.ubuntu.com/6227584/
17:12 zat        who wants to test the new rollback?
17:19 VanessaE   fwiw:  if I open a second client, fps is quite normal there, even while the first one is spinning its wheels.
17:20 VanessaE   (27 fps at 84 view range on complex landscape)
17:24 celeron55  VanessaE: is there absolutely nothing you can do on the slow client that would speed it up at all?
17:25 celeron55  oh by the way, what's the FPS now? 8)
17:27 VanessaE   FPS has actually come up a bit, 9.9
17:27 VanessaE   and no, I've tried everything short of turning off shaders and my texture pack, but as these weren't a problem in the past....
17:28 VanessaE    9.0
17:28 VanessaE   8.5...
17:28 VanessaE   it's all over the map0
17:28 VanessaE   -0
17:28 VanessaE   so to speak :)
17:28 VanessaE   (what an asshole ;) )
17:29 VanessaE   ( obscure Airplane! reference :) )
17:30 celeron55  i mean, anything without closing it
17:30 celeron55  i guess no then
17:30 VanessaE   well
17:31 VanessaE   there's one thing that helps some folks, lemme try
17:31 celeron55  what is the thing?
17:32 VanessaE   I teleported away from where I was, to a point about 500m away, well out of reach of my max view range and out of range of the loadable map range
17:32 VanessaE   fps jumped to 22, view range is still stuck at about 29.
17:33 VanessaE   range, range, range, range, FPS! FPS!, ra....  ok I'll stop now.
17:34 VanessaE   didn't last lonh
17:34 VanessaE   fps already dropped again, about 6.2 now
17:35 VanessaE   memory usage is up a bit, 4.1G but I have 16 and use a 64-bit setup, so that's a non-issue.
17:36 celeron55  okay, i'm pretty sure that instance won't give us any more information
17:36 VanessaE   ok
17:36 celeron55  i'm not sure what to do with logs
17:36 VanessaE   I was just about to ask :P
17:36 celeron55  it might be useful to have a verbose log sample (a 1000 lines) from when it works like it's supposed to, and when it doesn't
17:37 celeron55  there's probably nothing to be seen but who knows
17:37 VanessaE   preparing such now.
17:37 celeron55  so just tail the log now and restart the client and tail it then
17:42 VanessaE   http://pastebin.com/tbPrxbXc  ~~  http://pastebin.com/YNdQhyeC
17:45 celeron55  yeah, nothing to be seen
17:46 VanessaE   it was worth a shot
17:47 celeron55  the only option left is adding more and more ScopeProfilers everywhere until one of them shows what exactly is taking all the CPU
17:51 VanessaE   which could turn into a big mess
17:51 celeron55  well, it's really trivial 8)
17:51 celeron55  (of course the resulting code won't be committed anywhere)
17:52 chlue      celeron55: this scope profilers can simply be dynamically added by writing a value with a different name?
17:52 celeron55  chlue: yes
17:59 celeron55  zat: where's it available?
18:05 zat        in my computer uh lol
18:06 zat        the code is not tidied yet, celeron55, but it works so far.
18:06 zat        it automatically migrates an old rollback.txt file to the new sqlite one.
18:14 zat        (should I clone the minetest repo in my own git server then commit it there or just initialize my empty repository with my commit?)
18:19 chlue      If you ultimate want to be able to make a pull request you should have a common history --> start with a clone.
18:21 zat        well I dont know how pull requests work exactly so I am doing as you suggest then, thx.
18:25 celeron55  well umm yeah, you should somehow commit your stuff on top of a fork of the main minetest repo
18:26 celeron55  that way your commit is easy to merge work on otherwise
18:26 celeron55  (usually people do that on github but that's not a requirement)
18:26 celeron55  s/merge work/merge and work on/
18:27 celeron55  pastebin the diff or something so that people can look at it
18:27 zat        what is better for you, fork or diff?
18:28 celeron55  uhm... well you're going to make diffs even if you have a fork
18:28 zat        yes lol
18:28 zat        make
18:28 zat        oops sosrry
18:28 VanessaE   bbl
18:28 celeron55  but a plain diff works too... but you probably want version control yourself too so i don't know why you'd want to just make diffs
18:29 celeron55  just make sure to publish a diff before you mess up something and lose it altogether 8)
18:30 zat        really... actually... I almost killed it all for not looking at what cwd I was going to do a checkout -f at :)
18:32 chlue      zat: in a nutshell, If you start with a clone in git then the tool can for example reliable assist you when upstream moves forward because it know how to 3-way merge stuff. If you start with an anonymous copy of the same codebase then git can only guess based on content.
18:32 zat        its that I am just tired of having to make an account for almost anything
18:33 chlue      as long a you keep your stuff local thats not needed. Just do git clone "path shown on github" and you have your local copy.
19:28 zat        Done: you can clone the test implementation from: git://bros.ivn.cl/mbarrera/minetest.git
19:55 celeron55  zat: doesn't work
20:00 zat        Sorry celeron55, I had to disable it for a while due t a bug, you can clone now.
20:47 celeron55  zat: you managed to get something terribly wrong in this
20:48 celeron55  most of your changes are in a "Translated using Weblate (Spanish)" commit
20:50 celeron55  here's the relevant diff for anyone who wants to take a quick look: https://gist.github.com/celeron55/6954775
21:26 zat        celeron55: I just applied the changes to the last commit avaiable, I did not know which one I should work on.
21:26 zat        However... it is just one file, src/rollback.cpp
22:34 zat        celeron55: ??