Minetest logo

IRC log for #minetest-dev, 2015-10-14

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

All times shown according to UTC.

Time Nick Message
00:13 hmmmm in any case I'm looking for people to do a second look at that specific chunk
00:13 hmmmm working with the code made me understand all of the changes and the rationale for them a lot more than just looking at the diff in a web browser
00:13 hmmmm but still I might've missed something
00:16 younishd joined #minetest-dev
00:28 eugd left #minetest-dev
00:48 Player_2 joined #minetest-dev
01:01 Miner_48er joined #minetest-dev
02:43 paramat joined #minetest-dev
02:45 paramat https://github.com/minetest/minetest/pull/3248 "ABMs: Make catch-up behaviour optional" by paramat, is tested and ready for review if anyone is able to
03:40 AnotherBrick joined #minetest-dev
03:42 waressearcher2 joined #minetest-dev
05:47 Hunterz joined #minetest-dev
06:08 OldCoder date time ERROR[main}: Client::handleCommand_ActiveObjectMessages: caught SerializationError:deSerializeString: couldnt read all chars
06:09 OldCoder hmmmm, kahrl, sfan5 highlighting because the 3 of you might be awake and the person affected deserves help...
06:09 hmmmm yes, it's a pretty standard error
06:09 OldCoder and that is a known core issue. Can any of you comment on the nature of numerous messages of this type and what to do?
06:09 OldCoder
06:10 hmmmm YOU can't do anything
06:10 OldCoder Right. Is the cause known?
06:10 hmmmm it's low priority because frankly it doesn't seem like it affects anybody outside of vanessae and it doesn't actually impact gameplay
06:10 hmmmm no, it's not.
06:10 OldCoder It is affecting me ATM. Apparently 100s of these. But if not known, not known.
06:10 OldCoder Can I simply comment out the message?
06:11 OldCoder The player affected, if it matters, is a 49 year old heart attack patient
06:11 OldCoder This is his therapy
06:11 hmmmm uhh... lol
06:11 OldCoder lol to comment out?
06:11 OldCoder Or therapy? MT is excellent for patients
06:11 OldCoder Socialization, creative outlet, pleasant hobby
06:12 hmmmm anyway if you're getting these errors frequently enough that it's actually causing an impact on gameplay then we should talk
06:12 OldCoder O.K. I will restart. What type of logging could I patch into the core if the problem resumes?
06:12 hmmmm on vanessae's server this only happened in momentary spurts around very specific areas for a short amount of time
06:12 OldCoder Happy to add printf
06:12 OldCoder Useful info
06:12 OldCoder Specific areas?
06:12 hmmmm oh no, you don't want to do that
06:12 hmmmm I tried that already
06:12 OldCoder printfs affect the problem?
06:13 hmmmm the printfs I'd like to add would effectively freeze your server
06:13 OldCoder They are quite volume oriented?
06:13 OldCoder 1M's ?
06:13 OldCoder Heh
06:13 OldCoder MY server is pretty fast
06:13 OldCoder Does that help?
06:13 hmmmm if this is as easily replicatable as you make it sound then I would just comment out that log message for now
06:13 kaeza joined #minetest-dev
06:13 hmmmm i'm sorta busy atm but i would like to solve it as it has been an ongoing problem since 0.4.10
06:14 OldCoder I must find out if it is replicable. It may be specific area. I will report back in 2 days if the problem continues
06:14 OldCoder Is this satisfactory?
06:14 hmmmm OK.
06:14 OldCoder I will restart and ask player to note patterns. We will then see if random or replicable.
06:14 OldCoder Thank you.
06:15 hmmmm I tried capturing this error on vanessae's server before
06:16 hmmmm because of the sheer volume of data it was near impossible to find anything of significance
06:16 hmmmm one thing i did learn is how incredibly inefficient AOMs are
06:17 hmmmm if you have a pipeworks mod enabled with many stacks of nodes or items, it's possible your server could be slowing down to a halt because of all the constant movement messages being sent
06:17 hmmmm slowing to a crawl* rather
06:18 nyje joined #minetest-dev
06:18 OldCoder Yes
06:18 OldCoder It's pipeworks, is it? hmmmm But thought I saw it on a regular world.
06:19 hmmmm unlikely
06:19 OldCoder I've turned off half of pipeworks on that world
06:19 OldCoder Ah
06:19 hmmmm it's always possible though, but high volume seems to perturb the condition
06:23 hmmmm ShadowNinja:  I'm about to push these three commits to upstream/master: https://github.com/kwolekr/minetest/commits/sncleanup
06:23 hmmmm more to come soon
06:23 OldCoder hmmmm, I'll be asleep in moments, but player reports issue is continuous and returns after restart. It is just his area.
06:24 hmmmm yea, I'm not going to be able to get to this tonight at all.
06:24 OldCoder Conditions may be ideal to debug the problem
06:24 hmmmm sorry.
06:24 OldCoder No sorry, just reporting
06:24 OldCoder Busy for a day or two myself. Noting that conditions may be ideal to debug the problem.
06:24 OldCoder Will comment in a day or two
06:24 OldCoder Zzz
06:24 nrzkt joined #minetest-dev
06:26 ShadowNinja hmmmm: Weren't you going to make the none->always change?
06:26 hmmmm yeah that can come later
06:27 hmmmm I think I already messed with your commit enough as-is
07:05 julienrat joined #minetest-dev
07:17 julienrat left #minetest-dev
07:28 est31 joined #minetest-dev
07:29 est31 note that parts of the inefficiency comes from the way SAOs get used in pipeworks
07:29 est31 way too much, even used if nobody is there
07:29 est31 and the other part comes from how the server handles it
07:30 est31 AFAIK VE's servers are set up to transmit SAO messages to the whole server
07:30 est31 even if the player is at the other end of the world
07:30 est31 but I might be wrong with this
07:32 est31 so I dont think the problem is the protocol, its more the implementation
07:32 est31 but great work you did hmmmm btw on the logging pr
07:33 est31 finally it got merged
07:35 nrzkt2 est31: yes messages are transmitter to the whole server, like detached inventories in unified inventory, 3d_armor and more... evenif the inventory affect only 1 user
07:37 hmmmm oh it's not done yet
07:37 hmmmm there are literally 5 more commits to go, but these are barely related to logging at all
07:49 Darcidride joined #minetest-dev
07:52 est31 hrmmm why is this code even there the first place
07:52 est31 https://github.com/minetest/minetest/pull/3248/files#diff-baa4d2d0ffafdec2b86baf065fba252bR587
07:53 hmmmm so it doesn't divide by zero the line below it
07:54 hmmmm or at least that's the rationale
07:54 est31 oh, thats an argument
07:54 est31 perhaps it should be commented
07:54 hmmmm in floating point math, that wouldn't cause an exception, the result would just be infinite
07:54 est31 yeah
07:55 hmmmm the person who wrote this either forgot to remove that when it was changed to a float from an int or didn't understand floating point math
07:55 thePalindrome joined #minetest-dev
07:55 est31 my initial thought was that intervals is an int, and you jump if the actual_interval is smaller than trigger_interval
07:55 est31 but then i saw "float"
07:55 thePalindrome left #minetest-dev
07:57 est31 the general problem I have with the PR is that instead of fixing a problematic feature we just offer a way to disable it
07:57 est31 it should be made real catch up
07:58 hmmmm I don't have a problem with it
07:58 hmmmm notice how it says "simple" catch up
07:58 hmmmm the implication being that there will be a more advanced version in the future
07:58 est31 this is the lua api
07:59 est31 we have to support it for a long time
07:59 est31 and not remove stuff without a good reason
07:59 hmmmm then maybe catchup should be an enumeration instead of a boolean?
08:00 est31 there is no real fast way to do enums in lua
08:00 hmmmm catchup = "none", "simple", "complete", etc.
08:00 est31 thats good old slow strings
08:00 hmmmm "catchup", "mustard", "mayonaise"
08:00 est31 heh
08:00 julienrat joined #minetest-dev
08:00 julienrat left #minetest-dev
08:01 hmmmm I hate it too trust me
08:01 est31 well, its time spent during registering
08:01 hmmmm at one point I was going to replace all the flags and enums with constants defined in a builtin file
08:01 est31 it can be converted to enum on c++ side
08:01 est31 during parsing
08:01 hmmmm it is
08:01 hmmmm that's how literally all other lua enumerations work :)
08:01 hmmmm getenumfield()
08:04 est31 so yeah i like the enum idea
08:04 est31 generally I think there should be several different kinds of ABMs
08:04 est31 ones which get called upon world load
08:05 est31 and only then
08:05 est31 they would be good for replacing etc
08:05 est31 then ones which get called on newly placed blocks
08:06 est31 and those which get called regularly like we have now
08:06 est31 for the newly placed blocks we can have the lavacooling abm
08:06 est31 (newly placed around)
08:07 est31 but that use case is done much better once abms implement what I propose in #3003
08:08 ShadowBot est31: Error: Delimiter not found in "An error has occurred and has been logged. Please contact this bot's administrator for more information."
08:08 est31 https://github.com/minetest/minetest/issues/3003
08:08 est31 "Do leafdecay scanning in the engine"
08:09 deltib joined #minetest-dev
08:11 est31 man, this sat around for 2 months now
08:11 est31 and I havent even started
08:12 * est31 lazy
08:14 alket joined #minetest-dev
08:18 kaeza2 joined #minetest-dev
08:26 est31 hrmm, any ideas, I want to rephrase the channel description of #minetest-dev on http://dev.minetest.net/IRC
08:27 est31 so that it includes something like "Development discussion of the core projects like Engine or Minetest Game"
08:27 est31 but that's too long
08:28 est31 (bc its not just about engine here, paramat always announces his game prs here for example)
08:28 julienrat joined #minetest-dev
08:28 nrzkt2 remove like and tell Minetest team projects
08:35 nyje joined #minetest-dev
08:58 celeron55_ joined #minetest-dev
09:22 Megaf joined #minetest-dev
09:26 celeron55 i guess there should be an easy option to just turn off the visible error log on the client
09:26 celeron55 i mean, most users can't do anything about errors anyway and they still want to see the chat
09:27 celeron55 (sometimes it just becomes unnecessarily difficult to debug a non-fatal error in a reasonaable amount of time)
09:35 julienrat joined #minetest-dev
09:54 crazyR joined #minetest-dev
10:28 julienrat joined #minetest-dev
10:48 proller joined #minetest-dev
10:56 Calinou joined #minetest-dev
11:34 Ritchie joined #minetest-dev
11:55 Taoki joined #minetest-dev
12:00 Ritchie joined #minetest-dev
12:03 proller joined #minetest-dev
12:21 Darcidride_ joined #minetest-dev
12:23 misprint joined #minetest-dev
12:35 rubenwardy joined #minetest-dev
12:39 proller joined #minetest-dev
12:48 Ritchie joined #minetest-dev
12:57 Player_2 joined #minetest-dev
13:18 julienrat joined #minetest-dev
13:24 est31 joined #minetest-dev
13:25 Darcidride__ joined #minetest-dev
13:25 est31 agree celeron55.
13:25 est31 Those who get the errors can't really do anything with them. Other applications usually have error reporting for this / metrics etc
13:26 nrzkt2 +1
13:26 est31 (i dont know whether thats a good idea for us however, I'm neutral here)
13:29 Darcidride joined #minetest-dev
13:33 waressearcher2 if someone takes source code of minetest and rewrites it in C, can it be considered a fork and it have to be also under same license or it would be a completely different source code and minetest's license do not apply anymore ?
13:33 VanessaE the license still applies.
13:34 waressearcher2 if its rewritten in any language the license is still applies ? why is it applies because what exactly is the same ? algorithms or variable names ?
13:35 VanessaE because you'd have started with the original source code, and being C++, I'm sure some of it would still work.
13:35 VanessaE now were you to read the sources and describe their operation to someone else who has never seen the source, and that person rewrites it, then it wouldn't be a fork
13:36 VanessaE "clean room implementation"
13:36 VanessaE works for closed-source software too in some cases.
13:36 julienrat joined #minetest-dev
13:36 julienrat left #minetest-dev
13:36 waressearcher2 what if persone never seen the source but only bunch of videos of gameplay and he write the clone that looks and feel exactly as in video, is that a "clean room implementation" ?
13:36 rubenwardy It does depend on what you mean by rewrite
13:37 rubenwardy also, this isn't dev-like chatter
13:38 eugd joined #minetest-dev
13:41 VanessaE waressearcher2:  yes, as long as the person writing the code has never seen the original sources
13:41 VanessaE why do you ask, anyways?
13:42 VanessaE ruben's right.  --> #minetest
13:44 julienrat joined #minetest-dev
13:48 julienrat left #minetest-dev
14:43 julienrat joined #minetest-dev
14:53 julienrat left #minetest-dev
15:07 younishd joined #minetest-dev
15:33 CraigyDavi joined #minetest-dev
15:34 zat joined #minetest-dev
15:34 nrzkt joined #minetest-dev
15:42 eugd joined #minetest-dev
15:50 Amaz joined #minetest-dev
15:58 proller joined #minetest-dev
16:00 hmmmm joined #minetest-dev
16:08 proller joined #minetest-dev
16:09 H-H-H joined #minetest-dev
16:10 Hunterz joined #minetest-dev
16:20 nyje joined #minetest-dev
16:26 julienrat joined #minetest-dev
16:34 jin_xi joined #minetest-dev
17:08 loggingbot_ joined #minetest-dev
17:08 Topic for #minetest-dev is now Minetest core development and maintenance. Last release: 0.4.13, Aug 20 2015. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.ru/minetest-dev/ http://dev.minetest.net/
17:10 OldCoder joined #minetest-dev
17:10 Kray joined #minetest-dev
17:10 Warr1024 joined #minetest-dev
17:12 Sokomine joined #minetest-dev
17:14 Soni joined #minetest-dev
17:22 Darcidride joined #minetest-dev
17:24 eugd can #3199 get merged? i need to reset my fork.
17:24 ShadowBot eugd: Error: Delimiter not found in "An error has occurred and has been logged. Please contact this bot's administrator for more information."
17:25 eugd https://github.com/minetest/minetest/pull/3199
17:28 Krock joined #minetest-dev
17:44 eugd joined #minetest-dev
18:56 ElectronLibre joined #minetest-dev
19:25 alket joined #minetest-dev
19:56 hmmmm I'm not sure what's holding it up
19:57 hmmmm I'll take a look at that PR and do a quick re-review
20:05 eugd joined #minetest-dev
20:30 alket joined #minetest-dev
20:49 DFeniks joined #minetest-dev
21:21 proller joined #minetest-dev
21:26 ElectronLibre joined #minetest-dev
21:32 alket joined #minetest-dev
21:48 proller joined #minetest-dev
21:57 misprint joined #minetest-dev
23:23 Fritigern joined #minetest-dev

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