Time Nick Message 01:25 paramat ~tell Zeno` https://github.com/minetest/minetest/pull/5625 01:25 ShadowBot paramat: O.K. 01:25 paramat #5625 01:25 ShadowBot https://github.com/minetest/minetest/issues/5625 -- Mgflat, Mgv7: Only create noise objects if needed by paramat 02:39 octacian #5626 02:39 ShadowBot https://github.com/minetest/minetest/issues/5626 -- Menu: Add install mod button by octacian 02:39 octacian paramat, sofar: ^ 02:39 octacian also, Calinou ^^ 02:43 paramat ok 03:15 Hijiri Does the attached entity visual position bug still exist 03:32 OldCoder minetest.org is being DDoSed right now, Perrtu. I am the target. No other machine in the host's network is presently under attack. 03:32 OldCoder Tell the trolls to consider their actions carefully. 03:32 OldCoder There are 117,000 sessions. This is the work of somebody in power; one of the core devs or their friends. 03:33 OldCoder nore, sofar, shivajiva, others; my host has good hardware. And I do prosecute. 03:33 OldCoder I've been through this before in other contexts. 03:34 sofar everyone, please do not respond 03:34 sofar I advise you to not reply to oldcoder in any way on these accusations 03:34 OldCoder sofar, I'll hire an attorney and subpoena 03:34 OldCoder RedHat 03:35 OldCoder based on your involvement. Think it's a bluff? 03:35 OldCoder They are not "accusations". Wish to speak with my host right now? 03:35 sofar I always take legal accusations seriously 03:35 OldCoder You should 03:35 Zeno` Why would a core dev have that many shells? 03:35 OldCoder And I am being DDoSed 03:35 sofar Zeno`: please, do not respond 03:35 OldCoder hang on, I'm getting more details 03:35 OldCoder sofar, hush 03:36 OldCoder sofar, justify your remarks if you are not involved 03:36 OldCoder But excuse me, host is updating his own scans 03:36 OldCoder We may be able to trace the network if it is not HMA 03:36 OldCoder This (a) is childish (b) will not hurt me (c) is attracting my attention 03:37 octacian #5627 03:37 ShadowBot https://github.com/minetest/minetest/issues/5627 -- Menu: Merge singleplayer and server tabs by octacian 03:37 sofar OldCoder: I have no obligation to justify any remarks 03:38 OldCoder It was a sensible point. And here is a sensible link for you to review: 03:38 OldCoder Moment, host is vetting it 03:39 OldCoder Host offers correction. It is 17,000, not 117,000. That is a lot of simultaneous connections, though. 03:40 OldCoder Fortunately, I am not hosting from a living room floor. My host is the equal of anybody you know. Anyway, here is the link: 03:40 OldCoder http://minetest.org/geethanks.txt 03:40 OldCoder Are those IPV4s known to anybody present? Perhaps they are RedHat IPV4s, sofar ? 03:41 OldCoder Note that 1,000s of connections are originating *simultaneously* from each of them 03:41 sofar https://www.maxmind.com/en/geoip-demo try it 03:41 Hijiri does redhat mean something other than the distro 03:41 OldCoder I had not really noticed, sorry about that, it is like a sneeze. I know how... 03:41 OldCoder Hijiri, the company 03:41 OldCoder sofar, I do know how to geolocate 03:41 Hijiri oh yeah 03:41 sofar I don't think redhat has AT&T U-Verse blocks 03:41 OldCoder I have been fighting DDoSers in my family for 4 years... My brother is the VP of Lab126 and uses Amazon for this purpose. 03:42 OldCoder sofar, I knew it wasn't you. I found your remarks to be out of line. 03:42 sofar obviously whoever is doing this is using a botnet 03:42 OldCoder AT&T is a surprise, though... 03:42 OldCoder Well, yes 03:42 OldCoder So, I won't be able to track the IPV4s 03:42 OldCoder But the originators will recognize them, correct? 03:42 OldCoder U-Verse is residential. I'd expected this to be Russia. 03:43 OldCoder Talking to host about that, excuse me... 03:43 sofar maybe someone has backdoored the minebest engine... 03:43 OldCoder No 03:43 OldCoder It was a "booter" server 03:43 OldCoder It was a "booter" service * Observe this: 03:43 OldCoder http://i.imgur.com/QvLWoJN.png 03:44 OldCoder A senior Minetester paid for 4 hours 03:44 OldCoder A 4-hour botnet slot 03:44 OldCoder My host is very, very good 03:44 OldCoder But he won't be able to identify the individual 03:44 OldCoder I'll offer concluding comments shortly, talking to him again 03:44 * OldCoder didn't know that these things came in specific timeslots 03:46 OldCoder Host is digging into this a bit. sofar, do you still wish to use the word "accusations" as though it was nonsense? 03:46 OldCoder Or to contest that somebody here, or who will be here tomorrow, did not purchase this? 03:47 OldCoder Somebody is PM-g #minetest-project'ers and telling them that the channel is "shutting down" 03:47 OldCoder And somebody has purchased a 4-hour professional botnet. Just as a gift for the Old Coder. 03:47 OldCoder How sweet. 03:47 Hijiri I didn't receive any PM 03:47 OldCoder I am in fact kind. I have in fact spent 5 years trying to help here - and succeeding. 03:47 OldCoder Hijiri, sorry 03:48 sofar OldCoder: I'm seeing you make accusations with legal ramifications 03:48 OldCoder I'll try not to speak in absolutes. Somebody is PM'g #minetest-project'ers - one at a time 03:48 OldCoder Not en masse 03:48 OldCoder And somebody has hired a botnet just for me. 03:48 OldCoder sofar, yes; and, respectfully, what is your point? Do you deny the accusations on the part of the core devs? 03:48 OldCoder You are no child 03:49 OldCoder This is no sandbox game 03:49 sofar it appears someone is DDossing you, this I can conclude as well, permitted any of the information you provided in here is accurate 03:49 OldCoder Do you contend, or even believe, that minetest.org is not the specific target? 03:49 sofar this is largely offtopic for this channel 03:50 OldCoder ... 03:50 OldCoder It is not unless you contend 03:50 OldCoder this did not originate right here 03:50 OldCoder Do you so contend? That *will* place it in your responsibility 03:50 OldCoder And *will* offer legal openings 03:51 OldCoder Host says DDoS is restarting 03:51 Hijiri well it's not directly related to development at least 03:51 OldCoder Excuse me briefly, checking details 03:51 OldCoder Hijiri, it is directed at me personally 03:51 OldCoder By this group 03:51 OldCoder I may need to respond 03:51 sofar you go and make your accusations if you so desire 03:52 OldCoder sofar, thank you for your civility 03:52 OldCoder Excuse me briefly, checking details 03:52 OldCoder sofar, just out of curiosity, do you offer an alternative explanation? 03:52 OldCoder Taking into account 03:52 OldCoder That somebody here has asked people 03:52 OldCoder To PM people in -project to the effect that -project is "shutting down" (exact quote) ? 03:52 OldCoder I will have the name of THAT person 03:52 OldCoder in 24 to 48 hours. 03:53 sofar you're misrepresenting statements by people, regularly 03:53 OldCoder Odd, that such PMs would be made even as a DDoS is going on, no? 03:53 OldCoder Indicate misrepresentation, please 03:53 sofar I don't have to 03:54 OldCoder Then withhold your assessment 03:54 OldCoder It is unreasonable to make it and to decline to discuss it 03:54 OldCoder My host is assessing router load 03:54 OldCoder If it is going to cost him anything significant, legal actions will be unavoidable 03:55 OldCoder Other actions as well 03:55 OldCoder It is suggested, politely, that those who are proceeding in this matter 03:55 OldCoder Be advised of the preceding points 03:55 OldCoder sofar, is anything that was just said unreasonable? 03:55 sofar yes, you're threatening people with legal action 03:55 sofar this is immature 03:56 sofar unethical as well 03:56 OldCoder Immature? This may cost my host cash and is a felony violation of CFAA 03:56 sofar I understand you are upset 03:56 OldCoder Hardly 03:56 OldCoder I am starting 03:56 OldCoder to pay attention 03:56 OldCoder Respectfully, people most likely 03:56 OldCoder do not wish my full attention 03:56 sofar but your tone and attitude are unreasonable and outside of the guidelines that have been set out for this channel 03:56 OldCoder Justify your remarks, including that point 03:56 sofar take your legal accusations elsewhere 03:56 sofar please leave this channel 03:56 OldCoder Kick me and I contact RedHat tomorrow 03:56 OldCoder Go ahead. I promise. 03:57 sofar you are mature enough to remove yourself 03:57 OldCoder I am being attacked right now. And my host is a friend. 03:57 OldCoder Excuse me while I check the resource load with him. 03:57 OldCoder sofar, I will phone senior RedHat people at home if you involve yourself. 03:57 OldCoder And provide 03:57 OldCoder Information you may not be aware that I possess 03:57 OldCoder My friend needs me 03:58 OldCoder Excuse me a moment 03:59 OldCoder Waiting on his analysis. sofar, DDoS is a felony violation of CFAA. CFAA has teeth if one knows how to call the teeth. 03:59 OldCoder Do not threaten *me* 03:59 sofar I have made no threats of any form. 04:00 OldCoder You were preparing the framework to state that I was out of line to be present in this channel 04:00 OldCoder This channel that is involved in one of the more prosecutable felonies 04:00 OldCoder In the tech arena 04:00 sofar your current behavior is neither conductive to the project nor friendly 04:01 sofar it is by all means unwanted 04:01 OldCoder I am being DDoSed. By these people. Did I happen to mention that? 04:01 OldCoder That is by all means "unwanted" 04:01 OldCoder Respond to that, politely, or back off 04:01 sofar your complaining and threats of legal action are not in any way a benefit to the community nor do they help the project in any way or form 04:01 OldCoder sofar, does the DDoS do so? 04:01 OldCoder Hm. That almost rhymes. 04:01 OldCoder Repond to that simple question. 04:02 OldCoder Your statement is a non-sequitur, additionally. 04:02 OldCoder If the project commits a prosecutable felony 04:02 OldCoder You are hardly in a position to make the statement that you just did 04:02 OldCoder This is not a kid with a neighborhood botnet 04:02 OldCoder Somebody is putting effort into it 04:03 OldCoder Don't talk to me about "threats" when "threats" have escalated to somebody hurting my host 04:04 sofar again, you are misrepresenting peoples statements left and right, making legal threats, and generally being enormously disruptive to the project and community. 04:05 sofar please, just leave. 04:05 sofar nobody could do more harm to the project than you are currently doing. 04:06 sofar oldcoder, you had a great run. You poored your heart into the project for 5 years 04:06 sofar more than anyone else I think 04:06 OldCoder poured * 04:06 OldCoder sofar, moment, host is assessing 04:06 OldCoder You can be that reasonable 04:07 sofar today is the day you have the opportunity to reassess your commitment and change 04:07 sofar or, you can continue to do what you're attempting to do this very night, and continue the damage 04:07 OldCoder But the point about "legal threats" is bizarre. Um, a felony is being committed and you don't consider consequences... 04:07 OldCoder Stop that. Now. 04:07 OldCoder You will not hold me responsible for the actions of others. 04:07 OldCoder I'll never go there again. 04:07 OldCoder Take responsibility for your own choices. 04:08 OldCoder I've done naught but help, be kind, and stand up to bullies and trolls. 04:08 OldCoder Speaking of which, do hold the part about how I should not talk about calling the police... 04:08 OldCoder until I find out the status of what there is to call about 04:08 OldCoder Minute 04:08 sofar I can not and will not prevent you from calling any authority 04:08 OldCoder Yet you continue 04:08 OldCoder to position the fact 04:08 sofar I'm asking you to leave this channel 04:08 OldCoder that that is the right thing to do 04:09 OldCoder And I am telling you a felony has been committed 04:09 OldCoder And that if you or any identifiable person 04:09 * VanessaE peeks in 04:09 OldCoder Kicks me... I will file a police report 04:09 VanessaE OC, take it to another channel NOW. 04:09 VanessaE this is NOT the place for this discussion 04:09 OldCoder VanessaE, I am being DDoSed. Kick me and I file against sofar at RedHat. 04:09 OldCoder Wait, the DDoS may be ending 04:09 VanessaE I don't have the authority to kick you but I DO have the authority to bitch you out 04:09 OldCoder Will you excuse me long enough to check? 04:09 VanessaE OFFTOPIC, LEAVE. 04:10 OldCoder It is not offtopic. Justify the point. 04:10 OldCoder But let me check the status of my host's hardware first, please 04:10 OldCoder Or is that too civil a request? 04:10 VanessaE this is for core discussion of Minetest engine code and game code, not legal threads over a DDoS that isn't even coming from any of us 04:10 VanessaE no I will NOT wait, take it out of this channel immediately. 04:10 OldCoder VanessaE, that is two parts. #1 is defensible. Justify #2. You know it is one of the people here. 04:10 ShadowBot https://github.com/minetest/minetest/issues/1 -- GlowStone code by anonymousAwesome 04:10 ShadowBot https://github.com/minetest/minetest/issues/2 -- Burned wood 04:11 rdococ . o O ( wtf is going on? ) 04:11 OldCoder I have been put through enough 04:11 VanessaE I justify item 2 because even if it WERE someone here, this is still not the place to discuss it. 04:11 OldCoder rdococ, I'm being DDoSed by one of the core devs or their friends 04:11 OldCoder That is point #1 and a point worth considering, VanessaE 04:11 ShadowBot https://github.com/minetest/minetest/issues/1 -- GlowStone code by anonymousAwesome 04:11 OldCoder Regarding point #2, you know they are right here 04:11 ShadowBot https://github.com/minetest/minetest/issues/2 -- Burned wood 04:11 rdococ OldCoder, are you sure it was by someone on here? 04:11 sofar everyone, please do not respond to oldcoder. We've asked him to leave and take it elsewhere 04:11 OldCoder sofar, go one way or the other 04:12 OldCoder rdococ, fair point, I'll discuss that with you in PM 04:12 * OldCoder goes to talk to host friend 04:12 VanessaE OldCoder: I know nothing at all about who is or is not the source of your DDoS, other than I know I've no connection with it because I've been fecking asleep for the past 2 hours, and afk for a few hours before that. 04:12 OldCoder I sleep a lot too, Vanessa 04:12 OldCoder What is your point? 04:13 sofar OldCoder: please leave the channel 04:13 OldCoder Do you have any idea what all of this has cost me financially? 04:13 VanessaE no specific point, just making it clear 04:13 OldCoder sofar, you have made statements 04:13 sofar OldCoder: please leave the channel 04:13 OldCoder which call your right to make that request into question 04:13 sofar OldCoder: please leave the channel 04:13 OldCoder If you wish 04:13 sofar OldCoder: please leave the channel 04:13 OldCoder actual war 04:13 OldCoder sofar, you are flooding 04:13 sofar then leave the channel 04:13 OldCoder Is it to be actual war? 04:13 VanessaE and for the record, any good server host will not charge you for a DDoS. they'll mitigate it. I know, from having been a target myself a few times. 04:13 OldCoder VanessaE, that is not the issue 04:13 VanessaE and he's just repeating, you are flooding. now TAKE IT SOMEWHERE ELSE. 04:13 OldCoder This host is a friend. The issue is affecting him. 04:14 sofar OldCoder: please leave the channel 04:14 OldCoder These cowards are hurting a friend. 04:14 OldCoder sofar, you are flooding 04:14 OldCoder Is it to be war? 04:14 OldCoder Do you think I won't quit a job 04:14 OldCoder That is now in jeopardy because of all of this 04:14 OldCoder And do what I must? 04:14 OldCoder Leave me alone 04:14 sofar OldCoder: please leave the channel 04:14 OldCoder Stop PMg my people 04:14 OldCoder Stop DDoSing me 04:15 OldCoder Stop saying that I "weird out" people by being "too detailed" 04:15 OldCoder Just stop it 04:15 VanessaE OldCoder: ENOUGH! 04:15 sofar OldCoder: please leave the channel 04:15 OldCoder My words exactly 04:15 OldCoder The decision is the group's. If I wake up tomorrow... 04:15 OldCoder and I find that anything has been escalated... 04:16 OldCoder Which would include ex post facto ban from this channel 04:16 OldCoder It is war 04:16 OldCoder You people should be ashamed of yourselves 04:16 OldCoder Enough, yourself, young lady :P 04:16 VanessaE you need to prove it was someone from here before you make a claim, OldCoder. 04:16 OldCoder The reaction is good enough. But do you deny it, honestly? 04:16 sofar this channel is not the place for legal accusations 04:16 OldCoder Speak the truth 04:17 OldCoder sofar, oh, really? But this is the channel 04:17 OldCoder of the Gods 04:17 VanessaE how can I deny something for which I have no information one way or the other? 04:17 sofar neither factual or implied 04:17 OldCoder Isn't it? 04:17 Hijiri anywhere with third parties isn't the place for legal accusations 04:17 rdococ -_- 04:17 Hijiri since that could be defamation 04:17 OldCoder That is not a legal basis for your assertion 04:17 OldCoder I fear no defamation 04:17 OldCoder Ever; since 2012 04:17 sofar OldCoder: please leave the channel 04:17 OldCoder sofar, you are tedious. I need to talk to my friend. 04:18 sofar please leave minetest entirely 04:18 OldCoder But understand... if I wake up and anything has been escalated 04:18 rdococ NONE OF US TOUCHED YOUR STUFF. NOW GO. 04:18 OldCoder The escalation may be regretted 04:18 OldCoder Is that what is desired? 04:18 OldCoder Must it be so? 04:18 rdococ ... 04:18 OldCoder ... 04:18 rdococ well, I don't know if anyone else touched your stuff. 04:18 rdococ but I know I didn't. 04:18 rdococ SO GO. 04:19 sofar OldCoder: please leave the channel 04:19 OldCoder rdococ, then "BE POLITE" 04:19 OldCoder Overnight, best decide: peace or war 04:19 OldCoder And this is the place to announce that 04:19 OldCoder It is one way or the other 04:19 OldCoder This has got to end 04:19 OldCoder If war is necessary, BTW, I am not very nice about it 04:19 OldCoder IRL 04:20 OldCoder Ask about a dozen attorneys, if you like 04:20 OldCoder Who were sicced on me 04:20 OldCoder I'll send you their home phone numbers 04:20 OldCoder Say Hi for me 04:20 sofar OldCoder: please leave the channel 04:20 OldCoder sofar, I am not escalating. You are. 04:20 OldCoder Does NOBODY but me understand deescalation? 04:20 sofar I'm still the person who is kindly asking you to leave 04:21 sofar so, take it like an adult 04:21 sofar go walk it off 04:21 VanessaE I'm being as kind as I can tolerate right now as well. leave the channel. 04:21 OldCoder sofar, how dare you? 04:21 OldCoder "Take it like an adult" ? 04:21 OldCoder My host says it's continuing. They just lost one leg. 04:22 OldCoder Some system kicked in. 04:22 OldCoder $2,000. That's how much this has cost me. 04:22 OldCoder Not for the trademark. Lost work. 04:22 OldCoder Get off your high horse, sofar. What has this cost you? 04:22 VanessaE "lost work" 04:22 OldCoder Same to you, VanessaE. 04:22 OldCoder Yes, what about it? 04:22 VanessaE that argument won't fly here and you should know that. 04:23 sofar I advise you to cut your financial losses to a minimum 04:23 OldCoder Hm? It's not an argument. 04:23 VanessaE as for what this has cost me? hmm... well add up three years of server hosting costs, whatever that comes to. 04:23 OldCoder I'm taking care of a dog tonight. He's hungry. Woof. Shall a dog whimper that you may feel superior? 04:23 OldCoder He's scratching for food right now 04:23 OldCoder I mean this rubbish, VanessaE 04:23 OldCoder The last 4 weeks 04:24 OldCoder I'm going to feed a dog. Presumably he won't bite me for doing so. 04:24 OldCoder If only people were animals of this nature. 04:24 sofar go walk it off, with the dog? 04:24 OldCoder You go walk it off 04:24 OldCoder But I must feed him right now 04:24 sofar please leave the channel as well 04:24 OldCoder Walk off 17,000 connections? 04:24 OldCoder sofar, I'll leave when the DDoS stops. 04:24 OldCoder If I'm kicked, I initiate 04:24 OldCoder IRL actions, not limited to legal 04:25 OldCoder Hm, that came out wrong. Or maybe it didn't. 04:25 VanessaE yeah, that came out VERY wrong on a publically-logged channel...... 04:25 OldCoder I was being facetious, VanessaE 04:25 OldCoder It came out 04:25 OldCoder Exactly as I wished 04:25 OldCoder sofar, Corporate Counsels don't ignore me 04:26 OldCoder The last one tried 04:26 OldCoder to 04:26 OldCoder They brought in somebody who argues before State Supreme Courts 04:26 OldCoder After a few weeks 04:26 OldCoder He sort of went away 04:26 OldCoder I called him Christmas Day just to say Hi 04:26 OldCoder :D 04:26 sofar you're argueing in an IRC channel, you should probably get a really good lawyer 04:26 sofar also, please leave the channel 04:26 OldCoder sofar, I am not arguing at all 04:26 OldCoder Dog is waiting 04:27 sofar he needs you more than this channel does 04:28 sofar so, think of the dog, leave the channel, and take the dog for a walk 04:29 OldCoder sofar, you are being... the word is not polite. But those behind the attack are cowardly. 04:29 OldCoder And that assessment 04:29 OldCoder is perfectly on topic 04:29 OldCoder The core devs have developed hatred and prosecutable crimes 04:30 OldCoder It is on the new plan 04:30 sofar The core devs have developed hatred and prosecutable crimes 04:30 sofar defamation 04:30 OldCoder Fact 04:30 OldCoder I can document 04:30 OldCoder the first part 04:30 sofar also, please leave the channel 04:30 OldCoder In detail over 4 weeks 04:30 OldCoder Also, please give it a rest 04:31 OldCoder (1) I leave when the DDoS stops (2) If I am kicked, the results will be satisfactory in the end to me 04:31 OldCoder None of this was necessary. You just had to leave me be. Without deprecation, attacks, basically rubbing salt in the wound. 04:33 OldCoder I've got kids, playing, right now; other countries. How nice that their playground is to be smashed up. 04:33 Zeno` I don't like being accused of things I did not commit 04:33 OldCoder Zeno`, I am dealing with the group. You are an individual. 04:33 OldCoder There are many kind people logged in. 04:33 OldCoder They ought to see this. 04:33 OldCoder And to understand just how far this has gone. 04:33 OldCoder And just how wrong it is. 04:34 OldCoder "Oh, look, somebody is disliked... throw stones, throw feces... 04:34 OldCoder make the stranger be dead" 04:34 sofar just, stop 04:34 sofar please leave the channel 04:34 OldCoder sofar, I ask the same of you. The first part. 04:34 sofar you're just looking for someone to acknowledge you 04:34 sofar for what reason? 04:34 OldCoder sofar, I am being DDoSed. 04:34 sofar don't even bother answering that 04:34 sofar just, go walk your dog 04:34 sofar please leave this channel 04:34 OldCoder If you wish to make other statements, discuss them or hush. 04:35 OldCoder You please leave this channel 04:35 sofar I've got another one 04:35 sofar please leave this channel. 04:35 OldCoder And I have no dog. I was helping somebody. 04:35 OldCoder please leave this channel. 04:35 sofar you're being disruptive. 04:35 OldCoder you're being disruptive. 04:35 OldCoder Adult, much? 04:35 OldCoder Grow up, yourself 04:35 OldCoder I have taken enough from the group that has attacked 04:35 OldCoder For a solid month now 04:35 OldCoder It ends 04:35 OldCoder Or it escalates 04:36 OldCoder Which is it to be? 04:36 sofar your decision 04:36 OldCoder I will never again be the boy 04:36 sofar so, go make it 04:36 OldCoder How is it my decision? Shall I stand there 04:36 OldCoder my blood dripping on the floor again 04:36 OldCoder And smile sweetly 04:36 OldCoder So sweetly 04:36 OldCoder That I shall be told 04:36 OldCoder I am not being disruptive? 04:36 OldCoder Never again 04:36 OldCoder Ever 04:36 OldCoder The attacks stop. Including from you, sofar. 04:37 OldCoder Or we go to the next level. 04:37 XtremeHacker What did I walk into? 0_0 04:37 OldCoder XtremeHacker, I'm being DDoSed 04:37 VanessaE XtremeHacker: just back away slowly and close the door.... 04:37 OldCoder It is tedious 04:37 OldCoder XtremeHacker, best to stay in -project 04:37 OldCoder But it's up to you 04:37 * XtremeHacker LEaves -dev 04:38 OldCoder Host says he can tolerate even 17,000 simultaneous connections... Not sure indefinitely, though 04:38 OldCoder I might need to ask Bonsai Kitten about this 04:38 OldCoder Hm 04:38 OldCoder What sort of hardware is minetest.net hosted on? 04:39 OldCoder Does anybody know? 04:39 OldCoder What sort of network? 04:39 sofar are you threatening to DDos it? 04:39 OldCoder Is it really, really good? O_O 04:39 OldCoder Certainly not! 04:39 OldCoder It's just that 04:39 OldCoder If 04:39 OldCoder These people are randomly attacking Minetest 04:39 OldCoder And it is NOT just an attack on me 04:39 OldCoder Aren't the odds pretty good 04:40 OldCoder That they will take down Minetest.Net 04:40 OldCoder with something that makes 17,000 for a day or two 04:40 OldCoder Look puny? 04:40 OldCoder ^ Reasonable point 04:40 * OldCoder is shocked to realize that minetest.net may be at risk 04:40 sofar discussions about DDos-ing project servers (hypothetically) is certainly not appropriate for this channel 04:41 OldCoder Hypothetical is probably not on-topic. I contend that ongoing as in right now certainly is on-topic. 04:41 sofar your presence in this channel is disrupting the very goals of which it was created for 04:42 OldCoder sofar, do not even try to go there. Blame the victim. Sickening. 04:42 sofar Instead of discussing code and PRs, look at the channel now 04:42 OldCoder The victim is not at fault. 04:42 sofar you are doing more damage to the project than anyone else 04:42 OldCoder Rape is not justified by the fact that the victim is unable to fight back. 04:43 OldCoder Oh no O_O By saying I do not wish to be deprecated, dogpiled, lied about, DDoSed 04:43 sofar then go call your lawyer already, or call the authorities 04:43 OldCoder I am doing irreparable harm 04:43 OldCoder Oh no! 04:43 sofar but please, leave the channel 04:43 OldCoder sofar, it's martial arts 04:43 OldCoder The fact is, and I'm being pretty open about this 04:43 OldCoder My IQ is probably Hawkins level plus 5 04:43 OldCoder I learned this in 2012 04:43 OldCoder I could *see* everything 04:43 OldCoder laid out in legal actions 04:44 OldCoder And I realized that I could do *nothing* 04:44 OldCoder except when the other side moved 04:44 OldCoder Then they did 04:44 OldCoder And it was funny :D 04:44 OldCoder The trick is you don't let the other side choose the battlefield 04:44 OldCoder You will not be doing that to me 04:44 OldCoder The battlefield will not be DDoS 04:45 OldCoder And nobody in the project, including Perttu 04:45 OldCoder Hm; never mind the next part 04:45 OldCoder 05:00 rdococ why did I have to wake up 05:00 VanessaE go back to bed :P 05:00 rdococ I can't 05:00 rdococ once I wake up I can't go back to sleep 05:04 OldCoder rdococ, try melatonin 05:04 OldCoder and I've responded to your question in PM 05:05 OldCoder rdococ, VanessaE, sleep is off-topic 05:05 OldCoder :P 05:36 OldCoder so I'm gonna say one last thing: don't ever bring any of this noise into -dev again. the channel has a specific subject... if you are not discussing a patch or a pull request or a bug in the engine or _game or discussing making a change to it, or maybe making a quick one-liner joke, you're offtopic and need to take it somewhere else, regardless of what you think is or is not on-topic. 05:36 OldCoder VanessaE, deescalate. 05:36 VanessaE ^^^ private message sent to OC that didn't need to be repeated in here. 05:43 OldCoder Just as shivajiva's PM dump did not need to appear in #minetest-project 05:43 OldCoder nerzhul, good morning 05:43 OldCoder Excuse me, PMg with Vanessa 05:44 nerzhul hi 05:44 OldCoder o/ 05:53 nerzhul some noise on -dev unrelated to minetest development this night, hmmm 05:54 nerzhul can someone approve #5621 ? fairly trivial fix for an external contributor 05:54 ShadowBot https://github.com/minetest/minetest/issues/5621 -- Fix #5617 - respect message and reconnect parameters when shutting down immediately by orwell96 05:54 VanessaE wait, what? something ON-topic? 05:54 * VanessaE falls over 05:55 nerzhul guys feature freeze is in 1 month, please finish all milestone 0.4.16 pr :) 05:55 VanessaE there's another odd shutdown issue that should be looked at - the client likes to hang on the way down. Seems limited to singleplayer situations. 05:55 VanessaE hang to the point that I have to SIGKILL it. 05:56 nerzhul only when using shutdown command ? or in another way ? 05:56 nerzhul celeron55, can you add feature freeze date in topic (21st may), please ? 05:56 VanessaE well since it's singleplayer, usually ESC -> "exit to OS" 05:57 VanessaE I haven't noticed any problems with the separate server instance shutting down, other than that one bug that happened and was fixed, a few weeks back 05:57 nerzhul strange, is this triggered in every case ? 06:00 VanessaE noper 06:01 VanessaE it's pretty random 06:01 VanessaE next time I'll try to get a backtrace 06:01 nerzhul it's difficult :) maybe it can be related to some mods ? 06:04 VanessaE it could be, I don't know for sure 06:04 VanessaE however, 06:04 VanessaE since it requires a SIGKILL, probably not. 06:04 sofar #5619 updated, I think it's OK now 06:04 ShadowBot https://github.com/minetest/minetest/issues/5619 -- Add on_flood() callback. by sofar 06:34 nerzhul i will push the fairly trivial but performance improvement #5628 when travis build will finish 06:34 ShadowBot https://github.com/minetest/minetest/issues/5628 -- Fix various performance issues reported by cppcheck by nerzhul 06:57 celeron55 i'll add some core devs as ops here 06:58 VanessaE thanks. 06:58 celeron55 ask me later if you don't want to have +o, i'll just pick randomly for now 07:00 VanessaE meanwhile, a user over on #minetest asks for a filtering feature for the in-client server list, i.e. show only servers that have PVP and mod "foo" installed, that sort of thing 07:04 celeron55 how about that for now 07:05 celeron55 coming here with DDoS accusations is just stupid 07:06 VanessaE I *still* wonder who DDoS'ed him. 07:06 celeron55 he doesn't need to have this much attention 07:06 celeron55 i want people to be able to focus on minetest, not OldCoder 07:08 celeron55 maybe some script kiddie that happened to see some discussions; they tend to show their opinions that way 07:09 VanessaE perhaps. 07:12 celeron55 anyway, there is a limit for how much of my life i am going to waste with OC 07:13 red-000 wow my chat is spammed with OC 07:39 red-000 could someone review #5492? 07:39 ShadowBot https://github.com/minetest/minetest/issues/5492 -- [CSM] Add local particles and particlespawners. by red-001 07:52 nerzhul celeron55: in the topic on the channel, we have since many years a reference to whispers to you, is this always a reality ? 08:00 celeron55 it's been there since i created this channel 08:00 nerzhul is this always necessary to have this information ? :p maybe it promote you to be whisped :p 08:02 celeron55 (this channel was literally created to replace a long /msg with a contributor back then) 08:04 Zeno` hmm 08:06 nerzhul yeah, understood heh 08:06 nerzhul merging #5628 08:06 ShadowBot https://github.com/minetest/minetest/issues/5628 -- Fix various performance issues reported by cppcheck by nerzhul 08:06 nerzhul i didn't fixed all cppcheck issues yet, but ~65% are now done, the most criticals 08:07 VanessaE nerzhul: how much effect will those changes have 08:07 VanessaE (performance improvements always pique my interest) 08:07 nerzhul not very very huge, don't wait for a real difference, current problems needs a better approach than removing memory copis :p 08:07 Zeno` I guess cppcheck has been updated recently. About 3 months ago it didn't report much interesting at all 08:07 VanessaE fair enough. 08:08 nerzhul Zeno`: don't just do cppcheck, but cppcheck --enable=performance --std=c++03 src/ :) 08:08 Zeno` I'm meant to type all that?! :-o ;) 08:08 nerzhul for the network part it remove 4 packet copies in the connection threads, which is systematic and can a little bit improve latency 08:09 nerzhul Zeno`: heh, you can --enable=all if you want but some things are not very interesting (style issues for example) 08:09 Zeno` that sounds easier :D 08:09 nerzhul after this PR merging, in performance issue if i remember are 4-5 memory copies and ~40 constructor optimizations 08:10 nerzhul please note when switching to C++11 we will have many more errors to fix which are shown by cppcheck to improve a little bit more performance due to new compilers/standards 08:11 nerzhul oh nice ! 08:11 nerzhul https://gitlab.com/nerzhul/minetest/pipelines/7774250 08:11 nerzhul a full pipeline for Debian and ubuntu (and just a build for fedora i should add package & deploy steps :D 08:17 nerzhul celeron55: is this okay if i push #5629 for the gitlab daily build in the mirror repository ? 08:17 ShadowBot https://github.com/minetest/minetest/issues/5629 -- Gitlab-CI enhancements by nerzhul 08:17 nerzhul it doesn't affect github but i really want to continue and finish daily builds, and this is the second step :) 08:17 nerzhul maybe at the end make gitlab push some artifacts to github or somewhere else for end users 08:18 celeron55 if there's no other way to control gitlab, i guess it's okay 08:19 celeron55 it's kind of silly though 08:19 nerzhul .gitlab-ci.yml is equivalent to .travis.yml :) 08:19 celeron55 travis is silly also 08:19 nerzhul i prefer gitlab-ci it's a little bit simple to handle and docker usage permit to test on many systems, unlike travis which only uses ubuntu 08:19 celeron55 it's like they use our project for marketing by forcing their file to be there 08:20 nerzhul celeron55: heh 09:03 nerzhul what are those [off] everywhere 09:07 red-000 can someone kickban HiBitches 09:07 HiBitches I'll be back in 3 minutes 09:07 HiBitches Red, if they do 09:07 HiBitches alternative is to go +m on channel 09:07 red-000 go to your own channel 09:07 HiBitches There is a fallback plan for that 09:07 kaeza so mature 09:07 HiBitches red-000, PM, please, no sneeze, you are offtopic 09:07 HiBitches kaeza, I agree 09:07 kilbith how about to go fuck yourself 09:07 HiBitches PM, please 09:08 HiBitches This is offtime and I am lawabiding 09:08 HiBitches offtopic, sorry 09:08 red-000 -> #minetest-project 09:08 HiBitches PM, please 09:08 HiBitches Be good citizens 09:08 nerzhul VanessaE, oh i didn't knows that 09:08 HiBitches Ban or nothing. And ban means HMA. 09:08 red-000 Isn't ban evading agaist freenode rules? 09:08 HiBitches It surely is. 09:08 HiBitches Shocking. 09:09 HiBitches So is DDoS ofc 09:09 nerzhul i didn't ban just kick 09:09 HiBitches Sigh 09:09 HiBitches Let me sleep 09:09 VanessaE nerzhul: it also works to keep chat in an IRC channel from propagating to minetest server instances that are in that channel via kaeza's IRC mod. 09:09 nerzhul oh, it's because it's handled like this in the reading components then 09:10 red-000 nerzhul, oldcoder already got banned 09:10 HiBitches That was probably foolish 09:10 HiBitches as he is insane and not age 14 09:10 HiBitches and has sufficient funds 09:10 HiBitches to be rather inconvenient 09:10 HiBitches Oh no 09:10 HiBitches He hires detectives 09:10 HiBitches IRL 09:10 kaeza potato 09:11 HiBitches Did you know that? He's really quite a B*tch 09:11 HiBitches He likes to spread U.S. SSNs around 09:11 HiBitches for US people 09:11 red-000 ? 09:11 HiBitches vanessa, PM, truly; RBA did care for you 09:11 red-000 potato? 09:11 HiBitches I think 09:12 HiBitches It's a potat-Oh po-tah-to reference 09:12 VanessaE potato. 09:12 red-000 I'm confused 09:12 HiBitches It is a criticism of obsession 09:12 kilbith stop answering him already 09:12 red-000 why is everyone saying potato? 09:12 HiBitches A US meme of decades ago 09:12 kaeza potato! 09:12 HiBitches red-000: it's a meme 09:12 red-000 !?? 09:12 VanessaE hash browns. :) 09:12 HiBitches kilbith don't be a, um, bith 09:13 VanessaE because we're just being silly since the channel is slow right now :) 09:13 kilbith celeron55: please, shoot him down 09:13 kaeza nerzhul, ^ 09:13 HiBitches kilbith I'd be back in 2 minutes 09:13 red-000 I'm really confused 09:13 HiBitches Tomorrow I'll have 700 IPs 09:13 HiBitches Just as LM had 09:13 celeron55 what on earth is this now 09:13 HiBitches That is kid stuff 09:13 HiBitches It is potato meme 09:13 kilbith should we request a k-line to the Freenode admins ? 09:13 HiBitches kilbith this is a random IPV4 09:14 HiBitches Tomorrow I sign up for HMA 09:14 HiBitches After that I hire detectives 09:14 HiBitches Are you in the US BTW? But you are German 09:14 HiBitches *sad* 09:14 red-000 pretty sure ban evading is a k-linable offence 09:14 HiBitches It is, yes\ 09:14 HiBitches Indeed 09:14 HiBitches So is DDoS 09:14 nerzhul it's a very offtopic from an unpretty nick 09:14 celeron55 let's play this game 09:15 nerzhul i never see a such thing on IRC since 3 years... 09:15 celeron55 just whoever, i need to work 09:15 red-000 wow 09:15 kaeza op party? 09:15 shivajiva kk 09:15 red-000 thats a lot of ops 09:15 VanessaE have a better day, ce1 09:16 VanessaE er celeron55 09:17 kilbith wait, there's a better approach, give voice only to users with +v mode 09:17 nerzhul kilbith, it's not a good idea on this channel 09:17 nerzhul it's a more open channel than hub 09:17 nerzhul everybody should have a chance to discuss to developers, respecting rules 09:17 kilbith we don't need an "open bar" currently 09:22 IamSleepyFriend Hey 09:22 IamSleepyFriend Love abounds, hounds 09:23 IamSleepyFriend I do have a Visa, you know 09:23 IamSleepyFriend HMA when I wake up 09:23 nerzhul celeron you ban root@* ? :p 09:23 celeron55 mistake 09:23 nerzhul nobody can connect as root, security 09:23 nerzhul :p 09:24 celeron55 +m might be a good idea but we'll see 09:25 red-000 it's a bad idea to connect as root anyway 09:26 VanessaE don't most IRC clients and servers warn you if you do? 09:26 celeron55 well, root is probably just a temporary hacked IoT device anyway 09:27 kaeza or PuppyLinux 09:27 kaeza you are always root there, unless you specifically go through hoops to set up the damn thig properly 09:27 kaeza +n 09:36 red-000 or windows in a lot of common setups 09:37 celeron55 windows IRC clients don't tell it to the server 09:40 red-000 that was my ip 09:41 red-000 I was testing something 12:17 juhdanad #5186 is rebased, but I need help with pahole. 12:17 ShadowBot https://github.com/minetest/minetest/issues/5186 -- Soft node overlay by juhdanad 12:19 juhdanad My problem is that I use Ubuntu, and the provided pahole has a bug and does not work with the built object files. 12:20 juhdanad I also tried to build the patched version from source, but a required library is not available anymore. 12:20 juhdanad So I can not test if the memory layout of TileLayer and TileSpec is correct. 12:25 VanessaE pahole? 12:25 juhdanad An utility program to find holes in the memory. 12:25 VanessaE ok 12:41 nerzhul juhdanad, it's sofar who ask you to do that ? 12:43 juhdanad He did not ask me. I wanted to do this because of this line comment: https://github.com/minetest/minetest/commit/93c1d511e3880acf15a656f1653ce5853dd1d362#diff-f44f50a96d13d2421bd5301b7511ee26R288 12:52 sfan5 can't your use a chroot or vm to use pahole 12:53 juhdanad No. I can't compile it. 12:54 sfan5 yes that is why you should use a chroot with a different distro 13:04 Zeno` If you have to try a different distro then the optimisation is bogus 13:05 Zeno` actually that whole comment is bogus 13:05 juhdanad Well, I guess the best would be to use tile spec references instead of memory copy... 13:05 Zeno` that may be the case but it doesn't make the comment any more valid 13:05 Zeno` wait 13:05 Zeno` it's a pointer anyway :/ 13:06 Zeno` where is the memory copy? 13:08 juhdanad https://github.com/minetest/minetest/blob/master/src/mapblock_mesh.cpp#L678 13:08 juhdanad is an example 13:09 Zeno` I don't like it. Optimising that way is bad 13:10 Zeno` even worse, who knows what architecture this is being built on? 13:10 Zeno` it's a bit crazy 13:11 kilbith it's 50000% faster, Craig 13:12 Zeno` if it's only 50000% it's not worth it 13:12 juhdanad But it won't be easy to use pointers everywhere, as some parts of the tile specs are modified (rotation, color) 13:12 Zeno` juhdanad, huh? 13:13 juhdanad What do you not understand? 13:13 Zeno` Why you cannot modify an object that is pointed to 13:14 Zeno` that's kind of the point of pointers :/ 13:15 juhdanad So currently a mapblock mesh contains lots of tile specs. These contain redundant information - texture, animation. So it would be better if these meshes only had pointers instead of tile spec copies. 13:17 juhdanad If you have a flat block with grass, it creates 16 copies of the grass node's tile spec: 16 times the texture pointer and animation data. 13:23 Zeno` so remove the comment? 13:24 juhdanad I don't think it should be removed, as now TileSpec is better optimized. 13:24 Zeno` It is? 13:25 agrecascino uh 13:25 juhdanad Yes, some holes were removed by sofar. 13:25 agrecascino is anyone working on a client side lua implementation 13:26 Zeno` I don't get it juhdanad 13:26 juhdanad agrecasino: it is already implemented. 13:26 agrecascino oh snap 13:26 Zeno` anyway, will merge #5186 within 30 minutes 13:26 ShadowBot https://github.com/minetest/minetest/issues/5186 -- Soft node overlay by juhdanad 13:27 agrecascino heh 13:29 agrecascino so when did it like get implemented? 13:30 juhdanad To sum up: sofar reorganized TileSpec so it requires less memory. Meanwhile my PR also changes TileSpec. The only problem is that I can not check for holes in my PR, so it might not be as optimal as possible. 13:31 juhdanad agrecasino: #5088 13:31 ShadowBot https://github.com/minetest/minetest/issues/5088 -- Client side scripting/modding by nerzhul 13:32 agrecascino cool! 13:32 juhdanad also #5394 13:32 ShadowBot https://github.com/minetest/minetest/issues/5394 -- [CSM] Second Roadmap 13:32 agrecascino uh 13:33 agrecascino are like nametags modfiable via csm? 13:33 juhdanad Not yet. There are no client side object references yet. 13:34 agrecascino rip 13:35 agrecascino also 13:35 agrecascino can the client communicate to the server 13:37 juhdanad Not yet, but it is planned too. 13:37 agrecascino uh 13:37 agrecascino are there any plans yet 13:38 juhdanad You can read the second roadmap for the plans. 13:38 juhdanad Also you can help in coding if you want to. 13:39 agrecascino i would but my c++ code always ends up as object spaghetti 13:39 agrecascino but i was thinking, since my religion is wamp, what about wamp for messaging 13:41 juhdanad something like #5390? 13:41 ShadowBot https://github.com/minetest/minetest/issues/5390 -- CSM: messaging channels 13:42 agrecascino yes 13:42 agrecascino with wamp as the implementation on the protocol level 13:42 agrecascino because wamp 13:58 Taoki Thought I'd do my periodic asking for this question: Has any progress been made on shaders and graphics, including RBA's latest in-the-works codes? 13:59 Taoki Or more specifically: Is there any hope of seeing any of the following features make it into master soon: Extruded (3D-fied) plantlike, bloom shader, depth-of-field shader? 13:59 Taoki (those were the ones I was most interested in, and obviously still am) 14:02 juhdanad There's only one enhancement I know of: undersampling. 14:02 Taoki Interesting. Any screenshots? 14:02 juhdanad #5057 14:02 ShadowBot https://github.com/minetest/minetest/issues/5057 -- Undersampling by numberZero 14:03 Taoki Thanks 14:03 Taoki Cool. Can help with offering a pixelated, Doom-like art style :) 14:05 Taoki Woah... client-side Lua scripting is a thing now? Noticed it in one of the pull requests. Been away for quite some time it seems. 14:08 agrecascino Taoki, yeah, i stopped paying attention around 8 months ago, for various reasons 14:10 Calinou Taoki: current undersampling kind of sucks because you can't set it to be filtered linearly 14:10 Taoki Fair enough 14:10 Calinou I'd like that for my 4K laptop, while I want Minetest to render at 1080p (else it's unplayable), I don't want it to be blocky 14:10 Calinou the way it's implemented makes it non-trivial to implement that 14:14 agrecascino uh, i have to ask 14:15 agrecascino who thought it was a good idea to have 4k displays on laptops that can barely into 3d graphics at 1440p 14:15 Taoki Heh. Also, I assume that to this day, a working implementation for VBO was still not achieved :P 14:15 Taoki I think it's like, half a decade now since we're trying to get working VBO in master, for that essential performance improvement. 14:18 agrecascino wait 14:18 Calinou I think VBOs were merged? maybe enabled by default 14:18 agrecascino i have to ask, if we aren't using vbos, what are we using? 14:18 Calinou https://github.com/minetest/minetest/blob/master/minetest.conf.example#L329 14:18 Taoki Calinou: That would be wonderful 14:18 Calinou enabled by default, it seems 14:19 agrecascino i'm pretty sure minetest doesn't use opengl 1.1 14:19 Taoki agrecascino: Drawing by the CPU each frame, which was very slow 14:19 Calinou it's been here for at least a year, according to Git blame 14:19 Taoki Oh... amazing! 14:19 agrecascino i mean 14:19 Calinou agrecascino: Minetest uses whatever OpenGL is supported by the system 14:19 Taoki Ok... might have forgotten then 14:19 Calinou currently, I don't think any OpenGL 4.x features are used, although OpenGL 3.x may help in places, I don't remember 14:19 agrecascino my current attempt at a game engine uses gl 1.1 interwhatever mode calls 14:20 agrecascino but that's because opengl 3 is hard and a waste of time for me 14:20 Calinou immediate mode (and thus, OpenGL 1.x) has been obsolete since like 10 years 14:21 agrecascino just because it's obsolete doesn't mean it's not easy 14:21 Calinou well, it's like using mysql_* functions in PHP, it shouldn't be done in 2017 :) 14:22 Calinou and it's not like you can make anything good-looking with OpenGL 1.x anyway 14:22 Calinou (barring pixel art games, which are their own thing) 14:22 agrecascino obviously, but it's still super easy to use 14:24 agrecascino my laziness succeeds my reason in decision making 14:28 Taoki https://cloud.githubusercontent.com/assets/825418/25281686/c1c3319c-26b6-11e7-89ed-ccaafcc2b0c3.png Weird bug that just started happening on my machine. 14:28 Taoki Reported it here for now: https://github.com/minetest/minetest/issues/5631 14:33 agrecascino Taoki, surprisingly pretty 14:34 Taoki Still a bug, but true :P 14:34 agrecascino uh 14:35 agrecascino can client-side lua scripts play audio 14:36 Fixer agrecascino: 4k on laptops is a myth, 99% own 1366*768 garbage 14:36 agrecascino Fixer, it definitely exists 14:37 agrecascino or at least, too high-res-for-the-laptop-hardware laptops exist 14:37 Fixer i see even 1000 usd laptops has this 1366*768 garbage included... 14:37 Fixer what a joke 14:38 agrecascino ie: friend's laptop has like 2880xsomething screen 14:38 agrecascino Fixer, it's even worse that most are TN panels 14:38 Fixer yep 14:38 agrecascino also 14:38 Fixer TN 1366*768 garbage even on 1000+ usd laptop 14:38 agrecascino minetest has endian issues 14:39 agrecascino Fixer, i have a sager laptop with a 1080p screen 14:39 agrecascino worst tn panel evr 14:39 agrecascino ever* 14:39 agrecascino it's 18 bit with dithering 14:39 agrecascino the worst dithering i've ever seen 14:39 Zeno` mine is true 1920x1080 14:39 agrecascino to the point where dark grays produce weird line artifacts on the screen 14:40 agrecascino in linux i can disable the dithering, and the artifacting goes away, but it's still pretty bad 14:40 Taoki Perhaps client-side Lua also means a decent weather system. People tried doing particle weather in the past, but it sucked because you had to move a particle spawner on the server based on the player's position: The server lagged, rain would start dropping from another location. 14:40 Fixer Taoki: that coloring bug could be related to that recent coloring changes 14:41 Taoki Fixer: If no one else is getting it, it could be because I use Mesa / RadeonSI 14:41 agrecascino i want to see protracker in minetest 14:41 Taoki Hopefully it won't be hard to debug and fix 14:41 agrecascino or something that could read/write MOD files in minetest 14:41 Fixer Taoki: what is your version? 1ffb180? What is server you connecting? 14:42 Taoki Fixer: Latest Git master, local server (singleplayer game) 14:42 Fixer Taoki: try yesterdays build 14:42 Taoki Can I download it from somewhere else? 14:42 Fixer Taoki: win/lin? 14:42 Taoki Linux 14:43 agrecascino maybe i can finally implement nametag fading and stuff in a sensible way 14:44 Taoki That would be great 14:44 Zeno` protracker is not as easy as you'd first think 14:44 Fixer Taoki: you said you recompiled it, why can't you recompile one commit back, it is pretty easy? 14:44 Taoki Also, have any thoughts been given on a pixelated large font, similar to Minecraft? Not to copy MC per say... just always wondered if that would look and feel better. 14:44 juhdanad Taoki: maybe what you are seeing is #4986. Now it is possible to re-colorize nodes. 14:45 ShadowBot https://github.com/minetest/minetest/issues/4986 -- Hardware node coloring by juhdanad 14:45 Taoki Fixer: I guess I can try it, just don't wanna mess GIT up. What was the command to switch your local master back to a previous commit in it, but have that safely reverted once I do 'git pull' again? 14:45 Taoki Fixer: Also, which commit in particular do you recommend I try? 14:45 Fixer Taoki: i do it like this, go into /minetest and run: git checkout 2ad74a9 14:45 Fixer Taoki: then compile it 14:46 Fixer Taoki: after that you can go back by typing: git checkout master (if i remember correctly, if master does not work try head or origin).. 14:46 Fixer go back to newest i mean 14:46 Fixer i'm compiling engine right now 14:47 Taoki Done... recompiling now against that commit 14:52 Taoki Darn... it's recompiling everything. Might take a while 14:58 nore might be related to #5555 14:58 ShadowBot https://github.com/minetest/minetest/issues/5555 -- Some kinds of nodes are now colorized when they should not be. 14:58 Taoki Fixer: The problem no longer exists with that commit! So it is introduced after 2ad74a9e8b9da48aef62346de6cd55f304c6440d 14:58 Fixer kek 14:59 Fixer juhdanad: ^^^^ 14:59 nore Taoki: does your version have this commit? https://github.com/minetest/minetest/commit/6f641df8a52ccb84452a289416527e3d1a36621a 15:00 Taoki nore: Possibly. Hard to tell and I need to run soon 15:00 Taoki It seems the very latest commit has in fact introduced it 15:00 nore ah, the node overlay? 15:03 Fixer compiled, will check to 15:03 Fixer o 15:03 juhdanad I suppose you are playing on VanessaE's server. 15:05 juhdanad Then the reason of the colored nodes is https://github.com/minetest/minetest/issues/5555#issuecomment-293421136. 15:05 Fixer juhdanad: i have the same problem, checking if in vanilla too 15:05 juhdanad Testing... 15:05 Fixer got crasjh 15:07 Fixer juhdanad: confirmed on vanilla world https://i.imgur.com/qkzyGI2.png 15:07 Fixer and crash on exit 15:08 juhdanad I'm currently compiling, and will push a fix ASAP. 15:08 Taoki Need to go now. Glad I could help to idenfity this... good luck on finding the cause, and thanks for the help! 15:08 Fixer juhdanad: my config https://pastebin.com/raw/hUq3gZss 15:08 Fixer juhdanad: do you need a backtrace for crash on exit? 15:08 Fixer juhdanad: do you have crash when exiting from the world via menu? 15:09 juhdanad No, for me the game works as it should. 15:10 Fixer will get a backtrace then 15:17 juhdanad Okay, with your config I see the miscolorization, but not the crash. 15:18 Fixer juhdanad: join the world, type esc - click on quit the game 15:19 juhdanad I get back to the menu. 15:19 Fixer crap, i will post a backtrace anyway 15:40 juhdanad I found the cause of miscolorization, fixing... 15:40 Fixer backtrace of my crash https://pastebin.com/raw/84XS9fhP 15:40 Fixer will post an issue 15:41 Fixer 0x00000000007470ab in CItemDefManager::~CItemDefManager (this=0xa862b90, __in_chrg=) 15:41 Fixer at /media/sf_Exchange/minetest/build/minetest/src/itemdef.cpp:264 15:45 juhdanad Fixer: could you change wieldmesh.cpp:441 to 'scene::SMesh *mesh = NULL;', please? 15:45 juhdanad I hope it will fix the crash. 15:45 Fixer yep 15:45 Fixer will do 15:47 Fixer compiling 15:51 Fixer juhdanad: fixed the crash for me 15:52 juhdanad Great! 15:52 juhdanad There was an item which did not have an inventory image and was not a node. In this case the mesh pointer was not initialized. 16:04 VanessaE paramat: saw in the commit log, "Tools: Fix tool digging speed limit" ... reminded me, can someone please fix the digging speed on unknown nodes (so that it respects whatever tool is being used)? 16:09 juhdanad #5633 16:09 ShadowBot https://github.com/minetest/minetest/issues/5633 -- Fix after soft node overlays by juhdanad 16:40 red-000 are we banning all ip's oldcoder has control over? 16:41 VanessaE ordinarily I would suggest doing so 16:41 red-000 it's one of his sites 16:42 Krock am I not looking correct? where do all the ops come from? 16:42 Krock *opens logs* 16:42 red-000 He used it to contact me before 16:43 VanessaE Krock: celeron55 spread the ops around. 16:43 red-000 another one of his sites 16:44 * Krock looks at sofar and wonders why he's so strongly opposed to OC 16:44 benrob0329 Did you read the logs? 16:46 sofar no worries, it's hightly confusing what happened 16:48 juhdanad I don't think that it is wise to ban him. I'm sure that he has done a lot for this community. 16:48 Krock right, needs some time to filter out what's the actual problem. I wonder why he even came here 16:49 Krock juhdanad, yes he did but it must not be a perma-ban. Also: I think it would only be applied on #-dev 16:57 red-000 yes 16:58 VanessaE palemoon, the web browser? 16:58 red-000 yes 16:58 VanessaE huh. 16:59 VanessaE http://www.palemoon.org/ 17:02 sofar juhdanad: did you rebase your PR? I'll see if I can get the pahole output for you 17:03 juhdanad sofer: yes, and it was merged too. 17:03 juhdanad *sofar 17:11 sofar juhdanad: I'll still look 17:16 juhdanad sofar: thank you! 17:17 sfan5 merging #5633 since it's a trivial bugfix 17:17 ShadowBot https://github.com/minetest/minetest/issues/5633 -- Fix after soft node overlays by juhdanad 17:20 nerzhul nice sfan5 17:22 red-NaN linked to another OC site 17:22 red-NaN http://haggishell.com/ 17:23 Grandolf why are you banning everrything OldCoder? 17:23 garywhite Who is 88.198.98.192? 17:23 red-NaN Grandolf, he uses his servers to ban evade 17:24 Grandolf why ban him? 17:24 VanessaE Grandolf: he was banned for being extremely disruptive, then proceeded to evade the ban 2 or 3 times. 17:25 VanessaE that's a re-bannable offense in most channels, and is probably against this network's usage rules as well. 17:25 red-NaN ^ 17:25 Grandolf when was this? 17:25 red-NaN pretty sure you can get k-lined for that if you annoy the wrong people 17:25 VanessaE earlier today. 17:25 VanessaE http://irc.minetest.net/minetest-dev/2017-04-21 17:25 nerzhul 181 pr :( 17:25 VanessaE nerzhul: they'll get whittled down soon enough 17:25 VanessaE we are approaching feature-freeze after all 17:26 nerzhul sfan5, can you look at #5621 it's fairly trivial too 17:26 ShadowBot https://github.com/minetest/minetest/issues/5621 -- Fix #5617 - respect message and reconnect parameters when shutting down immediately by orwell96 17:29 sfan5 nerzhul: lgtm 17:32 nerzhul merged 17:32 nerzhul ty 17:38 juhdanad I was planning to create a 'save for older Minetest version' feature. What do you think? How should users access it? 17:40 VanessaE nerzhul: remember the thing I mentioned about minetest crashing on exit? 17:40 VanessaE (not related to the soft overlay issue) 17:41 VanessaE it's not limited to singleplayer. I've got a client hung right now, just exited from doing a bunch of running around on my survival server. sorry I didn't think to run it in gdb 17:41 DS-minetest (#5604 is also very trivial) 17:41 ShadowBot https://github.com/minetest/minetest/issues/5604 -- lua_api: Correct craft recipe param `type` to `method` by DS-Minetest 17:41 VanessaE BAM, got gdb to attach to it 17:42 sfan5 it it a debug build? 17:42 VanessaE and...it's a release build :-/ 17:42 nerzhul VanessaE, is this not related to the PR merged by sfan5 ? 17:42 VanessaE http://pastebin.ubuntu.com/24428119/ 17:42 sfan5 no that's something else i think 17:42 VanessaE nerzhul: it is not. this is that crash bug I mentioned before the soft overlays went in 17:42 sfan5 looks like a deadlock to me 17:43 nerzhul sfan5, yes 17:43 VanessaE is that bt at all useful without debug symbols? 17:43 sfan5 yes 17:43 VanessaE oh good. 17:43 sfan5 debug symbols "only" tell us which file & line 17:43 sfan5 normal symbols are already better than nothing 17:44 sfan5 CurlFetchThread and the main thread (httpfetch stuff) seem to be waiting for eachother 17:44 VanessaE er, why? 17:45 sfan5 that's the question 17:45 VanessaE why is curl even getting involved in shutdown? 17:45 sfan5 curl threads need to exit 17:45 VanessaE oh right 17:45 sfan5 cf. curl_parallel_limit = n in .conf 17:56 VanessaE want anything else from this hung client before I kill it? 17:57 sfan5 nah 17:57 VanessaE ok. 17:57 VanessaE killed. 17:59 VanessaE bbl 18:04 ThomasMonroe VanessaE, sfan5, sofar, rubenwardy, nore, red-NaN, i need to talk to you guys for a bit 18:04 ThomasMonroe in private 18:05 sfan5 ...to all of us? 18:05 red-NaN diffrent channel? 18:05 ThomasMonroe im making a channel right now, just wait a mom 18:05 red-NaN just join one 18:05 red-NaN and send everyone the name in a pm 18:16 Krock ThomasMonroe, you're summoning quite many people there.. 18:16 ThomasMonroe ik Krock 18:16 nerzhul you are privileged guys :p 18:17 nerzhul maybe he will tell you that sofar insult me like the tell me this morning 18:17 nerzhul he* 18:17 Krock nerzhul, huh? could you please re-phrase that sentence? 18:18 nerzhul this morning OC whisper me on IRC and told me sofar insult me in PM when he talked with him 18:18 nerzhul (sorry for talking like a spaghetti :p) 18:18 Krock ah okay. now I get it :) 18:18 sfan5 nerzhul: don't you mean baguette ;) 18:18 Krock no, you're not italiano. only these guys can talk spaghetti 18:18 nerzhul :p 18:18 nerzhul lol 18:19 * Krock imagines sfan5 talking like beer.. 18:19 nerzhul beer is belgium not german haha :D 18:19 sfan5 :/ 18:20 Krock oh really? but.. oktoberfest! 18:20 nerzhul german is deustch kalidad :D 18:21 Krock #5604 looks good but it's PITA to have both params used in our functions 18:21 ShadowBot https://github.com/minetest/minetest/issues/5604 -- lua_api: Correct craft recipe param `type` to `method` by DS-Minetest 18:21 Krock would rather like to see a fix for the functions themselves first before starting with the documentation 18:21 benrob0329 #5586 is light stored in a signed int? 18:21 ShadowBot https://github.com/minetest/minetest/issues/5586 -- Add negative value for node light_source. 18:22 Krock unsigned byte 18:22 benrob0329 Oh, nvm 18:23 benrob0329 How hard would it be to change to a signed byte? 18:24 sfan5 hard 18:24 benrob0329 is light source seperate from light level? 18:25 paramat 'stored' in what way, in the world database or as a light source parameter? 18:25 sfan5 doing this kind of change requires deep knowledge of the light system, which you don't seem to have; sorry 18:27 Krock signed byte is nonsense, regarding how we store the light values right now: 4 bits for daylight, 4 bits for at night, thus light max = 15 18:27 Krock alone in the backwards compability would be a challenge (to not forget anything) 18:28 DS-minetest (also negative light is unrealistic) 18:28 Krock > unrealistic > running with 20 n/s 18:29 paramat mapnode light levels are positive, so using signed ints would mean we are using twice as much memory as we need to 18:29 Krock btw, int16 or int32? 18:31 benrob0329 sfan5 true, I was thinking a negative source would simply subtract said light rather than add it, but from reading light.cpp/h this probably wouldn't be easy. 18:32 paramat nope the lighting method is built around positive values, negative ones would not automatically work 18:33 paramat so i'm closing the issue :] 18:33 benrob0329 The light value would remain the same, but light_source would change and extra clauses added 18:34 Krock that's not as simple as you may think 18:35 paramat closed 18:54 celeron55 the only change to the datatypes of the light system might be extending it to 16 bits total, but it might be completely hazardous to performance as light propagation is recursive and probably gets really expensive after something like 15 nodes 18:54 celeron55 the only remotely reasonable* 18:55 celeron55 well except for modifying of the meaning of the light banks a bit (which is already a PR?) 18:57 celeron55 (also, making mapnode larger from its current 4-byte size is probably also hazardous to performance) 18:58 celeron55 it would be nice if subgames or something could decide such performance tradeoffs but it's not really possible to implement anything like that 18:58 sfan5 should've used java 19:06 sofar juhdanad: TileLayer holes are OK, it's not so bad 19:06 sofar 2 holes of 6 bytes, no big deal 19:06 sofar it's 2 bytes larger than a cacheline though without the holes 19:06 sofar so maybe we can save 2 bytes and get it in one cacheline? 19:07 sfan5 is it the same for AMD cpus or are we optimizing for intel only? 19:07 Hijiri I don't think I've reviewbegged today yet: #5612 19:07 ShadowBot https://github.com/minetest/minetest/issues/5612 -- Allow overriding tool capabilities through itemstack metadata by raymoo 19:07 sofar it's all 64bytes afaik 19:07 paramat will merge #5623 in 2-3 hrs if no objections 19:07 ShadowBot https://github.com/minetest/minetest/issues/5623 -- Sneak glitch: Set default to false by paramat 19:08 sofar sfan5: even if it would be e.g. 32 bytes, then you'd still be 3 instead of 2 cachelines 19:08 sfan5 hm, and ARM-based cpus? 19:08 sfan5 indeed 19:08 sofar 66 bytes for a struct is just a big overhead for those last 2 bytes 19:08 sfan5 i guess no chip designer is weird enough to have odd-sized cache line 19:08 sfan5 s 19:08 sofar afaics arm is 64bytes cacheline as well 19:09 sofar cat /sys/devices/system/cpu/cpu0/cache/index0/coherency_line_size 19:10 sofar if we go to 16bit shader id's we can do it 19:11 sofar or 16 bit texture ids 19:12 paramat #5625 is trival mapgen stuff requested by zeno so will merge later once cleaned up and tested 19:12 ShadowBot https://github.com/minetest/minetest/issues/5625 -- Mgflat, Mgv7: Only create noise objects if needed by paramat 19:47 * Fixer .oO(please fix mod packs and mod loading in general) 20:22 Fixer I've noticed torches are no longer instamined, is that intentional? 20:23 sfan5 i don't think so 20:24 sfan5 caused by https://github.com/minetest/minetest/commit/858c72297409ca54092c3c908279d2a05fef58cd 20:25 Fixer added issue in mtg https://github.com/minetest/minetest_game/issues/1706 20:25 sfan5 it looks like instant dig is no longer possible after that commit is that correct? 20:26 Fixer i suspect that 20:28 Fixer since just few builds ago it was ok 20:28 sfan5 i think i even approved that pr 20:28 sfan5 hmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm 20:30 Fixer node is removed without tool wear after 0.15 seconds (torch) 20:30 Fixer 0.15 seconds? o_O 20:30 Fixer it is quite visible 20:30 sofar it's 10 frames 20:30 sofar at 60fps 20:30 sfan5 what did the removed code do 20:31 sfan5 then 20:31 Fixer it is quite annoying, since you need to hold the button a bit more than you really want %) 20:43 Taoki Wow... latest Git master has some very pretty mgv7 biomes. I assume the abstinence toward considering v7 incomplete is over by now :) 20:46 sfan5 mgv7 is default since last release 20:46 paramat happy you like it 20:46 red-NaN could someone review #5492? 20:47 ShadowBot https://github.com/minetest/minetest/issues/5492 -- [CSM] Add local particles and particlespawners. by red-001 20:47 paramat just need a new surface node and ferns for coniferous forest 20:48 nerzhul particlespawner constructor is just crazy xD 20:49 juhdanad paramat: for the /fixlight command, is "Resets lighting in the area between pos1 and pos2" a good description? 20:50 sfan5 red-NaN: can you please not fix the code style of entire files in your pr 20:50 sfan5 thanks 20:50 Taoki Indeed I do 20:50 Taoki Apparently however, Minetest constantly crashes 20:51 juhdanad Taoki: when does it crash? When you exit the game? 20:52 Taoki No... sometime randomly during the game 20:52 Taoki And it's not a Lua crash to the main menu... engine crash 20:52 nerzhul sfan5, as we don't change the particle code source very often, why not no ? :) 20:53 nerzhul a memleak was introduced today juhdanad 20:53 nerzhul in your PR 20:53 juhdanad I know, but I already fixed it, sorry. 20:53 nerzhul cppcheck showed me an error when i fetched before preparing my 5th PR on cppcheck performance fixes 20:53 nerzhul temp_buf->append(v, mesh_buffer->getVertexCount(), indices, 20:53 nerzhul mesh_buffer->getIndexCount()); ? 20:54 nerzhul is this fixed ? 20:54 juhdanad This is not, I will check. 20:54 nerzhul cppcheck reports 20:54 nerzhul [src/mesh.cpp:409]: (error) Memory leak: temp_buf 20:54 nerzhul [src/mesh.cpp:418]: (error) Memory leak: temp_buf 20:55 nerzhul in EVT_STANDARD you returned temp_buf, it's fine 20:55 nerzhul for others you return clone which is NULL because we never set it in the function 20:55 nerzhul i think it's not normal i think no ? 20:55 nerzhul also in this function the u16* case is unneeded 20:55 nerzhul mesh_buffer->getIndices() is already u16* 20:56 nerzhul please send us very fast a fixing PR 20:58 juhdanad I'll try. 21:02 Taoki nerzhul: Is that related to my crash too? 21:02 juhdanad If you use tangent vertices, probably yes. 21:02 nerzhul i'm pretty sure 21:02 nerzhul juhdanad, you hsould never return null i think ? 21:03 juhdanad Yes, I intended to return at the end of each case. 21:03 Taoki tangent vertices? Is that an option? 21:03 Taoki But in any case, gladly awaiting a fix 21:05 juhdanad nerzhul: how can I enforce Minetest to use tangent vertices? 21:06 nerzhul no idea, use taoki world ? 21:07 nerzhul if i look at the code 21:07 nerzhul before temp_buf->drop(); was called 21:07 nerzhul and dst_mesh added temp_buffer 21:08 nerzhul in every case 21:08 nerzhul and we returned clone 21:08 nerzhul with your pr 21:08 nerzhul dst_mesh was renamed clone 21:08 nerzhul we never drop 21:08 nerzhul and in the working case we return temp_buf 21:08 nerzhul i think we should return temp_buf everytime and remove clone 21:08 nerzhul or re_use dst_mesh with addMeshBuffer 21:09 nerzhul juhdanad, https://lut.im/3LkJxDGyx4/we6bLDzXdltoTizg.png 21:10 nerzhul Taoki, you can modify code right ? 21:10 nerzhul and you crash is systematic, right ? 21:10 Taoki Depends on what code. And it's random but consistend, happens unpredictably 21:11 nerzhul it should be triggered on specific event using those vertices, how many time (mean) before it happens ? 21:12 Taoki Not sure what you mean by times. It happens sometimes. 21:12 Taoki Although I think it happened twice in the same point once 21:13 juhdanad Okay, I enforced tangent vertices and I see the crash. Now I'll attempt to fix... 21:13 juhdanad Taoki: probably an item's mesh causes the crash. Or a special node on the map. 21:14 Taoki Makes sense. Glad you found a way to test 21:14 nerzhul juhdanad, first remove clone and return the temp_buf (and rename it to another name if it's not temporary, or restore prvious behavour) 21:14 Taoki I've been away from Minetest for a bit like I said... would have been harder, as I might have forgotten a few things 21:15 nerzhul nore, be careful you merged this , i know the PR was very huge :( 21:15 nerzhul when i will finished with CPPcheck fixes i think i will provide a PR to add a cppcheck verification to show the static errors found, without this a random crash + memleak could happen 21:16 nore nerzhul: the soft node overlay? it was not me who merged it... 21:16 Taoki 777 MB of RAM. I don't think Minetest should be using that much... 21:16 Taoki So might confirm the memleak as well 21:17 nerzhul oh sorry nore, i misread github log you have same colors as juhanad :p 21:18 nore lol :D (although not exactly, mine is more pink :p) 21:18 nore no problem anyway :) 21:19 nerzhul i will push #5635 after travis build pass, it will fix the master LINT problems (i think a merged PR branch was before lint) + some code style errors i introduced in previous cppcheck pr + some memory copy fix + constructor fix etc 21:19 ShadowBot https://github.com/minetest/minetest/issues/5635 -- Fix various performance issues reported by cppcheck + code style (CI) by nerzhul 21:20 nerzhul and some ugly spaces added by juhdanad :p 21:20 nerzhul "return dst_mesh; " 21:25 benrob0329 shouldn't clearobjects only remove item entities, and a clearentities command be added? 21:25 juhdanad nerzhul: #5636 21:25 ShadowBot https://github.com/minetest/minetest/issues/5636 -- Fix a memory leak by juhdanad 21:26 nerzhul assert before return if it should not happen 21:26 juhdanad You mean 'assert(false)'? 21:26 nerzhul also remove break after return, it's unreached 21:26 nerzhul yes 21:26 juhdanad Or rather with sanity_check? 21:26 nerzhul sfan5, ? 21:27 nerzhul what do you think ? 21:27 nerzhul i think it's a pure devel error, it should not be reported to user in release mode ? i don't see how can that happen 21:28 nerzhul you can use sanity_check yes, why not 21:32 sfan5 return null sounds fine to me 21:32 nerzhul sfan5, you prefer a crash when using the object than an assertion for a unhandled switch case ? 21:33 nerzhul i don't really know where this pointer will finish but it can be difficult to track 21:33 sfan5 then throw in a sanity_check(false); 21:33 juhdanad cloning should be always successful. 21:34 nerzhul yeah 21:34 nerzhul guys i found a very interesting Lua call ! 21:34 nerzhul https://github.com/minetest/minetest/blob/master/src/script/lua_api/l_server.cpp#L488 21:34 sofar I've lost internet at home 21:34 nerzhul maybe we can remove it no ? 21:34 nerzhul i don't see any interest to have a lua call to trigger segfaults, zerodivision and exception 21:35 nerzhul sapier told: "Catch some error situations when mod used without thinking about it" 21:35 nerzhul wtf 21:35 nore nerzhul: it's only these in debug builds and I guess it's there to check if error handling works well, probably? 21:35 juhdanad nerzhul: updated. 21:36 nerzhul nore i don't see how you want to handle a segfault :p 21:36 nore ehm 21:36 sfan5 nerzhul: that function looks useless 21:36 nore yeah, maybe remove it 21:36 nerzhul sfan5, i will push a commit to remove it 21:36 sfan5 and clang does actually not allow you to cause a segfault by deref'ing a NULL pointer 21:37 nerzhul exact 21:37 nerzhul nore 21:37 nerzhul note* 21:37 nerzhul it's called in games/minimal/mods/errorhandler_test 21:37 nerzhul i remove it ? 21:37 nerzhul it's related to /exceptiontest in minimal 21:37 sfan5 yes 21:38 nerzhul okay i will provide a trivial pr for this 21:39 Taoki Found another bug BTW... 21:39 Taoki Hit a mob once: Their model becomes white, but doesn't go back. Every time you hit an entity and damage it, it becomes whiter until you reconnect. 21:41 Hijiri are you sure it's not a bug with the mod? 21:42 Taoki Don't see how it could be, so yeah. I don't colorize entities or anything 21:42 Taoki Although I used colorized mixed textures 21:43 Taoki Appended with ^ in the texture string. But that's all 21:43 Hijiri what does the code look like? 21:43 Hijiri oh 21:43 Hijiri no punch code? 21:43 Hijiri that does sound like an engine bug then 21:44 juhdanad I see it too, will investigate... 21:45 Taoki Other than that, the glass_framed drawtype does some z-fighting, when you see the backface against a wall. That's really unimportant though, but just thought to mention 21:45 nerzhul sfan5, sofar #5637 for this change 21:45 ShadowBot https://github.com/minetest/minetest/issues/5637 -- Remove lua crash triggering by nerzhul 21:49 paramat juhdanad that seems ok to me 21:49 nerzhul juhdanad, you should rebase :) 21:49 * sofar kicks 45 min internet outage 21:50 nerzhul and i will merge it after CI pass after rebase 21:51 juhdanad nerzhul: you only removed (u16*) casts, right? 21:52 sofar juhdanad: #5638 21:52 ShadowBot https://github.com/minetest/minetest/issues/5638 -- Reorder TileLayer. by sofar 21:52 nerzhul yes 21:53 sofar juhdanad: ideally, we shrink TileLayer 2 bytes, though! 21:54 juhdanad sofar: sorry for not responding so long, there are a lot of bugfixes to do now... 21:54 paramat Taoki tangent vertices is enabled when you use bumpmapping/normalmapping/parallax occlusion shaders 21:54 Taoki paramat: Oh... I use those in that case 21:55 nore sofar: I'll try to modify it not to copy things that are not needed 21:55 sofar juhdanad: don't worry 21:55 nore (i.e. you don't need to copy the whole vector, and copying a vector is probably expensive even if there is nothing in it) 21:56 sofar nore: it's gonna be hard, maybe shrink shader_id to u16? but that comes from the protocol I think 21:56 nerzhul copy an empty vector is not expensive as memory area is less than 8 bytes :) 21:56 nore nerzhul: vector is already 24 bytes long, and I think it allocates another area for the new vector 21:56 nore I mean the vector struct is 24 bytes long 21:57 juhdanad nerzhul: rebased. 21:57 nore including a pointer to data 21:57 nerzhul nore, can you show me the code line ? 21:57 nerzhul juhdanad, nice, i will merge it after CI 21:57 nerzhul but CI seems very busy 21:58 nore nerzhul: hmm, it's the animation vector is TileInfo that is now TileLayer I think 21:58 nore *in 21:58 juhdanad sofar: the alpha bits of the layer's color is not used currently. 21:58 juhdanad That's 1 byte. 21:58 sofar we need 2 bytes, and alignment :) 21:59 nore alignment is more difficult 21:59 nerzhul sofar, does that change really something at this moment as we have many other areas to improve ? :p 21:59 sofar yes, this is a really hot struct in the client 22:00 nore but anyway I'll try to modify the code to make it copy only what is needed, I think it would get the hot part to 20 bytes only 22:00 nore so 2 of them would even fit in a single cache line 22:00 nerzhul is the vector size dynamic or static ? 22:01 nore (well, 3 if we didn't have these alignment restrictions) 22:01 nore nerzhul: doesn't change for a fixed TileLayer, but size depends on the TileLayer 22:02 nore (i.e static as does not change within each instance) 22:08 juhdanad nerzhul: I've found the reason why entities become continuously brighter when you punch them: it's a const reference. 22:08 Fixer sfan5: funnily enough i have impression that tool dig speed was increased, for example diamond shovel or pickaxe 22:09 nerzhul juhdanad, can you be more precise 22:10 nerzhul i don't see why a const ref can trigger that 22:10 juhdanad When some time has elapsed, the entity wants to return to its original texture. 22:10 juhdanad Therefore, 'updateTextures(m_previous_texture_modifier)' is called. 22:10 nerzhul okay, and ? 22:11 juhdanad Then it does: m_previous_texture_modifier = m_current_texture_modifier; 22:11 nerzhul const ref just prevent copying the texture modified but it in updateTexture it's copyied, it's done 22:11 juhdanad And m_current_texture_modifier = mod; 22:11 nerzhul sorry but not relevant 22:12 nerzhul mod is a reference to m_previous 22:12 juhdanad Where mod is a reference to m_previous_texture_modifier. 22:12 nerzhul and m_current_texture_mod = mod; copy the reference content 22:12 juhdanad Bot mod was overwritten in the previous command. 22:12 nerzhul m_previous_texture_modifier = m_current_texture_modifier; 22:12 nerzhul m_current_texture_modifier = mod; 22:12 nerzhul i think the problem is that 22:13 nerzhul no ? 22:13 nerzhul oh okay 22:13 nerzhul i see 22:13 nerzhul exact 22:13 nerzhul we don't copy it, and we use it without swapping 22:13 nerzhul anoying side effect 22:14 juhdanad So the parameter there is not const, because you have a non-const pointer to it too. 22:14 nerzhul i'm thinking about it, wait a minute :) 22:14 juhdanad (at content_cao.cpp:1325) 22:14 nerzhul if (mod != m_previous_texture_modifier) 22:14 nerzhul m_previous_texture_modifier = m_current_texture_modifier; 22:15 nerzhul that should do the trick 22:15 nerzhul can you try it ? 22:16 juhdanad Well, it works, but I don't understand yet why. 22:17 nerzhul with my fix ? 22:17 juhdanad Yes. 22:17 nerzhul it's simple, if you set previous with current, when you set current to mod, current = previous = current => no change 22:18 nerzhul but it's not the right fix i think 22:18 nerzhul copy the string solve the problem, but it's not very elegant to cut the previous override, because previous is not current then 22:19 juhdanad No. There are three different values there: the previous, the current and the next (mod) texture modifiers. 22:19 juhdanad *No, it is not the right fix. 22:20 nerzhul i'm searching if it's possible to keep the const ref benefit 22:20 nerzhul i see how it's possible, jsut copy the variable before usage 22:21 juhdanad But if you copy it, there is no point in sparing a copy. 22:21 nerzhul in 2 usecase on, in 3 usecases yes 22:21 nerzhul 3 vs 3 22:21 nerzhul then, rollback this sounds the only correct solution 22:22 nerzhul very tricky bug 22:23 nerzhul or we can do better 22:23 juhdanad Maybe keep the const ref, but copy the string when you call the function with the member! 22:24 nerzhul yeah i think about that 22:24 nerzhul but no real benefit, another copy 22:25 juhdanad No, there is a call 'updateTextures(m_current_texture_modifier + "^[brighten")' 22:25 nerzhul yes u know, there is 3 calls like this, and three cases to member variables 22:26 nerzhul https://pastebin.com/KckCCDr0 ? 22:26 nerzhul swap if we are working on members, else do the regular way 22:26 sofar https://gist.github.com/sofar/21b8d3b70a83103c9c0cb9bf59bbaf21 <== all Pahole output of minetest 22:27 nerzhul sofar, i you want to pahole all minetest i think i will pass clang-format on all files, as it will do same thing on PR :p 22:27 nerzhul (joke) 22:28 nerzhul sofar, don't the compiler reorder things for CPU ? 22:28 juhdanad nerzhul: looks good to me. 22:28 nerzhul juhdanad, you tested it ? 22:28 juhdanad Not yet, wait... 22:28 nerzhul i'm waiting for travis on #5637 no problem :p i will merge it in 5 mins, travis will finish 22:28 ShadowBot https://github.com/minetest/minetest/issues/5637 -- Remove lua crash triggering by nerzhul 22:28 sofar nerzhul: this is from the binary, and no, reordering doesn't happen except for initializers 22:29 nerzhul sofar, it's strange that compiler doesn't try to make binary more efficient for their architecture 22:32 juhdanad nerzhul: the first line should be: 'if (mod == m_current_texture_modifier) {', shouldn't it? 22:32 juhdanad Unfortunately the patch doesn't work. 22:33 nerzhul okay, i will just revert to a copy, we will see later 22:34 sofar ItemDefinition is 6 cachelines of which 2.5 cachelines is just c_strings[32] 22:35 sofar ContentFeatures is huge, lol 22:35 nerzhul ContentFeature is a huge thing yes 22:35 nerzhul i'm pushing a const ref revert on updateTextures 22:36 sofar /* size: 4272, cachelines: 67, members: 57 */ 22:36 sofar 67 ... 67 cachelines 22:36 sofar it wastes half a cacheline on paddings and stuff 22:37 juhdanad sofar: when I write f.light_source, then all 67 cachelines are loaded? 22:37 sofar no, it doesn't work like that 22:38 sofar it'll load the part of the struct that isn't in cache, in 64byte segments 22:38 sofar for what is needed 22:38 nerzhul sofar, ContentFeature is not very often updated it's write once in majority cases 22:38 juhdanad Okay then! Fortunately ContentFeatures are passed as references. 22:38 nerzhul juhdanad, i pushed the revert on master for this const ref 22:39 nerzhul juhdanad, it was not the case everywhere 2 years ago :p hopeful Zeno` and me worked on removing that :p 22:39 nerzhul this struct is very dangerous i think the copy constructor should be removed if possible when using C++11 to prevent bad usages :) 22:39 juhdanad I'm sure Minetest became significantly faster when you did that. 22:39 sofar nerzhul: is ContentFeatures read a lot in hot (render/mesh update) paths? 22:40 juhdanad I only know it is read a lot at light updates. 22:40 sofar the trick isn't to make it smaller per se 22:41 juhdanad Oh, yes, mesh update also reads content features! 22:41 sofar the trick is to put the hot items in it all together in a single cacheline 22:41 sofar suppose that mesh and light update only need 10-15 items 22:41 sfan5 nerzhul: i think some of your fixed introduced this warning http://sprunge.us/YFQM 22:41 sofar and now they're in 13 of the 67 cachelines 22:41 sofar maybe we can make that 2 cachelines 22:42 sofar and reduce memory pressure by a factor 5+ 22:42 nerzhul sfan5, why my compiler doesn't report it... 22:42 sfan5 clang 4.0 on arc 22:42 sfan5 h 22:42 sofar which change? 22:42 nerzhul i think my clion uses gcc atm instead of clang 22:43 juhdanad Good night to everybody! I hope I successfully cleaned up everything after myself. 22:43 nerzhul it was reseted when testing jsoncpp defaults heh, i switched, it's better now 22:43 nerzhul okay i see, a cleanup to remove 22:43 nerzhul oh no, just a const ref to re-link :) 22:44 nerzhul it also permits to show gui::IGUIElement* m_parent; in guiFileSelectMenu warning, nice 22:44 nerzhul lol 22:45 nerzhul compilers are crazy 22:45 nerzhul if you set m_parent(parent) in constructor it reports m_parent as unused 22:45 nerzhul and if you do m_parent = parent in constructor it doesn't 22:46 nerzhul i'm testing it and i will push the double warn fix 22:46 sofar nerzhul: `class Client` 22:46 sofar watch this 22:46 sofar you won't believe this: 22:47 sofar /* size: 134088, cachelines: 2096, members: 69 */ 22:47 sofar /* sum members: 2045, holes: 16, sum holes: 940 */ 22:47 sofar lmao 22:47 sofar none of it is hot, though, I think 22:47 nerzhul rah cmake reseting RUN_IN_PLACE when changing compiler 22:48 nerzhul client is central but env is more used than client in terms of attributes 22:48 nerzhul also client has a ClientEnvironment object inside it 22:49 sofar that one actually sounds like a pahole bug 22:52 nerzhul okay pushing a trivial fix for warning + guiFileSelectMenu header clean on unused attributes (never called anywhere) 22:53 nerzhul pahole is not simple i think :) 22:54 sofar using it right is hard 22:54 sofar we stumbled on TileSpec being bad due to profiling 22:55 nerzhul i also merge #5636 (travis is yellow, but C++11 build will finish soon and is not concerned by this change). 22:55 sofar if the problem doesn't show in profiling it's probably not so bad 22:55 ShadowBot https://github.com/minetest/minetest/issues/5636 -- Fix a memory leak by juhdanad 22:56 nerzhul i'm off, see you tomorrow guys, thanks for your help, and good luck sofar but don't forget we have 180 PR now (argh) 22:56 nerzhul i don't know which project have the most PR/issues but: https://github.com/ansible/ansible >1000 pr :p 23:01 Fixer !tell nerzhul I wonder if your building system can be used for making flatpaks... https://forum.minetest.net/viewtopic.php?f=42&t=15336 23:01 ShadowBot Fixer: O.K. 23:56 paramat +1 for #5630 will merge later 23:56 ShadowBot https://github.com/minetest/minetest/issues/5630 -- Add /fixlight chat command by juhdanad 23:58 VanessaE so it does something similar to worldedit's `//replace air air` trick? 23:59 nore VanessaE: much more efficiently