Minetest logo

IRC log for #minetest-dev, 2017-03-24

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

All times shown according to UTC.

Time Nick Message
00:03 red-001 you know there is a PR to re-add the sneak elevator?
00:04 OldCoder red-001, there are multiple proposals, if you're talking to me. But I'm addressing the larger picture.
00:04 OldCoder Not just the sneak thing
00:04 red-001 just letting you know
00:04 OldCoder Thanks
00:04 OldCoder Thoughts are appreciated
00:04 red-001 since thats the url, I haven't finished reading it yet
00:05 Billre where do you want the thoughts?
00:05 red-001 OldCoder, whats your option on CSM?
00:06 OldCoder Billre, IRC PM is fine
00:06 Billre ok
00:06 OldCoder red-001, I've been editing that document for about... Hm. 4 hours.
00:06 OldCoder Need to take a short break. Back in a bit.
00:06 red-001 sure
00:10 paramat my comments on the doc: all changes affect the feel of a game, trying to draw a line somewhere for +5 is not practical. however, i admit i probably merged it too quickly and more expressed approval may have been needed. it was +4 -2, but more devs may support the intention but have just not expressed that, so the number approving may be higher than 4
00:11 paramat we need more devs to state their opinion on the PR's intention
00:11 paramat also i suggest not bothering editing such a doc, just everyone discuss in this channel
00:14 paramat are people's comments edited? do you have their permission to publish PMs?
00:18 fireglow joined #minetest-dev
00:18 fireglow[m] joined #minetest-dev
00:22 twoelk OldCoder, as you used html for the document you could have provided links to the quote sources so that one could study the context things were said in.
00:22 rubenwardy twoelk: was in PMs
00:23 twoelk ok, still same
00:24 twoelk nvm then
00:25 paramat joined #minetest-dev
00:28 paramat .. i'm not saying it should be reverted, we are now moving towards the best possible situation: undesirable bugs fixed and the popular side-effects kept as optional and intentional features
00:30 red-001 ^
00:30 twoelk I'm not really convinced it was a bug, it had an inner logic, an ingame logic not a model of reality.  :-D
00:31 twoelk and it was somewhat unique
00:32 * twoelk goes of to climb some sneak ladders on some obscure servers
00:54 OldCoder twoelk, the quote sources were IRC PM with permission; there are no links
00:55 twoelk was told, so nvm the comment
00:55 OldCoder paramat, red-001 As I said in the HTML, not wading into this based solely on the PR. sofar and I are close to agreement... the real issue is how to have more of a sense of a team
00:56 OldCoder One reasonable suggestion is that certain types of changes should need higher +approval in the face of opposition
00:56 OldCoder and definitely more discussion
00:56 twoelk maybe a devplayonserverday?
00:56 OldCoder Yes
00:56 twoelk like once a month or so
00:56 OldCoder and at the same time perhaps lesser changes should be easier to get into core
00:56 OldCoder Ones that are acknowledged not to be dangerous
00:57 twoelk milage may vary
00:57 OldCoder red-001, did you ask about CSM? if so, identify acronym
00:58 OldCoder all, I haven't talked about a spoon or other utensil because of sneak. I just think the above: some changes should be handled more carefully and others should be easier
00:59 rubenwardy OldCoder: client side modding
00:59 OldCoder Consensus may not be the absolute rule, but look at what people care about, go easy on stepping on a game that they may love, and allow little tweaks into the core if they don't damage things
00:59 OldCoder rubenwardy, go on
00:59 OldCoder I'm talking specifically about core dev
00:59 rubenwardy !g minetest client side modding good or bad site:forum.minetest.net
00:59 ShadowBot rubenwardy: [Mod] MAPP - map for Minetest - optimized! [0.2.4] - Minetest Forums - It is a shame when good mods are discontinued and forgotten (like this mod and the mods Mauvebic made.) It would really be great if this could be done client-side. The client has most of the data already. (A. Savage, Mythbusters) I'm not modding and/or playing minetest anymore. (9 more messages)
00:59 OldCoder rubenwardy, go on?
01:00 twoelk https://github.com/minetest/minetest/issues/5394
01:00 OldCoder R
01:00 rubenwardy !title https://forum.minetest.net/vi​ewtopic.php?f=47&t=17074
01:00 ShadowBot rubenwardy: Client-sided modding: Good or bad? - Minetest Forums
01:00 OldCoder Ahhh
01:01 twoelk texturepacks are allready csm ;-P
01:01 * OldCoder was thinking context-sensitive migration or something
01:02 OldCoder Well, it would permit the distribution of tailored clients, right?
01:02 OldCoder As well as worlds?
01:02 OldCoder Would complicate things, in some cases, if different players had different modsets
01:02 OldCoder But what argument could there be to rule out such a feature?
01:02 OldCoder It adds to the power of the platform
01:03 twoelk security problems?
01:03 OldCoder Right
01:03 OldCoder But almost anything can be misused
01:03 betterthanyou711 hi
01:03 OldCoder betterthanyou711, you are +1 older O_O
01:03 OldCoder People can recompile clients and do what they wish already
01:03 twoelk also not all clients are the same, some are potatoes as said somewhere else
01:03 betterthanyou711 +1 older requires a certain age to be told
01:04 OldCoder I don't buy security as a reason to skip the feature. Maintainability perhaps.
01:04 betterthanyou711 or ye old to fold like the mold on bread?
01:04 OldCoder If the API stabilizes
01:04 OldCoder betterthanyou711, PM
01:07 OldCoder red-001, I don't know enough yet to comment authoritatively. If it is difficult to stabilize and maintain, I might have reservations. In general it sounds like another useful tool to have in the box.
01:08 paramat more approval required for larger or more controversial changes is already how we work, i felt +4 was enough, but yes merged to soon
01:08 OldCoder Progress, then. See the flip side. Don't be so strict about lesser things but *yes* I understand caution on the flip side. Things that are option-able should be less controversial. Options are good.
01:09 OldCoder If this change was safely option-able there would be no issue. Regrettably this apparently was not the case for the 1st version.
01:10 OldCoder But my own remarks, again, are not specifically about the PRs. I wouldn't spend whatever capital I have on that. This is about the larger picture.
01:10 OldCoder The core devs are skilled, investing time, do not take orders, all of this is understood... at the same time the project as a whole needs to be seen as more of a team
01:11 paramat yeah, more testing was needed and more effort made on keeping an option, i did actually ask for more effort to address the complaints
01:11 OldCoder
01:12 OldCoder C55 did hope that the core devs would, at a higher level, spend more time in the other channel. This seems like a sensible point.
01:12 OldCoder Communication, when positive, is rarely a bad thing
01:12 paramat and please consider not posting blank lines, it's irritating
01:12 OldCoder paramat, it's part of me. Not conscious.
01:12 OldCoder And less irritating than other things that are accepted
01:12 OldCoder Plus, it's logical
01:13 * OldCoder has never understood why *three* reasons for separators are not sufficient
01:13 twoelk mileage may vary ;-P
01:14 paramat some lines of yours in that document are unreasonable and untrue in the way they portray devs
01:15 twoelk well it is personal opinion
01:15 OldCoder paramat, they are perceptions that many hold
01:15 OldCoder as were those in the charter for #minetest-project
01:15 OldCoder The fact that the perceptions are broadly held ought to be understood
01:15 OldCoder C55 himself holds them :-)
01:15 OldCoder Or some of them
01:16 OldCoder <celeron55> but keep in mind that it really seems core devs are kind of alienated from the way of thinking that you and I do
01:16 OldCoder paramat, ^ :D
01:16 OldCoder paramat, response to that is welcome!
01:16 paramat i'm seeing quite a few overly-negative perceptions over this
01:16 OldCoder <paramat> Surely the speaker is mistaken. Oh, wait, he's the one who created the project.
01:17 octacian joined #minetest-dev
01:17 twoelk OldCoder that feeling of them and us is not unique to mt-deevs and players, it seems to come naturally
01:18 twoelk -e
01:18 twoelk what is needed is suggestions and mechanics how to keep the alienation as little as possible
01:19 paramat as i wrote before i feel things are not quite as bad as c55 is suggesting
01:19 Zeno` why are the perceptions overly negative?
01:20 twoelk because no gains of the fix are explained well?
01:20 Zeno` Also, 4 devs approved that PR? Which 4 devs are those?
01:21 paramat just a few comments by some that suggest devs 'do not listen at all' or similar
01:21 OldCoder paramat, fair enough
01:21 OldCoder and twoelk ditto
01:22 OldCoder Core devs might spend more time in -project and understand that -project should rank slightly higher than random GitHub posters
01:22 OldCoder -project was created for this purpose. It is not a modding channel per se.
01:22 paramat 3 official approvals and 1 approval of concept, so i guess +4 for concept, +3 for code and concept
01:22 twoelk erm, -project is what it became
01:23 twoelk a nice place
01:23 OldCoder It is a place the platform as a whole to be discussed, for common ground to be reached
01:23 OldCoder twoelk,  it was designed intentionally to be a nice place. And a steering committee.
01:23 OldCoder This was its charter.
01:23 OldCoder Additionally, suggestion is for a tiered approval system. Dangerous or controversial changes, up the bar.
01:24 OldCoder And less dangerous or controversial changes, especiall those that are option-able, lower it.
01:24 OldCoder Three suggestions, therefore, looked at this way. Are they unreasonable?
01:24 twoelk so techical stuff in -dev and steering-chatter in-project?
01:24 OldCoder twoelk, technical is fine in -project but it is less formal than -dev which is intended to be solely technical
01:25 OldCoder We do respect the desire to focus in this channel
01:25 OldCoder but perhaps the platform as a whole and directions could be discussed more with the group as a whole - which is, remember, vetted
01:25 OldCoder -project is moderated. -dev IIRC is not.
01:26 OldCoder So, I have those three suggestions, I think. sofar started out as pretty blunt and firm but I think he's leaning towards agreement.
01:27 OldCoder Note the 3rd change. It shouldn't be too difficult for amateurs to get their changes into core. Especially if they're minor and option-able.
01:27 OldCoder There are risks involved but not so high and this is a way for the codebase and the team to grow
01:27 OldCoder
01:29 paramat we already scale required approvals for PRs
01:29 twoelk actually I don't find the way minetest is developing that bad. Only thing I would suggest is to call the next release 0.5.0 for all the map/mod/gameplay changing stuff
01:30 OldCoder 0.5.0 seems a separate discussion
01:30 twoelk or rather breaking than changing
01:30 Shara In fairness, I don't think getting minor changes in via PR is too difficult. It's the big controversial issues specifically that  leave me with a headache, when it seems like there are extreme views and no middle grounds on offer.
01:31 VanessaE I agree with twoelk
01:31 paramat #5440 is ready and tested woo
01:31 ShadowBot https://github.com/minetest/minetest/issues/5440 -- Map generation limit: Make per-world by paramat
01:31 VanessaE you can't legitimately call it 0.4.x anymore I think
01:31 OldCoder paramat, it wasn't really understood that this change would be as controversial as it was... that is clear. If it was seen as one minor voice objecting... maybe that voice should have been listened to and the matter reviewed further.
01:31 OldCoder 0.5.0 is here?!
01:31 twoelk it would make clearer that rather sigificant changes are to be expected
01:31 VanessaE OldCoder: only if the others here think it is.
01:31 OldCoder Not backwards incompatible, I gather?
01:32 OldCoder VanessaE, and not here but -project
01:32 OldCoder That is one of its points
01:32 VanessaE I won't speak to that.
01:32 OldCoder Here, discussion may tend to be circular at times, small group, no experience with worlds
01:32 OldCoder You - You personally know exactly what I'm referring to
01:32 VanessaE minetest has changed enough in various ways that it does make sense to call it 0.5.0 now.
01:32 twoelk old clients won't see all colors for example
01:32 OldCoder Well, that change at least seems not too controversial
01:33 VanessaE but that's immaterial to the sneak issues.
01:33 twoelk security breaks many old mods
01:33 OldCoder VanessaE, I'm not very good at mixing threads if things aren't clear - 0.5.0 is not germane to sneak? Yes, you are correct.
01:33 OldCoder Or colors are not relevant?
01:33 rubenwardy Server modding was added in 0.4. Client should be 0.5, imo
01:33 VanessaE OldCoder: nevermind.
01:34 VanessaE rubenwardy: +1
01:34 paramat we knew it would be controversial, objections were listened to, but yes more work was needed (which is being done now)
01:34 OldCoder paramat, thank you for your remarks
01:34 OldCoder My own suggestions are general and you and sofar both seem to feel they are reasonable - or 2 of them; you have not commented, I think, on #-project
01:34 paramat i agree with 0.5 btw
01:35 OldCoder Perhaps consider spending more time there
01:35 paramat heh i spend a lot of time there
01:35 OldCoder Very well
01:35 OldCoder Good for you to receive positive feedback as well as negative, you know
01:35 OldCoder You know how much people enjoy your work
01:36 paramat 0.4 was a new default mapgen, we've just changed that
01:36 OldCoder Another mapgen, then?
01:36 OldCoder Oh well
01:36 * OldCoder prepares for territory breaks
01:36 OldCoder And old one is not option-able, I assume?
01:36 paramat i mean mgv7 is another reason for 0.5
01:37 OldCoder Hm? v7 is like 2 years old, isn't it?
01:37 paramat more, but only now default and reasonably complete
01:37 twoelk OldCoder just the default is changed - if v6 is declared it wont change in running maps if I understand correctly
01:37 OldCoder v7 is default and has evolved... Plus it has been 0.4 for about 4 to 5 years, right?
01:38 OldCoder twoelk, right, but v7 will have territory breaks now, right? Which is fine
01:38 OldCoder Good to know about it
01:38 OldCoder Hm. Compare to Firefox version numbers.
01:38 OldCoder <Mozilla> We must seem cool. 1, 2, 3, 20, 25, 40!
01:38 OldCoder <Minetest> I am more sedate. 0.4.1, 0.4.2, ...
01:39 OldCoder <Minetest> Hm. After 4 to 5 years, time for 0.5.0
01:39 QwertyDragon joined #minetest-dev
01:40 twoelk interesting thing is 0.5.0 wasn't planned, enough breaking stuff just seems to have piled up imho
01:41 OldCoder Makes sense. Dot .0 is for breaking things.
01:41 OldCoder But network protocol and maps will remain backwards compatible?
01:41 VanessaE don't see why not.
01:42 VanessaE as compatible as they can be considering mods put new features to use that permanently change a map
01:42 OldCoder Then 0.5.0 on with the show no tale of woe the new era is here pour a beer
01:42 twoelk if a map is opened with new server it cannot be read by old server iirc
01:42 OldCoder twoelk, OK
01:42 VanessaE there's that, too
01:42 OldCoder But will it work with old clients?
01:43 VanessaE as long as the client is new enough to render whatever features are in use.
01:43 twoelk except that they are colorblind :-D
01:43 OldCoder So, there is a colors change
01:43 VanessaE like the param2 color feature, can't do that with any client prior to 20170123
01:43 OldCoder And can't move worlds back and forth between old and new servers
01:43 OldCoder But can connect and play
01:44 VanessaE yeah.
01:44 OldCoder Hm. Chance to break the Android clones. But they'll catch up.
01:44 VanessaE perfect.
01:44 VanessaE there it is guys
01:44 VanessaE the perfect reason to call it 0.5.x :D
01:46 Zeno` backwards compatibility doesn't really come into the equation when talking about semantic or semi-semantic version numbering
01:46 OldCoder Zeno`, how so?
01:46 Zeno` just because 0.5 still mostly supports 0.4 doesn't mean that 0.5 should never have had the version bump
01:47 Zeno` e.g. photoshop 10 can still open photoshop 8 stuff
01:47 Zeno` doesn't mean that photoshop 10 should not be called photoshop 10. It should be called photoshop 10
01:47 OldCoder Zeno`, Hm? Sure.
01:47 OldCoder No argument
01:48 * OldCoder is mostly amused that MT has gone +.1 for years while Mozilla and others think it matters to go +1 or even more
01:48 twoelk the idea is that old servers will not be able to open new maps - they are forced to update
01:48 OldCoder Or just run their existing maps
01:48 OldCoder If old clients can connect up to a point to new worlds
01:48 OldCoder that is backwards compatible
01:49 twoelk then you end like redcrab
01:49 OldCoder Let's avoid THAT O_O
01:49 OldCoder Not quite that bad
01:49 OldCoder In fact this is the opposite in some respects
01:49 OldCoder I'm not hearing serious compatibility issues yet
01:49 OldCoder New servers can run old worlds
01:50 OldCoder Old clients can be used - up to a point - with new worlds
01:50 OldCoder This is most of what I'd ask for, right there
01:50 OldCoder #1 is the only crucial part...
01:50 ShadowBot https://github.com/minetest/minetest/issues/1 -- GlowStone code by anonymousAwesome
01:51 twoelk lol
01:51 VanessaE heh
01:51 OldCoder Should 0.5.0 switch to proller's network code and will it fix the basic problem?
01:51 OldCoder Anonymous Awesome?
01:51 OldCoder Good nick
01:56 twoelk btw <If this doesn't happen, ultimately, I'll launch a spoon,> does sound somewhat like a threat, and the rest near there like you assemblying some troops.
01:56 twoelk might not have been the intention
01:57 OldCoder twoelk, it's a simple consequence
01:57 OldCoder And the troops were and are assembled; it isn't my doing
01:57 OldCoder My goal would have been, or will be, to keep the project together to the extent possible
01:57 OldCoder Hence the term spoon as opposed to fork
01:58 OldCoder It's about finding the right balance. This isn't a corporate project - the core devs are volunteers and need to be respected on that basis.
01:58 OldCoder But may possibly be slightly out of touch, in some cases, with the rest of the team and world owners and players in particular
01:58 twoelk that's why I feel the tone in that phrase might be a little too threatening
01:59 OldCoder "threatening" ? This sounds like a discussion for #-project or PM as you prefer... It is what it is.
01:59 VanessaE there aren't any network changes of that degree, I don't think
01:59 OldCoder VanessaE, would 0.5.0 be the right time to switch to proller's code or was it rejected as unstable or incomplete?
01:59 OldCoder It's a major decision as it breaks old clients completely...
02:00 OldCoder But map backwards compatibility (and major mods, I think)
02:00 OldCoder is/are the only essentials
02:00 VanessaE I don't know if it was rejected or not, but now's not the time to think about that
02:00 OldCoder Not right at the transition to a major new branch?
02:00 * twoelk needs a multiscreen irc client to view more channels simultanely
02:00 OldCoder Isn't that when it would come in?
02:01 VanessaE imho save THAT kind of breaking change for 0.6.0
02:01 OldCoder Very well
02:01 OldCoder But it would be nice to fix the network... if his code does indeed do so
02:01 OldCoder twoelk, HexChat can detach windows
02:01 OldCoder So they can both be up at once
02:02 VanessaE I have very little problems with "the network" these days
02:03 twoelk hehe - just discovered I can put channel windows side by side - now I need a bigger monitor
02:03 OldCoder Right
02:03 VanessaE twoelk: do like I did and get three :D
02:04 OldCoder Or just overlay them so you can see bottom lines
02:04 OldCoder VanessaE, so Android is a separate issue?
02:04 VanessaE must be
02:04 OldCoder But the network code hasn't changed, has it?
02:04 VanessaE I personally don't care much about android
02:04 OldCoder Why would it be better for PCs ?
02:04 VanessaE all I know is I don't have significant problems with connecting, receiving media, etc.  it all just works.
02:05 rubenwardy I think that proller used enet, which was considered not a good choice by hmmmm
02:05 VanessaE yeah
02:06 OldCoder So, it was rejected based on merits and not compatibility?
02:07 OldCoder But with original protocol aren't downloads slow without curl?
02:07 VanessaE no one uses a build without cURL now.
02:07 OldCoder All right
02:07 VanessaE and anyone who does is just begging for it.
02:07 OldCoder Curl is for downloads. How is performance (on non-Android) in play ?
02:07 twoelk JT2 doesnT use remote media iirc
02:08 VanessaE JT2?
02:09 OldCoder ?
02:09 VanessaE oh just test
02:09 VanessaE bleh
02:09 twoelk :-D
02:09 OldCoder A non-compatible fork?
02:09 VanessaE there's no legit excuse not to set up a remote media server
02:09 OldCoder But answer question
02:09 VanessaE nah just some server
02:09 OldCoder Aside from downloads, how is performance?
02:09 VanessaE performance is adequate
02:12 OldCoder VanessaE, thank you
02:12 OldCoder proller had claimed massive improvements
02:14 rubenwardy might be worth penalising servers in the server list w/o remote media
02:15 rubenwardy meh
02:15 rubenwardy for tiny servers it doesn't matter
02:15 VanessaE I could agree with that.
02:15 VanessaE no server owner who cares uses just media_fubar anymore :P
02:15 Shara Some servers run just fine without it though, so it would be kind of harsh for them
02:15 OldCoder It is an action that carries its own penalty, isn't it?
02:15 VanessaE how so?
02:16 OldCoder Without curl, world loads slowly
02:16 OldCoder <players> to heck with this, we will go play on Bacontown
02:16 VanessaE only the media will load slowly.
02:17 OldCoder Why penalize the servers?
02:17 OldCoder Note: Kids running worlds off their laptops can't do so anyway, can they?
02:17 OldCoder (curl server through router)
02:17 VanessaE good question.
02:17 OldCoder (on a residential connection O_O)
02:18 OldCoder My worlds use GB a day of media bandwidth
02:18 VanessaE that's all? :)
02:18 OldCoder Well, I will ask
02:18 OldCoder Wait
02:19 OldCoder Have asked, host will notify me at some point
02:19 OldCoder Anyway it's a lot by old standards
02:19 VanessaE actually I have no idea how much total transfer all the servers I host use.
02:19 OldCoder Too much for residential
02:19 VanessaE nah
02:19 OldCoder Hm?
02:19 VanessaE residential connections are usually good to 200+ GB/mo so 1GB/day is bupkis.
02:20 OldCoder Depends on how many GB. And go over and some try to hit you with $1,000s of dollars penalty
02:20 OldCoder It's ridiculous
02:20 VanessaE never heard of that.
02:20 Zeno` I am running MT on a raspberry pi and 14.4k modem
02:20 Zeno` nobody ever plays :(
02:20 OldCoder Comcast has something like that now. And Zeno` that is remarkable!
02:20 OldCoder Ah
02:20 OldCoder Anyway, consensus is that the network code, so maligned in the past, is good enough after all?
02:21 OldCoder Provided only that curl is used for downloads?
02:21 VanessaE it's good enough that there are, in my uninformed opinion, much more important issues to deal with.
02:21 OldCoder Fair enough
02:46 OldCoder VanessaE, FWIW my host states that I'm presently using about 250 GB per month
02:46 VanessaE sounds about right for, what, 20 server instances?
02:47 OldCoder Or 60+ GB per week. This includes both RX and TX. Hm, yes. Somewhat less. Perhaps 15 worlds.
02:47 OldCoder I'm running webservers too but they are low traffic except for MT.
02:47 VanessaE that's about how many I host as well.
02:47 OldCoder Similar bandwidth?
02:48 VanessaE not sure.
02:48 VanessaE probably lower than yours
02:49 Warr1024 joined #minetest-dev
02:49 VanessaE https://daconcepts.com/vanessa/ho​bbies/minetest/stats-monthly.html
02:49 VanessaE maybe you can glean something useful from that.
02:51 QwertyDragon joined #minetest-dev
02:53 OldCoder R
03:00 TheReaperKing joined #minetest-dev
03:00 Warr1024 joined #minetest-dev
03:06 Warr1024 joined #minetest-dev
03:10 OldCoder VanessaE, TY
03:16 sofar VanessaE: I see a leak
03:17 VanessaE sofar: yeah I know
03:17 VanessaE it's unclear what it is.
03:17 * VanessaE pokes Zeno`
03:17 Zeno` hmm?
03:17 VanessaE possibly liquid queue overloading?
03:18 Zeno` massif still shows liquid queue every growing yeah
03:20 Zeno` I could run massif again tonight to see if there is something new though
03:40 twoelk left #minetest-dev
03:50 Warr1024 joined #minetest-dev
04:06 Warr1024 joined #minetest-dev
04:22 Hunterz joined #minetest-dev
04:31 Zeno` Did #5406 actually get profiled?
04:31 ShadowBot https://github.com/minetest/minetest/issues/5406 -- Reduce memory & function cost of Game class functions by nerzhul
04:31 Zeno` because I don't believe that's true. Last time I profiled using pointers was quicker (due to I assume better pipelining or fewer cache misses)
04:32 Zeno` well, maybe it is now true since so much has changed. I just hate claims of "this is faster" without profiles to back the claims up I guess
04:48 sofar Zeno`: since it's all within the same class/file, and most of that function is really simple code, I think it's correct
04:49 Puka joined #minetest-dev
04:54 Zeno` I think it's correct, yeah :) Just maybe not faster. Not sure but I could profile I guess. Hardly seems worth it though :)
04:58 Zeno` looking at my old notes the cache misses may have been due to something else anyway
06:02 ^v joined #minetest-dev
06:09 Hunterz joined #minetest-dev
06:22 ^v joined #minetest-dev
07:12 nerzhul joined #minetest-dev
07:22 lisac joined #minetest-dev
07:36 nerzhul merging #5444
07:36 ShadowBot https://github.com/minetest/minetest/issues/5444 -- Add multiline support to colorize. by red-001
07:45 Calinou joined #minetest-dev
07:46 red-001 there is an issue with 5444
07:47 red-001 I forgot to change the variable name from "colour_code" to "color_code" in part of the code
07:48 red-001 I pushed a fix but it got merged already
07:48 red-001 nerzhul ^
07:59 nerzhul fixed
08:04 red-001 thanks
08:25 kilbith joined #minetest-dev
08:45 nrzkt joined #minetest-dev
09:43 glut32 joined #minetest-dev
09:46 glut32 Is it the good github for Irrlicht 1.9? https://github.com/zaki/irrlicht
10:18 nrzkt it's just a mirror, irrlicht uses svn only
10:41 nrzkt joined #minetest-dev
11:07 Darcidride joined #minetest-dev
11:39 numzero joined #minetest-dev
11:47 Fixer joined #minetest-dev
11:50 DI3HARD139 joined #minetest-dev
12:04 proller joined #minetest-dev
12:26 nnnn20431 joined #minetest-dev
12:36 lisac joined #minetest-dev
13:28 nerzhul joined #minetest-dev
13:37 RichardTheTurd joined #minetest-dev
13:51 DI3HARD139 joined #minetest-dev
14:10 octacian joined #minetest-dev
14:53 YuGiOhJCJ joined #minetest-dev
15:01 Puka joined #minetest-dev
15:07 Taoki joined #minetest-dev
15:17 Fixer joined #minetest-dev
15:19 twoelk joined #minetest-dev
15:22 AnotherBrick joined #minetest-dev
15:41 nerzhul is #5436 okay for merge ?
15:41 ShadowBot https://github.com/minetest/minetest/issues/5436 -- [CSM] Change command prefix to "." and add "help" command. by red-001
15:46 Darcidride joined #minetest-dev
15:50 Fixer joined #minetest-dev
16:15 paramat joined #minetest-dev
16:23 paramat celeron55 sfan5 Shara and all, i had a 'middle of the night half awake' realisation. please can we add an 'old sneak' setting, and have that decide the running of old or new code? there's no need for players to suffer like this. i don't understand why i didn't suggest this earlier =3
16:23 sfan5 from a code standpoint this is a stupid idea
16:24 sfan5 also my PR + the "sneak_gltich" override has almost the same effect
16:24 celeron55 well it has been suggested already
16:26 celeron55 theoretically there should be a clean interface that allows swapping the new and old code without much maintenance headache, but in practice that might not be the case
16:27 celeron55 it probably is the case here more or less
16:27 celeron55 even in the long run
16:31 twoelk like the enable_build_where_you_stand switch?
16:32 sfan5 it should be a server-side settings in this case
16:32 lisac joined #minetest-dev
16:42 paramat #5443 is not an acceptable replication, i tested it, it also doesn't replicate 2-node sneak-jump. it may be very difficult to replicate the old behaviour within the new code
16:42 ShadowBot https://github.com/minetest/minetest/issues/5443 -- Bring back sneak elevators/ladders by sfan5
16:42 sfan5 it does no replicate that because i did not code it to do that
16:43 paramat and yes, have the setting simply switch big blocks of movement code, not pretty but it works
16:43 sfan5 thats more than just "not pretty" that's horrible
16:44 sfan5 and that will not fix damage evasion for 99.99% percent of people
16:44 sfan5 because those will want sneak elevators too
16:45 paramat i tested new sneak and it has many issues, it is not acceptable yet and may take time to get right
16:46 sfan5 did sneaking on snow work before?
16:49 paramat i know using the old code will not fix the undesirable bugs of the old code, those who want the old behaviour will just have to live with them, and will probably be happy to. doing this will keep people happy while we slowly improve new sneak code and try to provide good replications in the new code, if possible
16:50 paramat i don't know i never used sneak
16:50 sfan5 so what are these "many issues"? is it #5438 ?
16:50 ShadowBot https://github.com/minetest/minetest/issues/5438 -- Sneaking on 1/16th block doesn't, well, work.
16:55 paramat yes
17:00 Hunterz joined #minetest-dev
17:01 paramat i'm happy to write the PR for my suggestion
17:01 sfan5 paramat: please re-check my pr i fixed the complaints
17:03 paramat ok
17:04 paramat my suggestion would probably be a temporary thing to give us time to work on new code without players suffering
17:18 paramat #5440 is ready
17:18 ShadowBot https://github.com/minetest/minetest/issues/5440 -- Map generation limit: Make per-world by paramat
17:24 numzero joined #minetest-dev
17:40 sfan5 paramat: snow sneak does not work reliably in 0.4.16 either
17:41 sfan5 that covers 1 + 2
17:42 sfan5 3 is a bug indeed ("invisible wall")
17:42 sfan5 4 was explained already
17:42 sfan5 s/4\.16/4.15/ obviously
17:42 ssieb joined #minetest-dev
17:43 sfan5 you yourself said that this a dev version so breakage is expected, which means we don't need to (and shouldn't) add a temporary "switch code path" kind of thing
17:53 lisac joined #minetest-dev
17:57 lisac joined #minetest-dev
18:01 QwertyDragon joined #minetest-dev
18:06 nerzhul joined #minetest-dev
18:07 Krock joined #minetest-dev
18:07 Krock joined #minetest-dev
18:07 OldCoder I'm here to pass along a message from a middle-school player who's been here for 1.5 years.
18:07 OldCoder First, thanks again to those who talked yesterday. Core devs feel that some of the perceptions people have had are unfair. I felt, and it appears that C55 and others agreed, it was a good point to look at communications and processes.
18:07 OldCoder The message from the player is about the sneak feature:
18:07 lisac joined #minetest-dev
18:08 OldCoder "i liked it a lot. so do a lot of other players. i use it for my mines. most players who know about it like it."
18:08 OldCoder I myself didn't use it. My own requests are:
18:08 OldCoder (1) proceed slower with changes that have serious objections unless they're option-able
18:08 VanessaE you're late to the game, OldCoder.
18:08 VanessaE there's already two code proposals to reinstate it.
18:08 OldCoder Just finishing a message
18:08 OldCoder I know that
18:08 * OldCoder would like to finish
18:08 * OldCoder did not come here specifically yesterday about the sneak issue
18:08 OldCoder I myself didn't use it. My own requests are:
18:09 OldCoder (1) proceed slower with changes that have serious objections unless they're option-able
18:09 OldCoder (2) proceed *faster* with changes that are simple, unobjectionable, and option-able
18:09 OldCoder (3) continue to get to know others outside the core devs in #-project
18:09 OldCoder
18:10 OldCoder Commitment and skill on the part of core devs is noted and appreciated. My own input comes down to the three sensible suggestions listed right above
18:10 lisac joined #minetest-dev
18:10 OldCoder If these suggestions are taken, team will continue to grow and things will go more smoothly
18:10 OldCoder Thank you again
18:10 OldCoder VanessaE, late not, I think :P
18:10 OldCoder
18:11 * OldCoder feels that few will consider the three suggestions to be unreasonable
18:14 celeron55 i would guess everyone agreed to those way before you mentioned them, and way before the current issue too
18:15 OldCoder celeron55, call it a reminder, then. As you yourself felt was appropriate.
18:15 celeron55 it's just not so easy; having objectionable PRs hanging around is very tiring and trying to understand people is also tiring
18:15 OldCoder This is true. I can try to help more myself.
18:15 OldCoder Nothing has been said about taking every PR. The suggestions are sensible and "everyone agreed to those".
18:16 OldCoder I'd remind you, celeron55, that you yourself used the word "alienated" to describe this group. The word is too strong; there has been considerable improvement.
18:17 OldCoder 2 years ago, intruders were often shot on sight. Today, they're almost offered tea and a comfortable chair.
18:17 OldCoder So, I'd like to be clear that nothing I said yesterday myself was intended to be negative or even unfriendly in context.
18:17 OldCoder I don't feel that most of the group is "alienated" at all.
18:17 celeron55 i still think that's more of a perception thing than what actually has happened or happens
18:18 OldCoder Again, you used the word "alienated" yourself. Your own perceptions might be too negative.
18:18 celeron55 i don't know every english word; make up a better one if you know one
18:18 OldCoder No need. I think yesterday was positive.
18:19 OldCoder And it seems helpful to restate suggested guidelines
18:19 OldCoder BTW don't fall back on the "English word" excuse, celeron55. Your English is, and always has been, perfect :-)
18:19 lisac joined #minetest-dev
18:20 OldCoder Hm, somebody might wish to ban lisac temporarily with the message "Fix cycling and then PM"
18:21 celeron55 eh, whatever, i'm so tired i should probably say nothing
18:21 OldCoder celeron55, your input and that of others was very helpful. I think, again, this has been a good thing. I relate to tired BTW.
18:21 lisac OldCoder, Sorry, I'm testing out a new DE.
18:21 lisac It keeps crashing
18:22 OldCoder lisac, Ah. Could you close most channels? You're cycling far more in my network.
18:22 OldCoder Then test away. PM if you need help.
18:23 lisac joined #minetest-dev
18:23 lisac OldCoder, Sorry, crash.
18:23 lisac good idea.
18:24 OldCoder Heh. celeron55, I'm about a mile myself underwater in IRL and business. No MT time for weeks except for 1 world. But invested a day yesterday and don't regret it. Discussions with everybody were positive.
18:34 Player_2 joined #minetest-dev
18:35 Gael-de-Sailly joined #minetest-dev
18:43 lisac joined #minetest-dev
18:49 kilbith joined #minetest-dev
18:50 kilbith 2nd approval needed on #5395, please
18:50 ShadowBot https://github.com/minetest/minetest/issues/5395 -- FormSpec: support custom size and spacing for slots in list by adelcoding1
18:50 kilbith important for my company
18:58 Krock Pushing MSVC compiling fix in a few minutes: http://pastebin.com/raw/2P0e9wQN
19:04 Grandolf joined #minetest-dev
19:04 Krock kilbith, I think that the examples are not neccessary there
19:05 Krock or what do you think about this?
19:05 kilbith yeah, that's what I told him
19:05 kilbith feel free to remove them if you decide to merge
19:12 kilbith if our contributions are not merged in a timely manner, we'll likely be less inclined to submit PRs in the future
19:13 kilbith and believe me we have extremely competent people in our side
19:14 kilbith (that worked on e.g. Nintendo or Ubisoft)
19:20 VanessaE > extremely confident  > Nintendo
19:21 VanessaE one of these things is not like the other :)
19:21 kilbith confident ?
19:21 VanessaE competent I meant :P
19:21 kilbith bullshit, Vanessa
19:21 VanessaE (not sure how that translated into "confident")
19:22 VanessaE dude, it was a joke (I was alluding to some saying the Wii U was a market failure, and my personal opinion regarding the Switch)
19:25 kilbith I could understand your grief on Ubisoft though
19:25 kilbith but, offtopic :)
19:25 VanessaE actually, I know zip about Ubisoft (except that they exist)
19:29 Grandolf joined #minetest-dev
19:35 Tmanyo joined #minetest-dev
19:45 numzero kilbith
19:45 numzero how had you made the screenshot you just posted?
19:45 numzero (in #5446)
19:45 ShadowBot https://github.com/minetest/minetest/issues/5446 -- Add shader-based post-processing
19:45 kilbith by pressing F12
19:46 sfan5 :D
19:46 kilbith ask VanessaE or Zeno`
19:46 numzero well, I mean the post-processing
19:46 kilbith they tried it
19:46 kilbith well I just used RBA' post-processing fork and made a new shader
19:47 numzero okay, thanks
19:55 Fixer rbas shaders were insanely slow for me :(
19:57 numzero they are complicated
19:58 numzero and 3-layer render targets seems to be used unconditionally in his PR
19:58 numzero that’s not necessary for basic effects
19:59 kilbith shaders is a touchy area, the slightest inefficient code in GLSL afflict your computer
20:00 numzero how do you think would tone mapping be faster in post-processing than in node shaders?
20:00 kilbith it's obviously more expensive by post-processing
20:00 kilbith you need to send the current frame to the GPU all the time
20:01 numzero er, why do you ever need to take it from the GFX?
20:01 kilbith it captures the frame, transforms it to a texture and ship it to the GPU
20:02 numzero I thought the render target texture remains in the video memory
20:02 numzero oh well, if you mean RBA code...
20:02 kilbith it does yes, but your frames are changing all the time
20:02 numzero I remember, he used some software processing there
20:02 kilbith so what you're seeing is a succession of textures on the screen
20:04 * numzero looks through the code
20:11 rubenwardy joined #minetest-dev
20:26 numzero tested #5057 with various settings (including undersampling=1, not settable through the GUI, but supported internally)
20:26 ShadowBot https://github.com/minetest/minetest/issues/5057 -- Undersampling by numberZero
20:26 numzero render-to-texture doesn’t seem to affect the performance by itself
20:27 numzero (I mean, “render to texture, then display that texture”)
21:34 Gael-de-Sailly joined #minetest-dev
21:39 paramat joined #minetest-dev
21:47 Fixer joined #minetest-dev
21:52 juhdanad joined #minetest-dev
21:53 kaeza joined #minetest-dev
22:06 proller joined #minetest-dev
22:07 juhdanad The monthly "devs on a server" idea should be done in reality.
22:08 juhdanad If regular players could meet with the developers in-game, they would be more motivated to participate in the development.
22:10 paramat i'd like a dev server, white-listed because i want to play on a server but don't like public ones with too many kids
22:12 juhdanad I'm not talking about a dev server: young kids who like programming and want to practice will be the future's contributors.
22:12 juhdanad They just need encouraging and help.
22:13 sfan5 and the devs will provide their already scarce free time for this? nah probably not
22:13 juhdanad Well... The idea sounded great anyways.
22:14 paramat yeah i know, my comment was a bit off topic :]
22:16 proller joined #minetest-dev
22:26 Puka joined #minetest-dev
22:29 Fixer last time i seen a dev in game was celeron%% in 2011-2012, RBA in 2016 (by my request), and turns out est31 played on ESM server
22:31 nerzhul joined #minetest-dev
22:37 Shara Zeno is often on servers (as a result of me commenting to him that I felt devs don't play enough to see the real issues server owners face)
22:38 Shara That was what helped lead to spawn bug finally getting fixed... (thanks again to sfan5 for that by the way)
22:38 paramat funny thing is on servers i'm like in IRL, shy, sensitive loner, so public servers scare me =3
22:41 Shara Just disable chat :)
22:42 Shara I added a /hidename command to one of my servers for people who prefer to play alone without being disturbed
22:43 juhdanad But then what is the point of the server if you play alone?
22:44 paramat hehe
22:44 Shara You can still see issues a lot of the time, but then I don't choose to play that way
22:45 Shara I think it's just that some issues a busy server might encounter won't be seen in singleplayer tests.
22:45 Shara Old spawn bug being one example.
22:46 paramat indeed i would like to interact, but only with well behaved white-listed players
22:47 paramat i occasionally joined servers in the past and would essentially run for the hills then never rejoin
22:47 proller joined #minetest-dev
22:52 Fixer juhdanad: for survival, you don't want to disclose your position, also - kids
22:53 Fixer yes, I'm grampy
22:53 Shara Fixer, it's why DL has /hidename for any trusted players.
22:54 Shara There are usually solutions, but it's also true that most devs are probably too busy to play the game themselves.
22:57 proller joined #minetest-dev
23:00 kaeza joined #minetest-dev
23:16 juhdanad Any news on #5433?
23:16 ShadowBot https://github.com/minetest/minetest/issues/5433 -- Add clang format & skip build if no source file modified by nerzhul
23:17 VanessaE juhdanad: hoping for the overlay patch soon, too :)  (yeah I know it depends on another one)
23:18 juhdanad VanessaE: I have finished rebasing just at this moment. You are precise...
23:18 VanessaE :)
23:19 juhdanad But there are still style issues, so it won't be ready today.
23:30 proller joined #minetest-dev
23:49 fireglow left #minetest-dev

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