Minetest logo

IRC log for #minetest-dev, 2014-09-26

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

All times shown according to UTC.

Time Nick Message
15:02 loggingbot_ joined #minetest-dev
15:02 Topic for #minetest-dev is now Minetest core development and maintenance. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.ru/minetest-dev/ http://dev.minetest.net/
15:06 khonkhortisan joined #minetest-dev
15:26 mos_basik joined #minetest-dev
15:30 ShadowNinja Zeno`: The disadvantage is that mods might do something like "if minetest.version[1] == 0 and minetest.version[2] == 4 and minetest.version[3] == 9 then <0.4.9 specific code> end" which would cause issues when it ran on a fork like freeminer.  The full version string would be usefull for informational purposes though.
15:32 ShadowNinja Eg, minetest.version_string = "0.4.9-123-gabcdef" and minetest.project_name = "Minetest"
15:37 Animetrom joined #minetest-dev
15:40 Hunterz joined #minetest-dev
15:41 proller joined #minetest-dev
15:44 Animetrom joined #minetest-dev
16:03 proller joined #minetest-dev
16:19 ignacio you can just tell people not to do that
16:19 ignacio this way, they still do it but with silly workarounds
16:22 ShadowNinja ignacio: I've argued that before, others have insisted that users will ignore the warnings.
16:23 * ShadowNinja rebases his get_version patch.
16:40 Amaz joined #minetest-dev
16:40 ShadowNinja kahrl, sfan5: Comments on https://github.com/minetest/minetest/pull/1673?
16:41 ShadowNinja (in-game key change menu)
16:45 rubenwardy joined #minetest-dev
16:47 Krock joined #minetest-dev
16:56 ShadowNinja ignacio: https://github.com/minetest/minetest/pull/1689
16:58 zat joined #minetest-dev
17:07 Krock ShadowNinja, what's the exact target of this function`?
17:11 Krock I would prefer a "minetest.features" because it would be much easier to use. Also, it's not possible to get the features of a -dev version
17:11 ShadowNinja Krock: Use you mean?
17:11 Krock yeah
17:11 ShadowNinja Krock: The version is **NOT** to be used for feature detection!
17:11 ShadowNinja Krock: It's informational, look at the text of the issue.
17:12 ShadowNinja It mentions /CTCP MinetestServer VERSION.
17:12 ShadowNinja Which is currently only possible with:...
17:13 Krock I see
17:13 Krock Still, it does not look like a world-chaning feature
17:14 mberends joined #minetest-dev
17:15 ShadowNinja Hmmm, I had a mod that looked for the minetest binary, opened it, and looked for "VER=" in it (minetest_build_info).
17:15 Krock Okay, but for me it does not look like a much used feature (except of IRC). Many information are already on the serverlist
17:15 ShadowNinja Not sure where it is now though.
17:15 ShadowNinja Krock: And it doesn't have to be major.
17:16 ShadowNinja The server list should actually use version_hash IMO.
17:17 sfan5 why that?
17:17 Krock it's easier to read the current format than a hash
17:19 ShadowNinja sfan5: Becuase it's more detailed.
17:19 sfan5 0.4.10-dev is precise enough
17:19 Krock ^ for normal users
17:24 Calinou joined #minetest-dev
17:25 stujones11 joined #minetest-dev
18:15 hintss joined #minetest-dev
18:22 ImQ009 joined #minetest-dev
18:37 ImQ009_ joined #minetest-dev
18:38 proller joined #minetest-dev
18:58 ImQ009 joined #minetest-dev
19:02 Amaz joined #minetest-dev
19:18 jin_xi ShadowNinja: i got some very basic (particle vs plane) collision affector working
19:18 jin_xi http://paste.ubuntu.com/8435086/
19:19 Krock how about https://github.com/minetest/minetest/pull/1690 ?
19:21 VanessaE jin_xi: yay!
19:35 Miner_48er joined #minetest-dev
19:51 Krock Could someone please explain me those errors? http://pastebin.com/x97PXt4h
19:51 sfan5 1) fix your compiler 2) fix your compiler!
19:52 Krock 3) ignore _very_ helpful messages
19:52 Krock seems like something with the chars and strings is broken
19:53 sfan5 did you try reverting the settings const change?
19:54 Krock the. what?
19:54 Krock https://github.com/minetest/minetest/commit/cd64a92a8cde573404e8af9e220d4e1c6ce6453e ?
19:55 sfan5 yes
19:56 Krock I'll try
20:02 Krock "fatal: Your index file is unmerged" >.<
20:04 Krock meh. Have to revert all the commits
20:29 Krock ehm. wrong channel
20:40 * Krock slaps ShadowNinja with #1691
20:42 Calinou git push --force
20:42 Calinou (don't do that on public repositories)
20:46 chchjesus joined #minetest-dev
20:59 casimir joined #minetest-dev
21:14 ImQ009 joined #minetest-dev
21:21 NakedFury joined #minetest-dev
21:23 Megaf Hi all, is the second time that I get this crash
21:23 Megaf *** glibc detected *** ./minetestserver: malloc(): smallbin double linked list corrupted: 0xa0ee3828 ***
21:23 Megaf Aborted
21:25 GhostDoge joined #minetest-dev
21:27 Megaf and again we have theses meaningless error messages
21:28 VanessaE blame the folks who work on glibc then
21:28 VanessaE they're the ones who came up with that error message, not Minetest
21:28 VanessaE it's still a Minetest bug, obviously
21:46 EvergreenTree joined #minetest-dev
21:54 Calinou error messages can't always be meaningful
21:55 Calinou “segmenation fault” isn't meaningful to the beginner
22:15 kaeza joined #minetest-dev
22:30 kaeza diemartin> what is the general opinion in splitting `minetest.conf` into sections? This would a) enable the user to quickly know which settings are used by server/client/both, and b) I think it would lead to faster lookup in most cases
22:30 kaeza . of course, for backwards compat, settings in the "root" section would be used if the setting doesn't exist in the wanted section
22:31 kaeza sorry for repeating, but there are no logs from yesterday/early today
22:31 Megaf minetest.conf? No need to split it!
22:31 Megaf Just dont split it!
22:34 RealBadAngel veto for splitting idea
22:34 VanessaE veto for splitting, but vote in favor of having a different default filename for the server if one is found, e.g. minetest-server.conf or so
22:35 VanessaE if not found, get everything from minetest.conf as usual.
22:35 VanessaE I've had to field a few support requests from users setting up servers where stuff didn't work exactly right because they had client-side settings in a server config.
22:36 RealBadAngel what we do really need is reorganized menu, with more common settings aviable in GUI
22:36 RealBadAngel rest of them are hardcore, freaks will find a way on how to use them
22:37 RealBadAngel atm we dont have a single cell aviable to add any more settings
22:37 kaeza my proposed format is like this: https://gist.github.com/kaeza/1866cbca9a06cf7ba9a2
22:37 kaeza (simplified)
22:37 VanessaE kaeza: oh G*d no
22:37 VanessaE ANYTHING but that method
22:38 RealBadAngel at this point im rising what c55 proposed, simplified menu with single player/multi/settings at the beggining
22:38 VanessaE (note: "ANYTHING" does not imply allowing for json or XML or anything similar :P )
22:38 RealBadAngel and get rid of those stupid tabs
22:38 kaeza VanessaE, what is wrong with it?
22:38 VanessaE kaeza: in practice, I find it hard to read, hard to find stuff within such a file when it gets long
22:39 VanessaE not enough visual separation between the sections and their content
22:39 RealBadAngel config file is not a newspaper
22:39 RealBadAngel and shouldnt be touched without knowledge
22:39 VanessaE we have a filesystem - let it do its job.  just like these damn apps with MDI.  we have a desktop that can do the same thing - let the window manager do its job!
22:39 kaeza well, the current way is harder to read, with options pertaining to client and/or server intermixed
22:40 RealBadAngel current way is a nerd way
22:40 VanessaE kaeza: no argument there - all lumped into one disorderly file IS harder to read, I'll grant you.
22:40 RealBadAngel and it shall stay so
22:40 VanessaE kaeza: I'm saying we'd be better off with a minetest-client.conf, minetest-server.conf, minetest-modfoo.conf, etc.
22:41 VanessaE if we split at all
22:41 kaeza another option could be #include config-server.conf <_<
22:41 RealBadAngel all what is safely useable for greenhorns shall be aviable in main menu
22:41 VanessaE RealBadAngel: even the well-seasoned can get confused by a poorly-laid-out config file :P
22:43 kaeza also related: <diemartin> my main concern with settings is traversing an entire hashtable of maybe 100s of settings is waaay slower than say, look up a section in a list of maybe 10s of sections (at most), followed by a lookup in a table of say < 10 settings in most cases (with exception of [server] or [client] maybe)
22:43 RealBadAngel as i said, no reason for them to dig in files
22:43 kaeza nevermind the "traversing the hashtable" thing
22:44 RealBadAngel instead of making nerd way more nerd friendly we need greenhorns friendly way
22:44 RealBadAngel and i keep repeating this since quite a while
22:44 RealBadAngel mt is totally unfriendly for newbies
22:45 RealBadAngel ie mod names and folders
22:45 kaeza ^this, 100% agree
22:45 VanessaE hear hear,
22:45 VanessaE this needs solved.
22:46 RealBadAngel theres an easy solution, i already said that
22:46 RealBadAngel single file within mods folder containin its name
22:47 RealBadAngel so folder name wont matter at all
22:47 VanessaE yep
22:47 kaeza hence #1062
22:48 RealBadAngel and i will veto any try to release .11 until its done
22:48 VanessaE https://github.com/minetest/minetest/pull/1062
22:48 VanessaE btw, where is that damn bot?
22:48 kaeza (with changes suggested, of course)
22:49 RealBadAngel not that way kaeza
22:49 RealBadAngel obtain the name from the specified file, thats the right way
22:49 RealBadAngel cutting the folders name is risky
22:50 kaeza >Allows mod authors to provide a mod.conf file, to specify the mod name directly (and can later be used for other information about mods).
22:50 RealBadAngel make it obligatory
22:50 RealBadAngel and remove cutting part
22:50 kaeza no way
22:51 RealBadAngel thats the way
22:51 kaeza there are TONS of mods which won't be updated accordingly
22:51 RealBadAngel ofc
22:51 RealBadAngel but later on everything will work
22:52 kaeza I could agree with making it obligatory for e.g. .12 or .13, while showing a warning from .11 onwards
22:53 RealBadAngel next xmas? forget it
22:53 RealBadAngel we dont have so much time
22:53 RealBadAngel folks are running away from mt because of such DUMB things
22:53 kaeza but there's just too many "bug reports" about "foomod-master not working!!!!1!!"; do you want to make it worse?
22:54 RealBadAngel yes i do, for a short period of time
22:54 RealBadAngel all the old mods, unmaintained shall go to old mods section
22:54 VanessaE better to reject the mod out of hand as "just doesn't work anymore"
22:54 RealBadAngel active modders will update own quickly
22:54 VanessaE than to make users think it's minetest that's what's bugging out
22:55 RealBadAngel and active mods shall go immediately into modstore
22:55 RealBadAngel period
22:55 VanessaE and it would not be at all hard for those of us here to fork and fix
22:55 VanessaE we're just talking about a single file that can be added to any mod
22:55 VanessaE fix it, post to the forum, get a forum moderator to add it to the first post in lieu of the original repo
22:55 jin_xi so, here are my results with collision enabled irrlicht particles: http://youtu.be/51-OZOSNSqs
22:55 VanessaE simple.
22:56 RealBadAngel VanessaE, yes its as simple as that
22:56 VanessaE yes it's a lot of mods, but this is something that should be fixable in a couple of afternoons if a few of us work together on it
22:56 jin_xi but i dont really understand. geometry is weird, and they still fall after some bounces.
22:56 RealBadAngel jin_xi, thats nice
22:57 RealBadAngel can you make them replace our ones?
22:57 jin_xi thats why im trying this
22:57 VanessaE jin_xi: do I read this correctly that those affectors aren't really having a huge effect on your fps?
22:58 jin_xi i don't think so, but then its just a single plane to check for now
22:58 RealBadAngel jin_xi, also you should take care of lighting for particles
22:58 RealBadAngel read light at pos and apply that for particles
22:59 jin_xi this experiment showed me some weird stuff, water and clouds are rendered in front of particles
22:59 jin_xi and yes, light needs fixing too
23:00 jin_xi the clouds thing makes a really trippy effect, its like in an escher drawing
23:01 RealBadAngel propably wrong rendering order
23:02 RealBadAngel btw, with my recent changes it seems i get like 2x frame rate
23:03 VanessaE what changes?
23:03 RealBadAngel i rewrote mapblock mesh
23:03 VanessaE ohh
23:03 jin_xi nice
23:04 RealBadAngel its buggy but its working
23:04 RealBadAngel speed up is amazing anyway
23:04 VanessaE got the code somewhere? :)_
23:05 RealBadAngel yup, here ;)
23:05 VanessaE weisenheimer :P
23:05 RealBadAngel i will make it aviable in a few days
23:05 RealBadAngel its too much on revolution
23:13 jin_xi anyways whatever it is im doing with particles, i am gonna need help.
23:15 RealBadAngel jin_xi, how can i help you?
23:16 jin_xi well, regarding particles collisions code, problem is that i do not really understand it...
23:17 RealBadAngel me neither, mainly because i havent saw it even ;)
23:17 RealBadAngel any links to the code you have saw/used?
23:17 jin_xi http://paste.ubuntu.com/8435086/ this is the whole affector
23:18 jin_xi but as you can see in the video particles somehow still fall trough the plane after a while
23:19 jin_xi also i dont get the geometries involved with BS
23:22 VanessaE jin_xi: my first thought is a rounding error
23:23 VanessaE (or rather a precision error)
23:30 VanessaE wait, could it be possible that your assignment at line 13 is turning out to be ever so slightly negative every so often?
23:30 VanessaE er line 18?
23:30 VanessaE (in your paste)
23:30 VanessaE I get the impression that getIntersectionWithLimitedLine() wants its two line coords to be in order
23:33 jin_xi what this function does is return (getIntersectionWithLine(linePoint1, linePoint2 - linePoint1, ...
23:35 VanessaE hm
23:36 VanessaE just making guesses.
23:56 RealBadAngel im reading the docs now
23:57 RealBadAngel http://irrlicht.sourceforge.net/docu.net/Irrlicht.Core.Plane3D.GetIntersectionWithLimitedLine.html
23:57 RealBadAngel 2nd point shall be next position of particle
23:58 RealBadAngel so the line will be current pos -> next pos of particle
23:58 RealBadAngel if line will intersect with plane it will return true and fill the intersection point

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