Time Nick Message 00:41 paramat hmmmm in case you didn't see these.. http://irc.minetest.ru/minetest-dev/2014-12-02#i_4046984 http://irc.minetest.ru/minetest-dev/2014-12-02#i_4047185 00:42 hmmmm hmmm 00:42 hmmmm well people did not start getting std::bad_allocs 00:42 hmmmm i'd say this is due to my latest patch... ugh another instance of things working fine for me 00:52 hmmmm can somebody give me a backtrace of this exception happening to them, the bad parameter, and the actual parameters they have configured? 00:53 paramat ^ VanessaE ? 01:31 Sokomine any solution yet for the "Some exception: "One or more noise parameters were invalid or require too much memory"" problem? 01:32 Sokomine ah, ok, current topic of conversation 01:36 Sokomine hmmmm: can't find any information regarding the bug that looks relevant. any hints as to what to look for in particular? else i'll try to pinpoint it to a mod.. 01:43 VanessaE I haven't obtained a backtrace, all I got so far is that it happens on every one of my worlds that involves a standard mapgen. New York City world (I think kilbith did that one) has a flat map, and works. newly created worlds work, too. 01:44 VanessaE worlds that don't work include my current VE-Survival and VE-Creative servers. 01:44 VanessaE what files do you need? just map_meta.txt? 01:45 paramat Soko the bug seems due to hmmmm's recent noise params changes, not a mod 01:46 VanessaE (I mean of course my local copies of those two worlds; I don't run git HEAD on the production servers) 01:46 Sokomine paramat: seems so, yes. even without any mods at all, i do get only the error message 01:46 Sokomine ethereal is affected as well, and that's v7 mapgen 01:47 VanessaE hmmmm: http://pastebin.com/1cns5KRX 01:47 VanessaE params from VE-Survival 01:48 VanessaE versus the aforementioned working NYC map, http://pastebin.com/W5xAW6V7 01:48 VanessaE versus a newly-created world: http://pastebin.com/fc62A61P 01:49 VanessaE which also works. 01:52 Sokomine perhaps that ethereal world that fails for me was initially created using v6. singlenode worlds work. worlds with other mapgens also seem to work. the old v6 worlds all seem to fail 01:59 hmmmm oh 01:59 hmmmm OLD v6 worlds you say 01:59 hmmmm yup i get a failure 01:59 hmmmm i'll check it out 02:00 hmmmm getNoiseParamsFromValue might be busted somehow 02:01 VanessaE hmmmm: not THAT bloody old .... 02:01 hmmmm welp, np.octaves is 2877 for a random old map 02:01 hmmmm pfth 02:01 VanessaE FAIL :P 02:01 hmmmm and now guys 02:02 hmmmm it's OKAY to have crashes like this 02:02 hmmmm it's the bleeding-edge development build 02:02 hmmmm it's not like it's some sort of hideous failure 02:02 VanessaE hmmmm: not when it breaks nearly every single world in existence ;) 02:03 hmmmm your world isn't getting nuked 02:03 hmmmm which you should keep periodic backups of anyway 02:04 VanessaE well considering I don't run bleeding edge on my production servers, and I *do* take daily backups (I keep a week's worth), and rollback is also enabled, I'd say I've at least got that covered. :) 02:08 hmmmm alright looks all good 02:09 hmmmm there, sorry 02:10 VanessaE hmmmm: take a closer look at that third paste I linked 02:10 hmmmm versus a newly-created world: http://pastebin.com/fc62A61P this? 02:10 VanessaE yes. 02:10 hmmmm what's wrong with it 02:10 VanessaE are those repeated foo = followed by foo = { stuff } lines intentional? 02:10 hmmmm yeah 02:11 hmmmm i was having an internal debate whether or not they should be there, because technically they should 02:11 hmmmm if a setting has a subgroup, should i not write the value if blank? 02:11 VanessaE they're intended to mimic how Lua does it? 02:11 hmmmm what if that value is legitimately blank? 02:11 hmmmm not really 02:12 hmmmm so I was considering constructing SettingsEntry structures with some odd character like a y with an omlaut over it to signify "this setting really hasn't been set" 02:12 VanessaE eh 02:12 hmmmm but that's stupid, because if it is actually blank and supposed to be blank, then get() will return a blank setting regardless 02:13 hmmmm so the logic for this is really simple and i'll do it right now in fact 02:13 VanessaE it's not my place to debate what absolutely *must* be there, but were I directing someone how to edit that file if they wanted to change a parameter, well, you can imagine how easily it will be to confuse them. 02:14 VanessaE -ly 02:14 VanessaE easy* 02:20 paramat ah cool a fix 02:21 hmmmm "did hmmm even test this shit??!" nope 02:21 hmmmm i forgot to write a unit test for the noiseparams stuff 02:57 Sokomine thanks! seems to work fine again 02:58 Zeno` hmmmm, what is stof? 03:00 Zeno` c++11 stof? 03:00 Zeno` Please don't use c++11. :3 03:33 kaeza some interesting read http://ask.slashdot.org/story/14/12/02/007206/ask-slashdot-non-coders-why-arent-you-contributing-to-open-source 03:35 Zeno` hmmmm, perhaps we can add something in util/numeric until c++11 is supported by every platform MT supports? Even then it can probably stay in util/numeric (referring to stof()) 03:36 hmmmm Zeno`, stof() is not an issue. it's defined by minetest. 03:36 Zeno` I couldn't find it (sorry) 03:36 hmmmm util/string.cpp 03:37 Zeno` wow... I must have grepped wrong. 03:38 Zeno` left out the -r (oops) 03:43 Zeno` I'm not sure I like stoi and stof (and assert for that matter) being "overridden" like this 03:44 hmmmm stoi and stof predate C++11 03:44 hmmmm assert is defined by the C++ spec as a MACRO that can be overridden by #undef #define assert() whatever 03:44 Zeno` yes, I'm talking about #define stoi mystoi and #define stof mystof 03:45 Zeno` stof is not in the standard until C++11 though 03:45 Zeno` stoi has been around since the dawn of time 03:45 Zeno` and just because assert /can/ be overridden doesn't mean it /should/\ 03:46 Zeno` kind of makes people seeing atoi and expecting 'man atoi' to be what they're looking at bound to be confused (IMHO) 03:47 Zeno` same with assert. You should be able to man assert and expect that behaviour :/ 03:50 Zeno` Similarly you should be able to go to the C or C++ standard, look at those functions and expect that behaviour 04:25 paramat hmmmmm your mgv7 rivers have underwater entrances to vertical water fissures that extend downwards for 1000s of nodes. these were a nice surprise 05:02 hmmmm paramat, they're definitely not intentional 05:07 paramat very cool accident then, i love the rivers, although they need tuning. the curves they carve into a sandy seabed are beautiful 05:08 paramat they could be narrower and perhaps fewer of them 05:10 paramat i can see potential in mgv7 i'm inspired to tune it and make it good 05:10 hmmmm paramat... i want to get rid of mgv7 and kind of start over again 05:11 hmmmm it's just that there are more important things to fix like schematics and luavoxelmanip 05:11 hmmmm and biomes 05:13 paramat many players will miss it though, maybe it could be left in for a while and you start again with v8.. 05:14 hmmmm yeah, i noticed the same with the previous "mapgen v7" 05:14 hmmmm lots of players actually like it, but you the mapgen maker do not, so you don't think very highly of it 05:14 hmmmm so you don't think twice before trashing it 05:21 paramat yes. this is ready for review/merge https://github.com/minetest/minetest/pull/1892 i re-added the underwater grass hack 05:23 paramat .. the title is misleading, i can't edit that, the hacks are there 05:23 hmmmm you can edit it 05:23 hmmmm git commit --amend 05:23 hmmmm it'll let you edit the commit message in the default system editor 05:25 paramat the commit message is correct though 05:26 hmmmm yup, looks good 05:28 paramat now i understand the underwater grass bug i feel i can still come up with some reasonable mgv5/v7 biomes and tree schematics for 0.4.11 05:30 hmmmm if we go in separate directions with biomes, how will registration work? 05:31 paramat um how about modular biome systems, a choice of multiple biome methods, to be used with the choice of base terrains 05:31 hmmmm yeah that's the point 05:32 hmmmm but different biome systems are going to have different biome parameters 05:32 paramat ah i see 05:32 hmmmm so when people make biomes for their mods they're not going to want to add two sets of biomes 05:33 paramat ok there needs to be a primary system 05:38 paramat anyway no worries just see how it goes with our experiments, see which is popular etc 05:44 paramat bbl 07:10 realbadangel hi 07:11 realbadangel sorry for being offline for a few days, had some problems with my box 09:27 celeron55 umm... i guess i just broke servers.minetest.net DNS 09:28 celeron55 i wonder what the IP was 09:29 celeron55 nearlyfreespeech.net just set a new world record in the obscureness of a DNS configuration interface; i have no idea how i even did that 09:30 celeron55 hmm, it might be 212.38.181.18 and 2001:1b40:5000:ff3::1c:a98e 09:35 celeron55 www.minetest.net will be now broken for however long the TTL of the DNS was 09:36 celeron55 because i can't make the old host accept minetest.net without possibly breaking the DNS due to automatic crap 09:37 celeron55 until then, this name works: http://minetest.metacrane.com/ 09:42 Zeno` ?? nslookup servers.minetest.net still works for me. I guess I'm just lucky 09:42 celeron55 servers.minetest.net unbroke now 09:42 Zeno` or they're fixed :) 09:43 celeron55 the default TTL of minetest.net was like multiple days 09:43 celeron55 so let's see if DNS servers actually wait that long... 09:43 celeron55 (what the fuck, i might ask, though) 09:45 celeron55 looks like www.minetest.net works already too 09:50 celeron55 minetest.net will be down up to 24 hours i guess (that was the TTL i think) 09:50 celeron55 (the TTL that was not set and neither seen by me, blame nearlyfreespeech.net's ridiculous defaults) 11:37 celeron55 minetest.net works now 11:37 celeron55 it's now a whole lot more responsive for me at least, might be slower for the americans this time 11:54 celeron55 oh fuck, the forum 11:55 celeron55 why didn't anyone tell me it's down too 11:55 celeron55 does it use the same IP as irc.minetest.ru? well i guess i'll try that 11:57 celeron55 doesn't seem like it 12:01 celeron55 should work after 10 minutes assuming dns servers actually follow the TTL 12:11 celeron55 and now dev.- and wiki.- will be working after a moment 12:14 ninnghazad thx, can access forum again 14:15 Amaz I have a bug where when I press f10, a black box comes down, but there is nothing in it... I can see the chat behind. Using latest (Or 1 minute old) git. Screenshot: https://cdn.mediacru.sh/onAAAMCIVFPg.png (Don't ask why I cropped the majority of the hud off. I must be tired!) 14:17 ninnghazad Amaz: thats the commandline 14:17 ninnghazad type commands in it, like /help 14:19 Amaz Nope. 14:19 Amaz Just comes up in the normal chat space. 14:19 Amaz I used to get the chat history in there. 14:53 celeron55 Amaz: probably the font is broken somehow 14:53 celeron55 whoever messed with that lately should have an idea why 15:01 Amaz Yeah. I noticed the font seemed to have changed... 15:01 sfan5 Amaz: sapier recently merged some patch with font engine stuff 15:01 Amaz Should I post an issue on github? 15:03 shadowzone To be honest I kinda like the new font. 15:05 PilzAdam font_size setting in minetest.conf seems to not work anymore 15:05 PilzAdam can anybody test this? 15:06 sfan5 I think before we can release we'll need to fix a ton of blocker bugs 15:06 PilzAdam shadowzone, there is no new font; it's the same as before 15:06 shadowzone Oh 15:07 PilzAdam this whole "fontengine" stuff is weird 15:07 PilzAdam how do I use my custom .ttf fonts? 15:11 PilzAdam the gui scale factor breaks the version text in the top left 15:11 PilzAdam and the progress bar when loading or exiting a world 15:12 PilzAdam oh wait, its not the gui scale factor 15:12 PilzAdam the font_size setting in minetest.conf is only applied to those 15:16 VanessaE [12-02 12:57] target_font_height *= g_settings->getFloat("font_size")/TTF_DEFAULT_FONT_SIZE; 15:16 VanessaE [12-02 12:57] can you add that line in guiFormSpecMenu.cpp L 1977 and try again ? 15:16 VanessaE [12-02 17:29] guiformspecmenu.cpp L 1935 15:16 VanessaE [12-02 17:29] use_imgsize = 0.53 * ... 15:16 VanessaE [12-02 17:29] change to 0.5555 15:16 VanessaE PilzAdam: ^^^^^ apply those changes. it'll help a bit. the latter part (0.53 --> 0.5555) is to work around a rounding error 15:16 VanessaE sapier and I discussed this at length last night. 15:16 PilzAdam Ill create an issue 15:18 PilzAdam #1896 15:18 ShadowBot https://github.com/minetest/minetest/issues/1896 -- font_size setting broken 15:18 VanessaE PilzAdam: sapier's one-liner ^^^ fixes that. 15:19 PilzAdam then he should push that to upstream 15:19 VanessaE I told him to after we tested it but he has additional changes to make I guess 15:21 VanessaE there, I added the necessary patch 15:28 PilzAdam was the scaling of the inventory modified? it seems to be larger 15:31 VanessaE it might have been; fonts therein are definitely scaled with window size now 15:32 VanessaE afaik everything scales with window size now 15:33 VanessaE d1d6a97b 15:33 VanessaE (except the hotbar) 15:34 PilzAdam the hotbar used to scale with the window size 15:34 PilzAdam why was that changed? 15:34 VanessaE idk 15:37 PilzAdam was it intentionally changed or should I create an issue for that too? 15:38 VanessaE I wish I knew :P 15:41 PilzAdam #1897 15:42 ShadowBot https://github.com/minetest/minetest/issues/1897 -- Hotbar doesnt scale with window size 15:42 VanessaE well is it a bug or not-a-bug ? ;-) 15:42 VanessaE (look at your labels) 15:43 PilzAdam I don't know if it was intentionally changed or not 15:43 PilzAdam its bug-state is in a superposition 15:43 VanessaE haha 15:43 VanessaE frankly, I'd call it a bug if it used to work but doesn't anymore. 15:44 VanessaE and it 15:44 VanessaE and it's not just you - it's a little too small even at 1600x1200 20" 15:44 ninnghazad hm, i think this wasn't planned - he could have disabled it, but his last commits, at a glance, do not seem to try to do so. 15:44 ninnghazad resizeHotbar and friends seem rather untouched 15:45 VanessaE yeah, even dreambuilder's 16-slot hotbar suffers 15:45 VanessaE I don't think this is intentional. 16:21 PilzAdam sfan5, https://github.com/minetest/minetest_game/pull/358 https://github.com/minetest/minetest_game/pull/359 16:22 sfan5 PilzAdam: both PRs are ok 16:24 PilzAdam according to http://dev.minetest.net/minetest_game_Development I am not allowed to merge the PRs? 16:24 sfan5 celeron55: did you, by any chance, break the other .minetest.net subdomains too? 16:24 PilzAdam do I have to get BlockMen or nore agree too? 16:25 celeron55 sfan5: i broke every subdomain and do i not have a list of them 16:25 sfan5 PilzAdam: I'd say those count as small fixes and I allowed you to merge them 16:25 sfan5 that explains my problem 16:25 sfan5 I guess I'll just add mine again 16:25 celeron55 you'll have to tell it to me 16:26 sfan5 are we no longer using afraid.org? 16:26 sfan5 or did that break too 16:26 celeron55 not since... like a year or something? 16:26 sfan5 TIL 16:26 sfan5 anyway: 16:27 sfan5 meow.minetest.net pending A 212.38.181.18 16:27 sfan5 meow.minetest.net pending AAAA 2001:1b40:5000:0ff3:0000:0000:001c:a98e 16:27 sfan5 sfan5.minetest.net A 5.231.56.127 16:27 celeron55 well if you don't know that, then was your subdomain broken for that long? 16:27 sfan5 no 16:27 celeron55 this is weird 16:27 sfan5 it worked until today 16:27 sfan5 (when hexchat decided to reconnect) 16:28 celeron55 i don't understand... whatever, i'll just add it 16:28 PilzAdam sfan5, done 16:29 sfan5 PilzAdam: thanks 16:31 PilzAdam sfan5, I found this old branch in my repo: https://github.com/PilzAdam/minetest_game/commit/8baf0d29e73c43bd55008f5af22f616d50391ff8 16:32 sfan5 celeron55: thanks for adding them;; maybe nearlyfreespeech have a fallback to the afraid nameservers when it didn't have success in resolving a .minetest.net subdomain 16:32 sfan5 s/have/had/ 16:33 sfan5 PilzAdam: if you can/want to update that it would probably be a good addition 16:36 PilzAdam hmmm.... I'll look into it 16:38 celeron55 i just realized the mail server must also be broken now 16:39 sfan5 we had a mailserver 16:39 sfan5 right 16:40 celeron55 uh oh, and the old mail field in freedns.afraid.org doesn't work here 16:41 celeron55 okay so now i seriously wonder what kind of DNS setup i actually had before this 16:44 PilzAdam sfan5, maybe the fuel thing isn't that good, since the register_craft() way allows replacements 16:44 PilzAdam or should we still add support for fuel in groups? 16:45 celeron55 umm... well, anyway, if something in the DNS doesn't work, please notify me 16:45 PilzAdam it would still be shorter and easier to add the fuel group for all items that don't want replacements 17:28 PilzAdam sfan5, https://github.com/minetest/minetest_game/pull/361 17:32 kilbith PilzAdam: there are some bugs with the ice alpha_channel 17:32 PilzAdam hm? 17:33 kilbith "x-ray vision" (you can see the caves through the terrain), dark shadows... 17:33 PilzAdam "Only merge if the engine finally fixes the glitches with half-transparent nodes." 17:34 kilbith i'm not sure that's the glitches you had in mind 17:35 Calinou set drawtype of ice to glasslike 17:35 kilbith aside that, 1m width of ice is not opaque IRL 17:35 kilbith -noy 17:35 kilbith -not 17:36 Calinou https://github.com/minetest/minetest_game 17:36 Calinou is reported as mostly Logos :( 17:36 Calinou where do we complain at GitHub for that? 17:36 PilzAdam Calinou, it is set to glasslike 17:36 PilzAdam kilbith, screenshot? 17:37 kilbith PilzAdam: i can't actually (Raspberry); just remembering those bugs 17:38 kilbith it's just specifical to the use of alpha 17:47 PilzAdam sfan5, https://github.com/minetest/minetest_game/pull/362 17:47 PilzAdam should I fix the code style there too? 17:50 PilzAdam also https://github.com/minetest/minetest_game/issues/363 18:10 VanessaE PilzAdam: see comment on that one. 18:11 PilzAdam VanessaE, what happens with texture packs? what if they only supply default_dirt.png with a different resolution? 18:12 VanessaE there'd be an issue then, or texture pack authors could get off the stick and update :P 18:12 VanessaE or you could name the textures just right and texture packs would probably just continue to work normally 18:13 VanessaE (e.g. an overlay that's named for a full texture would just overlay the full texture and the dirt would never be seen anyway) 18:13 VanessaE (I used to do that with default dirt with grass sides) 18:14 VanessaE I did them as overlays for forward compatibility with future dirt textures 19:55 pitriss hi please I want to ask.. how i can generate MT password hash from script? I thought about dokuwiki auth module using auth.txt from MT.. 20:00 PilzAdam pitriss, https://github.com/minetest/minetest/blob/master/src/util/string.cpp#L147-164 20:01 Krock iz base64 encoded 20:03 pitriss hmm i looked there but I'm not sure if i understood it correctly.. its usernamepassword sent to sha1 and then base64 encoded? becazse this seems as not matching in bash.. but thanks :) 20:58 kaeza pitriss, it's because in bash, sha1sum returns the "hexdigest", and not the raw digest 20:58 kaeza pitriss, this seems to work: https://gist.github.com/kaeza/34342260f74c1f5683db 20:59 kaeza (Minetest base64s the raw bytes of the digest) 21:00 kaeza use as `my_hash=$(hash_pw "$user$pass")` 21:01 pitriss ahh ok thanks kaeza:)