Time Nick Message 00:01 fluxionary i cleared out a couple gigs of ancient media from my minetest cache directory, i've been pondering whether to propose creating some sorts of (configurable) limits on it 00:03 muurkha fluxionary: yeah, I didn't mean to track an individual, though it might be possible 00:03 muurkha can you query a particular hash without actually sending them the media? 00:05 rubenwardy I believe the server sends a list of hashes the client needs, and then the client requests ones it doesn't have 00:06 rubenwardy theoretically, the server might be able to not send at that stage, but I suspect it'll prevent the client from joining the game 00:06 muurkha yeah, that's what it sounds like 00:07 muurkha although at that point the server knows your username 00:07 muurkha it's probably not an important privacy leak even if it can be probed nondestructively 00:17 diceLibrarian Ruben, did you see that stupid image I put in the buildings channel? 01:30 HanSaul thats correct if the server does not send the client will fail to connect 01:31 HanSaul i checked the cache folder and it does just seem to slap files in there by hash and names them after said hash 01:31 HanSaul which is fine 01:32 HanSaul i was considering cutting down the minetest_game but leaving it as-is would make my server load faster than serving up some botched version 01:32 HanSaul as those files would already be in most users local cache 01:33 HanSaul with all the png compression that takes place, I assume that only helps the connection speed for sending clients files, eventually they get unpacked into memory at full RGB 1 byte pixels or RGBA 01:34 HanSaul anyone running a server on an i9-13900k yet? 01:35 HanSaul and whats the biggest world file anyones ever heard of someone having? mine is 140 GB+ atm 01:36 HanSaul im guessing it will baseline somewhere around sea level map blocks being fully loaded 01:36 HanSaul whatever that number would end up being on average 01:37 HanSaul 30900^2 = 954810000 / 16^2 = 3,729,726 map blocks 01:39 HanSaul assume average mapblock is 8 mb? just guessing, thats 29,837,808 mb so about 30 TB 01:40 HanSaul just for full completion of mapblocks at sea level, but I am going to assume it would realistically be under 16 TB maybe more around 8 TB and with some house keeping nuking blocks outside protectors as Buckaroo does, probably under 1 TB easily 01:42 HanSaul I feel like minetest is unsustainable for popular servers, unless you do housekeeping as suggested above, and even then, protection options are... well those who do use them tend to use them incorrectly and there's still a very large percent who just dont use them 01:42 HanSaul if one is to take housekeeping seriously, they need to solve the protector issue first 01:43 HanSaul but then the only way to solve that is to give every node ownership which would probably make the blocksize even less sustainable 01:43 HanSaul so that protection becomes automatic for user placed nodes 01:43 HanSaul and shared areas can be configured in the inventory by the user 01:44 MTDiscord iirc ShadMOrdre or Skamiz Kazzarch once said that they emerged all of a map, or all of a map on surface level, thats probably in the top 10 of largest maps 01:44 HanSaul wow, did they specify how much diskspace that took up? 01:45 MTDiscord iirc it was in one of the many forum threads about expanding minetest world size if you want to go digging for it 01:45 HanSaul ill take a look thank you 01:46 HanSaul https://forum.minetest.net/viewtopic.php?t=19472 01:49 MTDiscord eh, not the exact thing i was thinking of 08:26 sfan5 assume average mapblock is 8 mb? just guessing, thats 29,837,808 mb so about 30 TB 08:26 sfan5 the raw uncompressed node data in a mapblock is 20 KB 09:23 MTDiscord Hi folks! 09:23 MTDiscord https://cdn.discordapp.com/attachments/749727888659447960/1032947200298319962/Capture.PNG 09:23 MTDiscord How can I create these suggested changes while I reviewing someone pr? 09:25 sfan5 https://a.uguu.se/xepcwXuK.png 09:25 MTDiscord but how to specify the line? 09:27 sfan5 you click on the plus sign that appears when you hover the right line 09:28 sfan5 you can also select a range by two shift-clicks as usual, then click the plus sign on the last selected line 09:30 MTDiscord the plus sign shows up only on the lines modified by the PR, but the line I want to comment hasn't been modified by the PR, but it should 09:30 MTDiscord and so the plus sign doesn't show up 09:31 sfan5 then you can't comment on it 09:31 MTDiscord ah! ok 09:31 MTDiscord thx 🙂 09:42 MTDiscord has the backport-5 branch been removed? 09:43 MTDiscord git says it's "stale". I still have it in my local clone of minetest/minetest repo 09:44 rubenwardy it's only in use when we're backporting to create a patch release 09:45 rubenwardy the release is then made in stable-5 09:48 MTDiscord ok. Thx! 13:15 MTDiscord Minetest icon is broken on Linux mint 21 when irrlicht sdl enabled 13:33 sfan5 known issue 13:44 MTDiscord oh sorry 15:10 MinetestBot 02[git] 04sfan5 -> 03minetest/minetest: Revise bump_version.sh script to address shortcomings (#12789) 139f0d884 https://github.com/minetest/minetest/commit/9f0d88407d9dadb1a8cf8c03131eda034f2f19e1 (152022-10-21T15:09:44Z) 15:13 MinetestBot 02[git] 04Desour -> 03minetest/minetest: Fix formspec focus (#12795) 137153cb8 https://github.com/minetest/minetest/commit/7153cb8a0bd6a0d405e2d84975812164233be4f6 (152022-10-21T15:11:41Z) 15:13 MinetestBot 02[git] 04snowyu -> 03minetest/minetest: guiChatConsole: fix the unicode characters crowded together on prompt… 138bdedd2 https://github.com/minetest/minetest/commit/8bdedd2bcf31ba9f8acd1457cfaeb652f64abf34 (152022-10-21T15:12:07Z) 15:13 MinetestBot 02[git] 04TurkeyMcMac -> 03minetest/minetest: Fix use of unassigned global "check" 13c78d565 https://github.com/minetest/minetest/commit/c78d565e011125a13c1e29584c04f2b97f0d6b3e (152022-10-21T15:12:21Z) 15:13 MinetestBot 02[git] 04Abdou-31 -> 03minetest/minetest: Fix typos and en_US/en_GB inconsistency in files inside doc directory… 137e11b8e https://github.com/minetest/minetest/commit/7e11b8eb721a9039e84ea51ec2e1f58d3c036849 (152022-10-21T15:12:39Z) 15:40 definitelya ù 16:18 MTDiscord thx sfan5 🙂 16:18 MTDiscord I've discovered a very powerful git tool: git grep 16:18 MTDiscord crazy 16:21 celeron55_ i often use grep -IsnR or some such, but the problem with that often is that it finds you all kinds of build files and hidden files that you don't care about. git grep gets around that IF all the files you're interested have been committed to git 16:21 celeron55_ interested in* 16:26 MTDiscord great 16:27 MTDiscord fixing common typos will be a piece of cake using git grep 16:39 MTDiscord https://cdn.discordapp.com/attachments/749727888659447960/1033057010935152742/necessary.PNG 16:39 MTDiscord wow 16:40 MTDiscord people tend to write it with 2 'c' 16:41 MTDiscord fortunately, those 3 are all what remain 🙂 17:36 definitelya Ok Minecraft ha sofficially won against Mineclone 2/5/-clonia/6/9/etc, they have a sniffa mob now, fam! 17:36 definitelya It's over for it. /s 17:37 MTDiscord snefler? 17:37 definitelya Is that what it's called? IDEK 17:41 muurkha a sniffa mob? 17:41 muurkha who has the sniffa mob, Minecraft or Mineclone? 17:43 definitelya Minecraft has it. 17:45 definitelya Keep your toes safe! 17:47 * muurkha looks with concern at toes 17:47 MTDiscord also how in the hell did a mob with the name "sniffa" get past mojang stockholm 17:48 definitelya xD 17:48 Pexin sniffalupagus 17:48 definitelya Like, why 17:48 definitelya ??? 17:49 definitelya Roller: I think its name is "Sniffer". Still... 17:49 definitelya I'm staying far away from that dumpster fire of a game. 17:49 MTDiscord oh thank god 17:54 muurkha so why does that mean Minecraft has officially won? 17:54 muurkha or, in what sense? 17:55 definitelya It's a joke. :P 17:55 MTDiscord glue sniffer mob 17:55 MTDiscord take my money microsoft 17:55 definitelya mhm 18:16 MTDiscord Is it fine to do only one PR to fix a large number of typos (100+) in many files (20+) at once? 18:25 MTDiscord Actually, 50+ files 18:26 MTDiscord code comment spellcheck I assume? should be fine 18:27 MTDiscord Minetest seems to have a general preference for large, squashed commits, so it seems sensible to me. 18:27 MTDiscord Nice 18:28 MTDiscord A question: are po folder exclusively for translation files? 18:29 MTDiscord yes, no touchie, they're supposed to be edited from weblate 18:30 MTDiscord Oh! No typo fixing on them? O.O 18:30 MTDiscord Why? 18:30 MTDiscord because they're supposed to be edited from weblate 18:30 MTDiscord I mean English typos 18:31 MTDiscord Even English sentences? 18:31 MTDiscord then you would edit the source strings 18:31 MTDiscord Ues 18:31 MTDiscord Yes* 18:31 MTDiscord in the source code, and then the locale files will be regenerated during the next feature freeze 18:32 MTDiscord All English sentences in those files are strings sources? 18:32 MTDiscord So i can fix typos in them? 18:32 Pexin if the key string is changed, wouldnt that just detach the translation strings so they would need to be created again from scratch, or re-attached manually? 18:33 MTDiscord they will need to be retranslated 18:33 Pexin I strongly suggest not touching those strings in the sourcecode either 18:33 MTDiscord Oh 18:33 MTDiscord That's weird then 18:33 MTDiscord I can't speak much about the engine (it handles translations a bit differently) but for mods, the translations are always keyed purely on the English text, so changing so much as a bit of punctuation invalidates all translations of that string and requires it to be retranslated everywhere. 18:34 MTDiscord I wouldn't be surprised, at least, though, if the engine's own strings suffer similar limitations. 18:34 definitelya Abdou: You should read CONTRIBUTING.md in the minetest/.github directory. 18:34 MTDiscord I think i did 18:34 definitelya Nice 18:34 MTDiscord "That's weird then" ...that's just how translations with the english text as key work, not really weird 18:35 MTDiscord I mean, as warr said, that's a big limitation 18:35 MTDiscord If fixing a typo needs to retranslate the whole sentence 18:35 MTDiscord But well 18:36 MTDiscord I won't touch the po folder 18:36 Pexin also dont touch anything inside S() calls 18:36 MTDiscord O.O 18:36 MTDiscord What's that? 18:36 MTDiscord lua translation function 18:36 definitelya Weblate is equipped with an interface to keep things easy. 18:36 MTDiscord it exists in the builtin game code, the rest of builtin uses main menu gettext functions that interface with the engine 18:37 MTDiscord I did a little bit of translation from that website 18:37 MTDiscord A tiny little bit 18:37 MTDiscord I hope i will do more 18:38 MTDiscord So po folder and S() function, anything else? 18:38 MTDiscord The sanest way to do translations is to not put any text in the source code, only key codes, and look up the text in each language based on the key. Then the English text (or whatever language is the "original" language) and then translations can be based on that. That requires a lot of planning in advance, though, so it's unsurprising that a volunteer-run project will have painted itself into a corner long before that. 18:39 MTDiscord Forcing English as the source language sucks too for other reasons. It doesn't let a non-English-speaker publish anything in other languages and hope it gets translated using the existing infrastructure, and it doesn't let you do stuff like deal with different dialects of English. 18:39 muurkha also it's an extra level of indirection that can make it harder to get the first version running 18:40 MTDiscord I don't get S() calls. Like in the definition of the function? or the passed params? 18:40 muurkha like you make more mistakes taht way 18:42 muurkha but another problem that can happen with English as base is where the same English phrase needs to be translated in two different ways depending on context 18:43 MTDiscord Oh yeah, I ran into that problem too: in the website, it doesn't show you the context of the sentence to be translated 18:43 MTDiscord Sometimes you need to know the context so you can translate accurately 18:43 muurkha so for example in English the phrase "digging a hole" can be used as an adjectival phrase (I am a dwarf digging a hole) or a gerund (Digging a hole is rarely a good way to escape mobs) 18:44 muurkha but in Spanish those are "cavando un pozo" and "cavar un pozo" respectively 18:45 muurkha even without getting into questions like gender and number agreement, which are much less of a problem in English than in most Semitic and Indo-European languages 18:57 MTDiscord Ah, yeah, I forgot the english string reuse problem 18:58 MTDiscord the worst one is when you end up with a translation string like "@1 @2" :-D 18:58 MTDiscord like, maybe @1 is meant to be an adjective, and @2 is a noun, and this is just replacing the old code that used to just concatenate them. 18:58 MTDiscord So to translate that into Spanish it might make sense to do "@2 @1" to fix the noun/adjective order 18:59 MTDiscord but such a highly generic string could be reused elsewhere. 18:59 sfan5 that just string concatenation with extra steps 18:59 MTDiscord You could also use "login" as a noun to refer to the user's login name on a form, and the same "login" as a verb in a button to actually submit that form, so there's one screen that alone might not be translatable in some langauges. 19:00 MTDiscord Yeah, for the "@1 @2" case the compromise I found was just to do like "@1 noun" for each noun. You actually end up with 1 fewer total strings, and the same amount of work for translators, but less ambiguity. 19:00 MTDiscord It's still though just one more complication that people need to know going into this. 19:01 sfan5 still bad 19:01 sfan5 ideally you need each combination be translatable differently 19:01 MTDiscord I mostly did that in my own game, but it's a trade-off 19:02 MTDiscord It makes more work, and more tedious and monotonous work, for translators 19:02 MTDiscord if I have 5 things that come in 5 colors, I only have to program 10 things but translators have to translate 25, so it's kinda unfair. 19:04 Desour can't we just use german everywhere and force everyone to learn german? 19:05 muurkha Warr1024: that's an ingenious solution to the head-order problem 19:06 MTDiscord The problem with clever solutions though is that they can have surprising failure modes. 19:06 muurkha I think there are very few languages where "login" the noun is the same as "log in" the verb 19:07 MTDiscord In English "login" vs "log in" is kinda ambiguous. Spacing in a way is pretty flexible. People complain a lot when they see "alot" as one word, but we have "away" and "ahead" and other words that are similarly structured and readily accepted ... and now I've started seeing "awhile" a lot and I suspect that's going to become a thing too. 19:07 Pexin lo gin. going to liquor store. 19:08 muurkha alot: https://hyperboleandahalf.blogspot.com/2010/04/alot-is-better-than-you-at-everything.html 19:08 MTDiscord haha, yeah, the solution to every "your login form isn't working" bug is "well add more gin then" 19:08 muurkha I think "awhile" dates back several centuries 19:08 muurkha 700 years: https://www.etymonline.com/word/awhile#etymonline_v_19022 19:09 MTDiscord Well, it's been dormant but it's reawakaning 19:09 muurkha "newt" is also interesting in this connection: https://www.etymonline.com/word/newt 19:10 MTDiscord newts have been around long enough that they should be oldts by now. 19:10 MTDiscord what a wonderful website 19:10 MTDiscord https://cdn.discordapp.com/attachments/749727888659447960/1033094992694099998/unknown.png 19:11 MTDiscord imageshack is it that has put that "no war" image on expired images right? 19:11 muurkha has it? it looks like "awhile"'s low point was about 01985, at a fourth of its peak popularity in 01856: https://books.google.com/ngrams/graph?content=awhile&year_start=1600&year_end=2022&corpus=26&smoothing=3&direct_url=t1%3B%2Cawhile%3B%2Cc0#t1%3B%2Cawhile%3B%2Cc0 19:12 muurkha I mean that graph shows some much higher points in the 17th century but I think they might just be random noise due to having relatively few books scanned from that period 19:13 Pexin https://pics.me.me/noone-cares-made-on-imgur-peter-noone-will-always-care-49636940.png 19:14 muurkha aw :) 19:15 muurkha it looks like "awhile"'s meteoric resurrection began about 01998 and peaked in 02011, since which point it's been pretty steady 19:15 muurkha at least in books! it'd be interesting to know if the Usenet corpus yields divergent conclusions 19:17 muurkha "awhile"'s long doldrums in the 80s were apparently exceeded by even lower rates of usage in the 01730s, before its dramatic resurgence in the 19th century up to almost double current levels 19:18 muurkha in short, your prediction that "awhile" will "become a thing too" are surely correct ;) 19:18 muurkha *is 19:21 MTDiscord I mean, awhile from now, alot of words are probably going to be like this, but living in afuture like that might give me aheadache. 19:41 muurkha in Spanish like 10% of common words are like that, incorporating the Arabic article: alacrán, alfajor, alambique, almohada, albañil, albaricoque, maybe alcatraz, alcohol, alcaucil, álcali, alcalde... 19:43 MTDiscord We need a way for the serverlist to take not only quantity, but also quality, into account when ranking servers, comparable to how reviews now play a significant role alongside download counts on CDB. (e.g. I believe the crypto server on top of the list may have many players, but the experience of each player still is inferior to that of, say, players on CTF) 19:43 Desour oh, german also has Alkohol 19:43 muurkha "alcohol" is from "the kohl", extended to the general idea of substances purified by distillation, and then later restricted to wine thus purified 19:44 muurkha what are the German words for albaricoque and álcali? 19:44 MTDiscord We have Alkali 19:44 Desour there are alkali metals 19:44 muurkha yup, those are the ones 19:44 Desour like natrium 19:45 muurkha right. "sodium" in English, "sodio" in Spanish 19:45 Desour what's albaricoque in english? 19:45 muurkha "apricot", from Catalan "abercoc", which comes from Arabic "al birquq" 19:45 muurkha amusingly the Arabic word "birquq" "apricot" comes originally from Latin! 19:46 Desour how ironic 19:46 MTDiscord Aprikose 19:47 muurkha that's probably from French (which got it from Catalan and thence Arabic) 19:47 muurkha I'm guessing German has "alambique" too 19:47 muurkha Alembik 19:47 Desour btw. according to wikipedia, the english kohl is black dye: https://en.wikipedia.org/wiki/Kohl_(cosmetics) 19:48 muurkha anyway all of these are examples of the same kind of article assimilation as "awhile" (700 years ago) and "alot" (now) 19:48 muurkha yes, that was the original maning of Arabic "al-kuhúl" 19:48 Desour the german kohl is a plant (collard) 19:49 muurkha huh, I thought collards were a different species! 19:49 muurkha German Kohl entered English in "kohlrabi" 19:50 Desour Brassica*? 19:50 Desour kohlrabi is a special sort of kohl 19:51 muurkha right, but the other Brassica cultivars already had English names 19:51 Desour (kohl is a whole genus) 19:52 muurkha collards, cabbage, broccoli, cauliflower, kale, Brussels sprouts, kohlrabi — all the same species, no? 19:52 muurkha I think also "cole slaw" derives from German Kohl 19:54 Desour idk, I don't know much about all this stuff. I get most of my knowledge from wikipedia 19:58 muurkha the 01851 edition of Cooley's Cyclopædia says of "alcohol": "From the Arabic al the, and kohol antimony. A term originally applied to several chemical preparations, presumed to be very subtile, or brought to the highest state of tenuity, but at the present day restricted to pure spirit of wine, of the strongest class." 19:59 muurkha unfortunately I don't have any 17th-century alchemical reference books to consult 20:00 muurkha https://www.etymonline.com/word/cole-slaw says it comes from Dutch "koolsla" 20:01 muurkha but English used to have "col" "cabbage" up to only a few hundred years ago, and it survived in some dialects 20:15 muurkha apparently in German alcohol used to just be called "Weingeist": "wine ghost" 23:02 MTDiscord Does the MT build process do anything like -march=native optimizations or something that would cause a build on one CPU to fail to run on a different CPU of the same architecture? 23:03 MTDiscord e.g. if I build MT on a newer x86_64 system, can that cause it to dump core with "illegal instruction" on a much older x86_64 machine? 23:03 schwarzwald[m] `grep -march CMakeLists.txt` ? 23:04 schwarzwald[m] I think the answer is yes. 23:04 schwarzwald[m] `grep "-march" CMakeLists.txt`* 23:05 schwarzwald[m] Hmm, nope, can't find anything like that. 23:16 MTDiscord Hmm, I need to figure out what I need to do to build a version that works on older CPUs then 23:23 schwarzwald[m] Warr, if you have a fast CPU for compiling, I'd recommend doing a fresh build (clean CMake Cache) with the verbose flag. 23:23 schwarzwald[m] That has been my goto initial debugging approach for builds and it has served me well. 23:24 schwarzwald[m] Could direct the output into a log file and grep it. If a flag gets added anywhere, like in a third party, it should be in there. 23:32 schwarzwald[m] Another approach would be to debug the executable and figure out which instruction is invalid and where it is, I think 23:34 schwarzwald[m] if(IS_I686) 23:34 schwarzwald[m] message(STATUS "Detected Intel x86: using SSE instead of x87 FPU") 23:34 schwarzwald[m] add_compile_options(-mfpmath=sse -msse) 23:34 schwarzwald[m] endif() 23:34 schwarzwald[m] Warr, found some sus stuff in the Irrlicht build. 23:34 schwarzwald[m] Can someone ping him? 23:35 schwarzwald[m] There's a similar thing for MSVC. I'll bet Minetest has something similar... 23:36 schwarzwald[m] Yep, Minetest does this as well. This is 1 candidate. Doesn't strike me as particularly likely but it's a possibility. 23:36 schwarzwald[m] I would check the verbose build output. If it includes an `-sse` flag or something that's a likely culprit. 23:37 schwarzwald[m] If something like that is the issue, working around it would most likely entail modifying the build by hand to remove the part that adds the flag. It doesn't look like there's a build option for these.