Time Nick Message 02:11 Exio by looking around, wouldn't be https://github.com/minetest/minetest/commit/f9a8efb99235c86eb61b6deb2bc034d5618e45ea useless with the cap in the speed? 02:53 Miner_48er any experts? 02:55 kaeza just ask your question 02:56 Miner_48er I need a way to make exceptions to the ip ban for certain usernames on shared ips. Any way to do this? 03:02 ShadowNinja Miner_48er: Not the channel but /mode +e 03:04 Exio /mode +e in MT? 03:04 Miner_48er ? 03:04 Miner_48er which channel? 03:05 ShadowNinja Wait, channel or Minetest? 03:05 ShadowNinja Miner_48er: PM. 03:08 Miner_48er ok 15:29 sfan5 We really need a /save command that forces saving the map, maybe also a command that saves the map as map_.sqlite so you can backup your world without resrtarting 17:11 celeron55 sfan5: more like /backup that saves a full "worldname.backup" 17:12 sfan5 celeron55: that was I meant 17:13 celeron55 it's okay to me at least 17:13 celeron55 i can think of a thing that will conflict with this, but it might be quite out of date already anyway: the database backend selection and leveldb 17:16 sfan5 I think a /backup is really urgently needed since, its not save to access map.sqlite while the server is running 17:16 sfan5 I accidently copied/tar'd map.sqlite while some server was running like 3 times and nothing happend.. 17:17 sfan5 but a "/backup" would be safer 17:17 celeron55 i have once copied a world between a running server and a development machine and it ended up corruped 17:17 celeron55 so yes, it isn't safe 17:17 PilzAdam will the server freeze while /backup is runnig? 17:18 sfan5 yes 17:18 sfan5 an env- and maplock is requires 17:18 celeron55 you need to think of the requirements of the people who want to automate their backups though 17:18 sfan5 *required 17:18 celeron55 ...and also the freeze too 17:18 PilzAdam have you any idea how long? 17:18 sfan5 well 17:19 celeron55 a lock file could make much sense 17:19 celeron55 in a world 17:19 sfan5 how it could be (maybe not 100% safe): write latest changes to map.sqlite, copy map.sqlite, resume normal operation 17:19 celeron55 but it kind of doesn't solve this problem 17:19 celeron55 sfan5: sounds good 17:20 sfan5 (assuming that the world doesn't corrupt when its simply open with sqlite [not written and not read]) 17:20 celeron55 the command processing should happen in lua, so there will be a lua api for making backups that people can use to make an automatic backupper mod 17:20 celeron55 -> win 17:21 sfan5 ^ what I just wanted to say 17:21 sfan5 something like minetest.save_map_backup() -> described above ; minetest.save_map() -> just writes tha latest changes to disk 17:21 sfan5 *the 17:23 sfan5 also: any comment about an sha1 api for lua? https://github.com/sfan5/minetest/tree/sha1_scriptapi 17:23 sfan5 *the 17:25 celeron55 what's the use for it? 17:25 sfan5 .. 17:25 celeron55 why is that a bad question? :P 17:25 sfan5 may come useful for a mod that auto-copied all mod resources to the client/cache folder 17:25 sfan5 no 17:26 sfan5 *auto-copies 17:26 celeron55 sounds hacky... i wonder if there is something non-hacky 17:27 sfan5 nah.. its more useful than its hacky ;) 17:27 celeron55 i guess some kind of content-based resource management, but i don't immediately come up with any such 17:27 celeron55 but effort should be focused on making minetest itself manage mod resources well 17:27 celeron55 not hacking around the core 17:27 sfan5 just an idea 17:28 sfan5 but maybe someone comes up with this super awesome idea© that needs sha1 17:28 sfan5 (in x*math.PI years) 17:28 celeron55 but including features that nobody needs is bad practice 17:30 celeron55 i think we need a "may be removed if nobody finds good use for this" flag in the documentation 17:30 sfan5 yeah 17:31 celeron55 altough, maybe it's kind of implied; but i fear that people will use sha1 without any good reason if it's put in without any warnings 17:31 celeron55 and also it can make people make all kinds of false security attempts that don't benefit anyone 17:31 sfan5 "-- Do use with care" 17:56 sapier Exio no speed limiting isn't enough as in exactly this step speed is calculated from acceleration and dtime. The way it's done atm doesn't limit acceleration to a fixed value. the faster your server is the higher accelerations are possible. 17:57 sapier an believe me this is working it fixed a problem beeing present way to long 17:57 Exio kk 17:59 sapier :-) sorry if I've been a little bit to harsh it was a hard fight to get this in and mobf won't work at all without it as mobf uses acceleration very often. without this fix a single overload situation can cause runnaway collision handling times 17:59 Exio you didn't soun harsh 17:59 Exio sound* 18:00 sapier ok good :-) 18:00 Exio but as the problem with super-accelerated-player => high ram / cpu usage 18:00 Exio (== "what that solved too") 18:00 sapier could be yes 18:00 Exio it seems like two fixes for the same time :P 18:01 sapier it generally fixes any runnaway collision handling due to overload situations and acceleration 18:02 sapier I don't like hard limits as they reduce flexibility ... this patch only depends on servers capability to deal with collision handling. if server is able to calculate accelerations > 100 in less than half a second this is still possible ... but I guess any object beeing accelerated that fast will instantly be unloaded :-) 18:03 PilzAdam game.cpp:2698 <- yes, should be removed IMO 18:04 Exio doit? :P 18:04 PilzAdam anyone against it? 18:04 Exio if it is deprecated and not used in any "mod" 18:05 sapier true PilzAdam does sign already use formspec or is this tbd too? 18:06 sapier 2719 is wrong too :-) "custom inventory view" doesn't quite fit current formspec capabilities 18:07 sapier PilzAdam do you have any issues left for formspec menu? 18:08 PilzAdam still havent tested 18:08 PilzAdam and minetest_game and minimal use real formspec since 0.4.3 or so 18:09 Exio https://github.com/minetest/minetest/pull/741 18:09 Exio what's up with that? 18:09 sapier should be checked and merged 18:10 sapier somehow i forgot my own additions on rebasing ;-) 18:10 Exio and #706 seems something what "should be tested" (just for adding support) 18:11 Exio but rebased? 18:11 Exio nvm, it only uses hud stuff