Time Nick Message 00:18 est31 hmmmm, a question. There is srp_verifier_new and it writes to its param "const char ** username" with "*username = usr->username;". I dont need that value, I'd best like to pass NULL to the method. But then it will do dereference null right? The example makes a local variable only for this purpose, not using it afterwards. 00:19 paramat hmmmm thanks for the commit, compiling latest results in a warning and an error: https://gist.github.com/paramat/5ac23a1e649a157bd77a 00:20 est31 paramat, wait, you mean master? 00:21 paramat yes 00:21 * est31 tries 00:22 est31 ouch 00:22 * est31 can reproduce 00:22 paramat phew 00:23 est31 breaking compilation is very bad: first all PRs break too and when people bisect they run into problems 00:29 est31 somehow it compiles for one configuration, and for another it doesnt 00:30 est31 pushing a fix if nobody minds 00:32 est31 in 4 mins 00:32 est31 (changing std::ifstream is(filename, std::ios_base::binary); into std::ifstream is(filename.c_str(), std::ios_base::binary);) 00:35 est31 ah now I know the reason 00:36 est31 hmmmm required c++11 lol 00:38 est31 pushed 00:44 paramat thanks 00:46 paramat now compiles 01:08 hmmmm I did? 01:08 hmmmm est31: learn C 01:09 hmmmm est31: I'd have to be looking at it to give you a real answer 01:10 hmmmm paramat, sorry bout that will be fixing in the next patch 01:11 hmmmm oh... est31 already did it 01:11 * hmmmm grumbles 01:11 hmmmm okay i'll rebase. 01:12 paramat no problem 01:18 est31 hmmmm, nvm bout my question. Would have only spared one line. 01:40 hmmmm ask away 01:40 hmmmm if you have any language questions at all, please feel free to ask 01:40 hmmmm this is how you get better 01:42 hmmmm btw, you should have [WIP] in your PR title 01:42 est31 I have the WIP label 01:42 hmmmm ohh 01:42 hmmmm i didn't look at the label 01:43 hmmmm okay, looking... you mean ver->username? 01:44 hmmmm this is somewhat wacky code. the style is weird and inconsistent. 01:44 est31 usr->username; 01:44 est31 and space idents 01:44 hmmmm that's in srp_user_new, you said srp_verifier_new 01:44 est31 sorry meant that 01:45 hmmmm hrm 01:45 est31 correcting again: meant srp_user_start_authentication 01:45 hmmmm I don't get it 01:45 hmmmm usr->username is a const char *, but then he disrespects that a couple of lines below by casting it to (char *) 01:46 hmmmm this doesn't seem like very high quality code to me... 01:48 hmmmm are you sure you don't need "username"? 01:48 est31 yes 01:48 hmmmm don't get what the point is of extracting it from SRPUser 01:48 est31 it has only been added so that you only have to store the SRPUser 01:49 est31 you can call ..._new at some other point 01:49 est31 then that method extracts it for you 01:49 hmmmm what the heck is the point of extracting the username from the SRPUser 01:49 hmmmm this makes no sense 01:50 hmmmm look, it does nothing but set username to the contents of that struct 01:50 hmmmm the caller could've just used usr->username if he wanted that 01:50 hmmmm I am starting to have my doubts about this library... 01:50 est31 the struct is "internal" 01:50 est31 it is only defined in the c file 01:51 hmmmm ahh i see 01:51 hmmmm I guess it makes sense then a little 01:51 hmmmm ahh that's why it's a const char 01:51 est31 ? 01:52 hmmmm okay, so if you're not using that value, you'd have to add an if statement to that 01:52 hmmmm you're saving a line, sure, but you're adding it somewhere else 01:53 hmmmm that's the way optional pointer parameters work. if (username)\n\t*username = thing; 01:53 est31 ok thanks 01:53 hmmmm yeh your question makes total sense now with context 02:04 est31 argh 02:04 est31 translatePassword has wstring as its second param 02:05 est31 and in most places it is used, there is narrow_to_wide(param2) 02:05 est31 internally, its doing wide_to_narrow 02:05 est31 gonna refactor this 02:12 est31 hmmmm, do you plan a second patch? 02:12 est31 l_mapgen.cpp:1042:19: warning: unused variable ‘ndef’ [-Wunused-variable] 02:12 est31 INodeDefManager *ndef = getServer(L)->getNodeDefManager(); 02:15 hmmmm ...yes. 02:15 est31 btw something else broke too "WARNING: Relative path not properly supported on this platform" 02:15 hmmmm you don't need to fix every single fart and cough within 5 minutes 02:15 est31 hmmmm, agreed 02:15 est31 that warning has time 02:16 est31 failure to build... bad 02:16 est31 due to the relative path problem I cant start the client 02:16 hmmmm that's not caused by my changes, that's for sure 02:17 est31 I'll fix it either way 02:17 est31 (not the warning) 02:17 est31 but the path detection problem 02:18 hmmmm okay 02:50 est31 gcc is doing quite weird stuff here 02:51 est31 clang too 02:52 est31 I have found out that the issue was caused because when you pass "--std=c++11" to the compiler, the preprocessor macro "linux" isn't defined anymore 02:52 est31 so this makes the code in porting.cpp believe we aren't on linux 02:52 est31 but it isn't limited to ++11 02:53 est31 99 has the same issue 02:53 est31 or 98 for c++ 02:55 hmmmm hrmm 02:55 hmmmm no comment 03:04 est31 pushing in 15 minutes http://pastie.org/10097239 03:04 hmmmm looks good to me 03:04 est31 then pushing faster :) 03:04 hmmmm in fact, I think the check for __linux was already there at one point 03:05 est31 blame it :p 03:05 hmmmm what happened to it? did somebody remove that thinking they're simplifying code? 03:07 hmmmm humm no 03:07 hmmmm it was always like that 03:08 hmmmm i added __linux to the porting.h things 03:10 est31 pushing now 03:11 est31 ugs 03:11 est31 srry wasnt compilation 03:12 est31 but path detection 03:12 est31 pushed 04:07 est31 Merging this small refactor in 15 minutes: http://pastie.org/10097280 04:07 est31 (removes trailing spaces & spares some string copies and conversions) 04:11 Zeno` :-o 04:16 Zeno` lgtm 04:17 est31 pushing now 05:03 hmmmm https://github.com/minetest/minetest/commit/3576f4fab7267ff6d110fe94b6f397d6471a6980 05:03 hmmmm https://github.com/minetest/minetest/commit/406d9ba87b9f6e57b86c6282bf157e3341aa195c 05:12 est31 ok my srp port now passes through valgrind 05:15 est31 needs ~40 MB per run, this is very disturbing. 05:15 est31 either its gmp's fault or the hash function's 05:16 est31 (valgrind shows this under "$num allocated") 05:16 est31 err 05:16 est31 40 KB 05:17 est31 that seems ok to me 05:33 hmmmm night. gotta get to sleep early because of an early-ass doctor's appointment 05:34 Zeno` have fun 05:34 hmmmm later 05:34 hmmmm pretty soon i'm going to take a break from FUCKIN' schematics 05:34 hmmmm i am frankly tired of it and it's in a decent spot right now 08:22 est31 #2618 08:22 ShadowBot https://github.com/minetest/minetest/issues/2618 -- 20s timeout when connecting to server by SmallJoker 08:32 est31 pushing fix that Krock suggested 08:33 est31 pushed 08:33 Krock So you include a new header to get that single line working? :3 08:33 est31 } catch { doesnt compile for me 08:33 est31 so I included that header 08:33 Krock oh. maybe again a fault from my C# experience 08:37 est31 about your PR Krock, what was the result of your discussion with hmmm? 08:38 Krock est31, I added some lines to set last_time to the current time. It's fine now. 08:45 est31 +1ed it 08:47 Krock pls4push 08:47 est31 if you can point me to a comment by hmmm where he agrees on your PR, I'll do it 08:47 est31 or any other coredev 08:51 Krock http://irc.minetest.ru/minetest-dev/2015-04-16#i_4225474 and a day earlier: http://irc.minetest.ru/minetest-dev/2015-04-15#i_4223935 08:52 est31 Thats not 100% what I need. 08:53 est31 but I don't see how it could harm people 08:53 est31 Its not such a big change 08:54 Krock It's tested and it works without problems 08:55 est31 thats not the issue 08:55 est31 the issue is that there is a rule 08:55 est31 2 coredevs have to agree 08:56 est31 I think it should be changed to the point: coredevs can push things upstream if they are obvious or trivial. 08:56 est31 (I mean PRs, the rule already exists for own code) 09:02 est31 but I guess those two comments you linked qualify as +1 09:03 Krock kinda much trouble for that simple change :3 09:27 Zeno` Krock: https://github.com/minetest/minetest/blob/f5ae8e837e854007f1164ccd81872b6fe506a69d/src/game.cpp#L2245 09:28 Zeno` ^-- why is that line necessary? 09:31 Zeno` I don't *think* it will cause problems, but strictly speaking getTimer()->getTime() should not be called unless immediately after the timer has been stepped or device->run() has been called 09:31 Zeno` there was no problem without it there 09:32 Zeno` (it won't crash or anything, I'm just not sure if there will be unwanted side-effects... there were none before it was added) 09:32 Zeno` oh well 15:36 hmmmm krock wanted to make sure that last_time was set to the current time if it had not been initialized 15:36 hmmmm so he tried something like last_time = fpscontrol->last_time or current_time 15:37 hmmmm but that doesn't translate to C 15:37 hmmmm fpscontrol is zeroed out so the last time is always 0 on the first run, maxing out the dtime when limitFps is called 15:38 hmmmm but really fpscontrol->last_time should be initialized to the current time so there's no jump on the first cycle 17:42 Sokomine hmmmm: i've got a problem with the current git version. minetest.create_schematic( start_pos, end_pos, nil, filename, nil) crashes the game 17:49 Sokomine hmmmm: it also crashes when using worldedit (mtschemcreate) 19:36 hmmmm hmm 19:36 hmmmm i'll take a look in a bit 19:51 Sokomine thank you 22:38 hmmmm oh yes oh yes! 22:40 hmmmm sokomine, fixed 22:40 hmmmm hmm odd 22:41 hmmmm i did not know you needed typeinfo and it worked for me so *shrug* 22:44 paramat updated and simplified PR for river water game#493 22:44 ShadowBot https://github.com/minetest/minetest_game/issues/493 -- Default: Add river water by paramat 22:44 hmmmm it seems to be getting backlash 22:47 hmmmm yeah, if we were to implement these kinds of physics in the engine, how would it be done.. good question 22:55 paramat the resistence in the comments was to the original PR which had fresh water, fresh river water and mixed water, i now agree that was too complex 22:57 paramat if anyone knows the best way for a bucket of normal water to be returned when bucketing river water, let me know, i'll add to this PR 23:00 Sokomine hmmmm: it works again. thanks :-)