Time Nick Message 01:27 sofar using the bundled jsoncpp, 5.0.0 doesn't compile o_O 01:27 sofar link, actually 01:37 paramat merging game#2327 to master in 5 mins 01:37 ShadowBot https://github.com/minetest/minetest_game/issues/2327 -- Properly count players in beds. by sofar 01:45 paramat done 01:46 GreenDimond oof ShadowBot is even slower 01:46 GreenDimond wait 01:46 GreenDimond no 01:46 paramat merging #8258 01:46 GreenDimond disgregard 01:46 ShadowBot https://github.com/minetest/minetest/issues/8258 -- Confirm registration GUI: Remove positional strings to fix Windows bug by paramat 01:47 GreenDimond I cant fking type 01:53 paramat oops that PR should be merged to 5.0.1 too 01:53 paramat so to branch 'backport-5' 01:54 paramat sorry. can anyone do that? 01:54 paramat #8258 01:54 ShadowBot https://github.com/minetest/minetest/issues/8258 -- Confirm registration GUI: Remove positional strings to fix Windows bug by paramat 02:11 p_gimeno paramat: may I PM you? 02:12 paramat yes 08:16 paramat the backport-5 branch still has '5.1.0' in client_lua_api.txt and menu_lua_api.txt 09:35 Krock #8350 09:35 ShadowBot https://github.com/minetest/minetest/issues/8350 -- Damage: Play no damage sound when immortal by SmallJoker 09:43 Krock merging #8306 and #8321 in ~15 mins 09:44 ShadowBot https://github.com/minetest/minetest/issues/8306 -- Document the `float` special group by DS-Minetest 09:44 ShadowBot https://github.com/minetest/minetest/issues/8321 -- Display pitch angle in debug menu by xerox123official 09:44 nerzhul nice 10:13 Krock merging.. 13:02 Krock p_gimeno: handling the deprecated functions in Lua would be much slower (ignoring duplicates) and aliased functions would have to be provided to builtin so that it could modify the metatable 13:02 Krock well, if that's even possible. C++ creates a new reference for executed callbacks, so adjusting the metatable might be not that easy 13:03 p_gimeno Krock: logging the deprecation is already slow anyway 13:03 p_gimeno I may work on a patch later 13:03 Krock the deprecated function handling should be less than a millisecond 13:05 Krock heck, 8us according to my tests using a debug build 13:05 Krock (June 2018) 15:27 p_gimeno Lua 5.0 is not supported, right? 15:27 nerzhul no 15:27 nerzhul we are in 5.1 only 15:28 p_gimeno ok, I ask because I see calls to luaL_openlib which was replaced by luaL_register in 5.1 15:42 p_gimeno *sigh* #8322 is yet another case where a fix for a tiny but critical problem is going to be so huge and touch so many areas, that no one will want to review it in time before conflicts are introduced, it will get dust and get harder and harder to rebase 15:42 ShadowBot https://github.com/minetest/minetest/issues/8322 -- deprecated_lua_api_handling = log causes crash for tail calls of deprecated functions 15:43 p_gimeno that's what frustrates me the most, there are areas that really need a redesign but redesigns are not expedited quickly 15:44 p_gimeno (if at all) 17:12 nerzhul yes it's difficult without breaking things 18:40 rubenwardy #8353 18:40 ShadowBot https://github.com/minetest/minetest/issues/8353 -- Fix serialization of std::time_t by casting to u64 first by rubenwardy 18:42 Krock rubenwardy: LGTM 18:43 rubenwardy cool, merging in 10 18:44 rubenwardy it's basically trivial 18:44 rubenwardy oh wait 18:44 rubenwardy I'll change that (u64) to a static_Cast 18:44 Krock how come? The datatype is of equal size, if not bigger 18:45 rubenwardy actually, maybe not 18:45 rubenwardy probably better to be consistent 18:45 rubenwardy C style casts are supposedly deprecated 18:58 rubenwardy another one #8354 18:58 ShadowBot https://github.com/minetest/minetest/issues/8354 -- Fix cast from const by accessing string data directly by rubenwardy 18:59 rubenwardy not sure if &str[0] is defined 18:59 rubenwardy probably is 19:01 rubenwardy [FAIL] testStreamRead - 0ms hmm 19:02 rubenwardy oh, that fails on master too 19:02 Krock were the unittests not updated after some recent changes? 19:05 rubenwardy ah, &str[0] is the same as doing &ptr[0] on a char* 19:06 p_gimeno &str[0] is the same as str, isn't it? 19:06 rubenwardy p_gimeno: no, as it gets a char* 19:06 rubenwardy this is core::string btw 19:06 rubenwardy like, irr::core::string 19:06 p_gimeno ah 19:07 rubenwardy the definition of [] is T& operator[]](const u32 index) { return array[index]; } 19:07 rubenwardy -] 20:20 nerzhul why not fix the serialization in the packet itself ? 20:20 nerzhul i'm not very fan about this fix 20:20 nerzhul oh you mean there is a truncation using the operator ? 20:24 rubenwardy nerzhul: compile error because std::time_t **is** u64 20:24 rubenwardy on some platform 20:24 rubenwardy also, time_t is u32 on some platforms 20:24 rubenwardy so it would end up with a bad packet 20:27 nerzhul right 20:30 nerzhul not fixed on all platform until 2038 :( 22:39 rubenwardy nerzhul: with my two PRs today, there's only one issue left in the milestone for 5.0.1 22:39 rubenwardy the noise segfault 22:41 sofar I've noticed that if you do any debug output during mapgen in a thread, you will hit the segfault quickly 22:42 sofar I wonder if lua causes it to crash 22:54 p_gimeno Thread interaction can cause all sorts of issues. I remember experiencing a very weird problem that could perhaps be explained by insufficient thread locking, #5375. I blamed the compiler at the time because the symptoms were crazy. 22:54 ShadowBot https://github.com/minetest/minetest/issues/5375 -- Wheat seeds grow as cotton (wrong ABM seems to be executed on catch-up)