Time Nick Message 00:39 paramat merging #8867 00:39 ShadowBot https://github.com/minetest/minetest/issues/8867 -- Various settingtypes.txt improvements by paramat 02:05 paramat merging game#2449 note: because Krock indicated support for a merge in IRC earlier 02:05 ShadowBot https://github.com/minetest/minetest_game/issues/2449 -- add_new_tnt_sounds by TumeniNodes 02:19 paramat oh and game#2452 thanks for reviewing =) 02:19 ShadowBot https://github.com/minetest/minetest_game/issues/2452 -- Spawn: Avoid spawning outside small worlds by paramat 10:21 ANAND I just realised what QuickTune is, and I'm amazed we have such a system in place for testing and debugging stuff. Kudos to C55 for this idea! 11:02 BuckarooBanzai ANAND: whats QuickTune? 11:49 ANAND BuckarooBanzai: It's a way to modify (numeric) variables at run-time, without having to recompile the code after every single modification. 11:50 ANAND It makes use of the quicktune_blah_blah keymaps, which I religiously ignored every time I open the "change keys" dialog. 11:52 ANAND See https://github.com/minetest/minetest/blob/4291961e09db7851fbc940eac58a1b919668d935/src/quicktune.h#L21 11:53 ANAND The example in L25 is exactly what I'm going to use this for, first. ;) 11:54 ANAND i.e. #8755 11:54 ShadowBot https://github.com/minetest/minetest/issues/8755 -- ItemDefinition: Add wield_offset and wield_rotation by ClobberXD 12:01 BuckarooBanzai Oh nice, macro magic :) 12:01 BuckarooBanzai Really useful function though.. 13:20 ANAND Indeed :) 15:27 DS-minetest is it ok if the whole mapblock is sent when block meta is sent? 15:28 DS-minetest set* 15:35 sfan5 probably 15:38 rubenwardy in the latest devs, the graphs are in a different order 15:38 rubenwardy this is messing me up 15:42 sfan5 Update frame [ms] and FPS are kinda redundant as graphs 15:43 sfan5 since one is 1/x of the other 15:47 rubenwardy FPS would include draw time too? 15:51 sfan5 hmm 16:05 DS-minetest is it a problem if a PR title lets people think that there's a new unread comment? 16:06 DS-minetest #8870 16:06 ShadowBot https://github.com/minetest/minetest/issues/8870 -- Block meta (1) by DS-Minetest 17:52 ANAND Krock: Re #8856, std::cout doesn't print anything to the console. I faced this issue for some of my PRs too, and I ended up using warningstream and/or errorstream instead. 17:52 ShadowBot https://github.com/minetest/minetest/issues/8856 -- Send cumulated inventory changes only each step by SmallJoker 17:52 ANAND I did revert the "Remove log messages" commit, btw 17:52 Krock cannot reproduce 17:52 Krock cout works perfectly for me 17:53 ANAND Meh 17:53 Krock do you have any local changes that redirect cout to another stream? 17:53 ANAND Not as far as I know 17:54 Krock does it also happen with xterm? 17:54 ANAND But what I've observed is that it correctly prints to the console when inserted in some parts of the code, while it doesn't, when inserted in other parts of the code 17:54 Krock "console" implies windows. in this case -- powershell? 17:54 ANAND I use Ubuntu 18.04 17:55 Krock same here 17:55 ANAND Weird 17:55 ANAND But the aforementioned log streams are pretty consistent 17:57 ANAND Krock: Does #8834 warrant a protocol bump? I'm still not exactly sure as to when the protocol is bumped, sorry. 17:57 ShadowBot https://github.com/minetest/minetest/issues/8834 -- [no squash] Send ActiveObjects along with definitions once connection is established by ClobberXD 17:58 Krock chceking 17:58 ANAND Thanks 18:00 ANAND Network-wise, the PR sends an existing unmodified packet to the client upon the latter's completion of Init2 18:00 ANAND Proto bump probably not required, but just wanted to be sure... 18:07 Krock noticed something else: why do you run getClient() twice in INit2 18:09 Krock oh no 18:10 Krock > m_state = CS_InitDone; @notifyEvent 18:10 Krock This one sits behind an event queue. TOSERVER_INIT2 adds this entry and now you're checking in the very same function whether it's set 18:10 Krock dangerous 18:11 ANAND Uh oh 18:11 Krock hold on. event() usunlike other events not a queue. iconic. 18:11 ANAND Oh yeah, it's just a bunch of nested switch-cases 18:12 ANAND The second getClient is to ensure that `client` is properly initialised in case the code skipped the previous getClient call 18:12 ANAND Dunno why I thought that way 18:13 Krock it looks like the inner loop code for sending entities is kept functionally unmodified 18:13 Krock in this case it's entirely compatible with the protocol 18:14 Krock You also assume that `client` is valid after the second call, and use the pointer without checking to get the protocol version 18:14 ANAND Yes, that's correct - I just moved the AO retrieval code to a separate function. Thanks for confirming. 18:14 Krock it was already like that before, so the first getClient() call ought to be enough for the entire thing 18:15 ANAND Yes, I'm removing the second one 18:21 ANAND Krock: https://github.com/minetest/minetest/compare/838bc8b949a60ec9a33b4a6cbbda1e0af7eedbba..06ccf05b89b479997dee99d5fb96faa91865f16a 18:22 Krock thanks 18:22 ANAND Re #8856, doesn't Lua inventories get modified at most only once every step? 18:22 ShadowBot https://github.com/minetest/minetest/issues/8856 -- Send cumulated inventory changes only each step by SmallJoker 18:23 Krock see example 18:23 ANAND I saw, but I don't understand why it is so :) 18:24 Krock there's plenty of calls within the same step. each inventory modification by Lua immediately calls to SendInventory 18:24 ANAND Oh! 18:24 ANAND Makes sense now 18:24 Krock !next 18:24 ShadowBot Another satisfied customer. Next! 18:24 ANAND I somehow saw that coming :P