Time Nick Message 07:27 nerzhul sfan5: #9769 is now ready to be merged. I restored the old initialization model, now we can merge it 07:27 ShadowBot https://github.com/minetest/minetest/issues/9769 -- Server class code cleanups by nerzhul 17:13 nerzhul sfan5, Krock, #9769 waits :( i need it before continuing my work, else it will be rebase powa 17:13 ShadowBot https://github.com/minetest/minetest/issues/9769 -- Server class code cleanups by nerzhul 17:26 sfan5 nerzhul: I didn't test that code snippet, not sure if that compiles 17:27 nerzhul testing now 17:27 nerzhul the function constness is fine 17:27 nerzhul no it doesn't, will fix & push the fix 17:28 nerzhul okay last comments are adressed 17:28 nerzhul the original code was totally out performant with many lookups, now it's pretty nice :D 17:28 nerzhul thanks for the various returns, i should have audited the original code, i will do that on next code move, not just reorganize 17:29 sfan5 doubt it makes a noticeable difference, but yeah no need to do it inefficiently 17:29 sfan5 did you push the changes yet? 17:29 nerzhul yep 17:29 nerzhul https://github.com/minetest/minetest/pull/9769/commits/58bb159f5af9cab91d8c16275961ed0cd9549fdd 17:30 sfan5 I see 17:31 sfan5 2020-05-07 19:31:09: WARNING[Main]: GUIInventoryList::draw(): The inventory location "detached:creative_trash" doesn't exist 17:31 sfan5 smells like a bug 17:31 nerzhul fresh minetest_game ? 17:32 nerzhul maybe before we sent all like ugly boys now we send only the real thing people wants to send this show bugs ? 17:32 sfan5 yes fresh minetest_game 17:33 nerzhul okay let me try, i tested on minimal was fine 17:33 sfan5 creative_trash should have owner == "" and be sent to everyone 17:33 nerzhul okay let me check 17:34 Krock INVALID peer ID and owner=="" 17:34 Krock it might skip/return too early somewhere 17:35 nerzhul yep 17:38 nerzhul where did you see that ? 17:38 nerzhul i have the Krock one 17:39 nerzhul note 17:39 nerzhul for the Krock warning, it doens't exist before 17:39 nerzhul Krock can you add the same warning on master and verify if it's present ? 17:39 sfan5 2020-05-07 19:30:59: WARNING[Server]: Failed to send detached inventories to peer_id=2, player not found. 17:39 sfan5 oh interesting 17:40 nerzhul see the diff from there: https://github.com/minetest/minetest/pull/9769/files#diff-ad60d65b34e16a3319296bb5d683acd6L2763 17:40 nerzhul and precisely here: https://github.com/minetest/minetest/pull/9769/files#diff-ad60d65b34e16a3319296bb5d683acd6L2717 (where the check was done before) 17:41 nerzhul we just ignore offline players 17:41 Krock nerzhul: sorry, didn't see it. just assumed it was likely that problem 17:41 nerzhul i added because if wanted to see if there is regression, but it seems not 17:41 nerzhul seems we just ignored and send them inventories after no ? 17:42 nerzhul the only problem is the inventory location sfan5 shown, how can i reproduce it ? 17:42 sfan5 1) enable creative mode 2) join minetest_game world 3) open inventory 17:42 nerzhul exact i got it after adding the priv 17:48 nerzhul ok found haha 17:48 nerzhul ... 17:48 nerzhul or not 17:51 nerzhul i think the problem is the peer id not found 17:52 nerzhul in fact we send the full inventory when peer is not found then we do incremental but there is nothing to push 17:53 nerzhul you know what, i will send the inventories on ready client and not in init2, where session is not ready 17:53 nerzhul there is no ense to do that early 17:53 nerzhul yeah now it's fine 17:54 nerzhul i don't know what in the previous spaghetti work make it work, but now it's clean we break thing hidden by old code 17:54 nerzhul and seinding the inventories with the playerlist when session is ready is better than seinding before entering world 17:55 nerzhul i pushed the code change, Krock, sfan5 can you test ? 18:05 sfan5 I don't see a commit 18:05 nerzhul oops pushed on wrong remote, it's fine now 18:14 sfan5 nerzhul: I prefer this fix https://0x0.st/i_id.png 18:14 sfan5 then the sending does not have to be moved to happen later 18:15 Krock is this thread safe? 18:15 Krock speaking of getClient()->getName() 18:16 sfan5 getClient is called from the serverpackethandler everywhere else without locks too 18:16 Krock that aside, I'd also prefer this suggested variant 18:17 nerzhul okay fine to me 18:17 nerzhul let me test it 18:17 nerzhul crazy refacto pr, again :D 18:17 nerzhul i hope the api abstraction layer will not last so long as there is not many things :D* 18:18 nerzhul commit pushed & test 18:18 nerzhul it's fine 18:19 nerzhul i really want to rewrok the session model of minuetest at a point 18:19 nerzhul as you see in this PR there is 2 way to gather the name, and they are not the same, with not the ame implications 18:20 sfan5 the client vs. player distinction is expected to be relevant when you deal with code that runs before the player has fully joined 18:20 nerzhul yep, i totally agree with that, but it's a nice melting pot in the say we develop 18:21 nerzhul the duplicate name (and maybe other thing) between the session and the player object is crazy in terms of dev 18:24 Krock RemotePlayer and RemoteClient should be merged at some point 18:25 Krock where the result contains a PlayerSAO pointer whenever the player is done initializing 18:46 nerzhul yep i think a thing like this is to be done 18:46 nerzhul it's no so easy 19:07 Syntaxxor Hey, all. I've decided to get back into working on shadowmapping. Been a while since my last attempt. Right now, my idea is to work with the plain renderer. Is there any issue with this? Should I be trying to work on the core renderer or something? 19:09 sfan5 uh, which renderer are you referring to? 19:09 sfan5 ah src/client/render 19:10 sfan5 shouldn't really matter where the code ends up, not like it's impossible to move later 19:14 Syntaxxor Alright. I actually realized that I should change what I'm doing so I'm moving it to the core renderer. 19:41 sfan5 Krock: http://sprunge.us/upwKNh?diff 19:43 Syntaxxor How does one add an option? 19:43 sfan5 in minetest.conf? 19:46 Syntaxxor Was trying to find that. Realized you need to close the game once first. Thanks! 19:51 Syntaxxor Alright. That doesn't exactly seem to be what I'm looking for. I would like to add an option for setting the shadow quality. It's not super important but would be nice to have. 19:52 sfan5 you add a default value to defaultsettings.cpp, you can fetch it any point using the right g_settings->get() method 19:52 sfan5 and to properly document the setting you can add it to settingtypes.txt (later) 20:01 Krock sfan5: LGTM 20:01 sfan5 thanks, pushing that 20:07 Syntaxxor That's what I needed. Thanks, sfan5! 20:33 sfan5 reminder that there's quite a few PRs in need of a second review https://github.com/minetest/minetest/pulls?q=is%3Aopen+is%3Apr+label%3A%22One+approval%22 20:35 sfan5 it might also be a good idea to start deciding which features need/should go into 5.3 20:37 nerzhul merging #9769 20:38 ShadowBot https://github.com/minetest/minetest/issues/9769 -- Server class code cleanups by nerzhul 20:38 nerzhul when do we expect a release for that ? 20:38 nerzhul i'd like to see my abstract API pr finished & merged before 20:43 sfan5 5.3 was originally target to be roughly one month after 5.2 20:55 nerzhul WE ARE A LITTLE BIT LATE HAHA 20:55 nerzhul oops caps 20:55 nerzhul why so close ? and not just a patch release branch ? 20:56 rubenwardy the changes to fix android aren't just bug fixes 20:56 rubenwardy and non-trivial back porting ends up painful 20:56 rubenwardy release early, release often 21:12 sfan5 Krock: aren't the type check errors supposed to have a backtrace? 21:12 sfan5 wait nvm 21:49 nerzhul rubenwardy i agree with you, if we don't do 1 month dev and 1 month QA freeze, PR merge is painful in terms of time it will be frustrating for contributors 21:50 nerzhul release often means be reactive on bug, but can make linux package owners crazy :D 22:14 Syntaxxor As I'm working with this, I'm remembering why I wasn't able to last time. Mainly, Irrlicht has little to no control over the actual rendering process and framebuffers are nonexistent. Is there any way around this? 22:41 Syntaxxor How can I print a value to the in-game console?