Time Nick Message 00:27 paramat merging 9910 01:24 ANAND Minimal HP change test assertions are failing again 01:24 ANAND I'm on 7ec76e53 01:25 ANAND That's probably just a couple of commits behind master 01:27 ANAND I "fixed" this by deleting the old world and creating a new one, but I'm getting this on the new world too. 01:28 ANAND I remember my HP being pretty low, the last time I was able to play in the world 01:29 ANAND I'm thinking the assertion failure happens bc the code doesn't properly handle the case of player having low HP already 08:26 sfan5 ANAND: those assertions also fail when you have damage disabled 08:56 nephele Any thoughts on a drawing api for mt? Basically a "hudpaint" or so callback in which a couple of calls are valid to precisely draw the hud, as in draw a line, draw a texture, draw a text (and get the drawn size), draw a color in a given space (like a square) etc 08:59 sfan5 I doubt that such direct drawing is going to be implemented for performance, security or other reasons 09:24 nephele I am just testing the water, why do you think that performance and security are issues here? Personally i would like such an api to be able to efficiently draw a hud without bending over backwards to use the hud elements there are now, so i am not really sure why performance would be an issue, likewise security 10:46 sfan5 nephele: performance would be because the client would have to call into Lua to draw the GUI very often, security is because you'd be exposing additional APIs but I don't expect that to be an issue in practice (it just has to be considered) 10:48 nephele it's only 60times per second usually, i don't expect the performance to be horrible (and if the drawer has a buffer he can just return that immidiently if nothing in his model has changed) 10:52 ANAND sfan5: I hadn't disabled damage in either my old minimal test world or my new one. While I didn't specifically note the events that led to the assertion fail in my old world, I did note that I had very low health just before leaving my world the last time. 10:55 ANAND The reason for the assertion fail might be due to my HP being clamped at 0 when the code expects it to go as far as (curr_HP - damage)? 10:55 ANAND I'm just speculating, but I suspect this might be the reason 10:55 ANAND nephele: Related #9962 10:55 ShadowBot ANAND: Error: That URL raised 10:56 ANAND Oh sorry 10:56 ANAND #9662 :P 10:56 ShadowBot https://github.com/minetest/minetest/issues/9662 -- Low-level HUD element to allow mods to "draw" elements on the screen 10:56 ANAND Does that satisfy your use-case? 10:59 nephele No, not really, the current api is giving me massive trouble the way it is designed now, especially with no way to get apparent sizes for textures, no way to get aparent sizes for text, and no way to have a reliable coordinate system (since it will scale with the scale factor) 11:00 nephele #also having to keep track of every element and unregistering it or deduplicating it with regards to the next state is a bit troublesome, but not a massive problem 11:02 ANAND Ohh, are you requesting for dynamic elements that allow modifications every tick? :O 11:03 ANAND It's a neat idea, really. Performance would be my concern as well - there would be a lot of Lua<-->C++ overhead as sfan noted. 11:03 nephele No, i don't need every tick neccesarily, although ideally a low-latency input would redraw as soon as the state changes, my main usecase is building a ui that does change based on my view of the state, it doesn't matter much whether i re-make the hud elements that will be drawn or draw them myselves that much 11:04 ANAND Mhmm 11:05 nephele I suppose the current system would work fine with some more primitives, exposing the scale factor /instead/ of scaling the coordinate system (and have a whenchanged callback for that), and a way to get the appart size of text which was laid out, aswell as getting the size of textures ahead of time (and laid out sizes) 11:06 nephele apparent* 11:11 sfan5 merging #9893 in 10 minutes 11:11 ShadowBot https://github.com/minetest/minetest/issues/9893 -- Cleanup of particle & particlespawner structures and code by sfan5 11:19 nephele ANAND, I did check that specific mr now, i really hate the api there, makes it much more complicated to use it, i would just have different primitives for square and triangle instead of shoving all into one element 11:21 nephele (also for current elements it's really wierd that one needs to use "text" or something to specify the texture, does that have any reason?) 11:24 nephele err, issue actually... not mr :F 12:19 Krock will push https://krock-works.uk.to/u/patches/0001-MacOS-Fix-environ-not-being-found.patch in 10 minutes 12:19 Krock #9904 afterwards, with removed baces 12:19 ShadowBot https://github.com/minetest/minetest/issues/9904 -- Check for valid base64 before decoding by Lejo1 12:22 sfan5 Krock: if you name it apple_environ anyway why not inline the *_NSGetEnviron()? 12:22 Krock yes, right. 12:30 Krock pushing (2) 15:20 Krock interestingly, std::vector + resize + memcpy is faster than std::stringstream << value 15:20 Krock vector: 100ms to append "test" 10 million times 15:20 Krock stringstream: 200ms 15:23 Krock though stringstream catches up fast in time with larger amounts of data 15:34 Krock also push_back vs emplace_back makes no difference at all 15:34 Krock at least for primitives 15:42 ANAND emplace_back doesn't make much sense when used with primitives 15:45 ANAND I'm sure it'd only be negligibly faster than push_back for small objects. 15:50 ANAND But std::vector being faster than stringstream comes as a surprise to me. That's very interesting. 18:17 appguru Tryin' to get PRs merged, see #minetest 18:18 sfan5 then why not do that here? 18:18 appguru Alright, I'll start with asking for #9903 to be merged. 18:18 ShadowBot https://github.com/minetest/minetest/issues/9903 -- Properly implement exposing the zoom key by appgurueu 18:19 appguru I think I have resolved any possible attribution issues as I merged raistlin's branch. 18:21 Lone_Wolf I get this crash in Nodecore with 5.3 it's happening to everyone else too https://pastebin.com/1udEPyXf 18:30 appguru it does "scene::IMesh *mesh = m_meshnode->getMesh();" which is later dereferenced for upright sprites, could be a null pointer 18:31 appguru "scene::IMeshBuffer *buf = mesh->getMeshBuffer(i);" could also be a null pointer 18:33 appguru Lone_Wolf: what are you doing to cause the error? 18:40 Lone_Wolf Walking around in an NC world 18:40 appguru how long? 18:42 sfan5 rubenwardy: random UI idea: in the bottombar with the games add a button that directly takes you to contentdb to browse and install games 18:43 rubenwardy that's in my main menu redesigns, effectively 18:44 sfan5 I expected a response like this 18:44 sfan5 but why not add this now until the redesign is ready in a few versions? 18:45 appguru I might get my mainmenu redesign done until 5.3 18:46 sfan5 the intended release date for 5.3 isn't in 6 months if you are expecting that 18:52 rubenwardy users don't understand the game bar 18:52 rubenwardy it's been shown a lot of times 18:52 rubenwardy it could be added, wouldn't be that hard, the whole thing needs to be improved though 18:53 rubenwardy although actually 18:53 rubenwardy could be an idea to keep it 18:54 sfan5 the idea is just to make CDB even more obvious with quick, small change 18:54 appguru How about making the content tab the first one? 18:54 rubenwardy yeah, adding a plus to such a list makes it obvious how to add stuff 18:55 appguru but the plus combined with the scroll buttons will lead to confusion 18:55 appguru unless you plan on using a scroll container 18:57 rubenwardy I'll make it so the last element is a plus, and it scrolls with the content 19:04 rubenwardy it probably should also be changed to a scroll container 19:36 Krock #9914 19:36 ShadowBot https://github.com/minetest/minetest/issues/9914 -- POC: Introduce SerializeStream helper class by SmallJoker 19:37 Krock scroll container end