Time Nick Message 04:12 ANAND Hello. More opinions needed on #7027 04:12 ShadowBot https://github.com/minetest/minetest/issues/7027 -- Add on_drop callback register. 04:13 ANAND ^ Feature request for a minetest.register_on_drop global callback 07:09 tomraceror hello 09:48 p_gimeno ANAND: does it refer to the kind of drop that happens when you press Q? 12:32 ANAND p_gimeno: Yes 12:33 ANAND But the global variant instead of the one in the itemdef 12:58 p_gimeno ok, just wanted to be sure I understood 16:43 sfan5 merging #8516 in 5 mins 16:43 ShadowBot https://github.com/minetest/minetest/issues/8516 -- Define operators == and != for ItemStack by ClobberXD 17:09 rdococ Is there a way to change how gl_Color is interpolated? 17:45 p_gimeno hm, DS-Minetest isn't on IRC, right? The Matrix-Euler conversion functions in #8515 should probably be using the core, or at least the same method as the core. 17:45 ShadowBot https://github.com/minetest/minetest/issues/8515 -- Add matrix helpers to builtin by DS-Minetest 17:48 p_gimeno The method is here: https://github.com/minetest/minetest/pull/8019#issuecomment-454976672 17:52 p_gimeno in the core, they are functions setPitchYawRollRad and getPitchYawRollRad defined in numeric.cpp 18:03 p_gimeno I also wonder why to deliberately drop precision when returning rotation matrices 18:19 p_gimeno if the intention is to handle sin(pi/2) and cos(pi), special casing them would be better, rather than reducing the precision under that of 32-bit float 19:14 paramat input needed on what to do in #8249 , see the discussion 19:14 ShadowBot https://github.com/minetest/minetest/issues/8249 -- Make autoforward simulate keys by DS-Minetest 19:22 paramat wow maybe we should have a 'formspec' issue label, it will get used a lot =/ 22:43 p_gimeno DS-Minetest> Minetest has this: [YXZ euler convention] <---- No it doesn't. Irrlicht uses XYZ Eulers; Minetest uses ZXY Eulers. set_rotation is using setPitchYawRollRad. The function getPitchYawRoll[Rad] is not used but it's tested in the unit tests. I may have screwed up the signs of the angles, because left-handed coordinate systems confuse me and Minetest conventions even more, but not the rotation order for sure. In fact your 22:43 p_gimeno from_yaw_pitch_roll applies the rotations in ZXY order. 22:45 p_gimeno The sign convention for yaw was made to be backwards-compatible, and the other two followed the same as yaw. 23:03 rdococ can't we just forgo all this confusion with angles and conventions and use quaternions? 23:03 p_gimeno if only... 23:04 rdococ you could try to convert between the two in lua I suppose 23:04 rdococ there needs to be an intuitive way to explain quaternions 23:05 p_gimeno re quaternions: yes you could, in fact stu says he's doing that: https://github.com/minetest/minetest/pull/8019#issuecomment-457778933 23:07 p_gimeno explaining quaternions is easy: they are like zipped rotation matrices, and don't mess with the components unless you know what you're doing 23:09 rdococ I think minetest should switch to using a "rotation" table with functions for creating and manipulating rotations in intuitive ways, but then just representing them as quaternions internally. 23:11 p_gimeno Irrlicht doesn't directly support representing them as quaternions internally, though, so that's an obstacle 23:11 rdococ the heck? 23:12 rdococ why would a game engine use anything other than quaternions internally? 23:12 rdococ especially euler angles... 23:12 p_gimeno there's been loads of discussion over that issue :) 23:12 p_gimeno in both the Irrlicht and the Minetest sides 23:17 rdococ why? to me it's a no-brainer, what is the argument against switching? 23:20 p_gimeno http://irrlicht.sourceforge.net/forum//viewtopic.php?p=293347#p293347 http://irrlicht.sourceforge.net/forum//viewtopic.php?f=7&t=52414&p=304417#p304417 - no answer from Irrlicht, maybe no dev interest. 23:21 p_gimeno in Minetest, extra code and maintenance burden because it delegates these things to Irrlicht 23:28 p_gimeno I don't think there's anything against adding quaternion functions to Lua, though. A matrix class is in the works (the object of the discussion above).): 23:29 rdococ I found some thin wrappers around existing libraries for quaternions and matrices 23:34 p_gimeno there's CPML, but it uses column major order instead of row major, and it screwed up matrix multiplication order (probably for that reason)