Time  Nick        Message
13:46 Krock       will merge #9503 in 15 minutes
13:46 ShadowBot   https://github.com/minetest/minetest/issues/9503 -- Fix some chatcommands not returning a value by Wuzzy2
14:01 Krock       merging...
15:12 rubenwardy  sfan5: pong
15:16 sfan5       rubenwardy: I compiled Minetest for the Raspberry Pi for use with the open GLES driver that apparently works on newer pis
15:16 sfan5       so if you have some time, It'd be nice if you could test it on yours https://kitsunemimi.pw/tmp/minetest-5.2.0-dev-rpi%2B.tar.gz
15:17 rubenwardy  ok cool
15:17 rubenwardy  I'm in a cafe currently but may be able to later
15:17 rubenwardy  although, I have no keyboard for the pi so hope I can still ssh in....
15:18 rubenwardy  since my keyboard broke, I've been using cli macro tools to move the mouse and type lol
15:18 rubenwardy  when I needed to rarely do that on the pi
15:26 rubenwardy  lol (irrlicht source code) https://rwdy.uk/Bmy65.png
15:29 Krock       only the GUI needs to send GUI events
15:31 rubenwardy  trying to work out how to make events pass through non-interactable elements
15:32 Krock       either you set the element temporary to invisible, or you return false on the callback
15:34 rubenwardy  the latter doesn't work
15:34 rubenwardy  because it only sends mouse clicks to the focused event
15:34 rubenwardy  and returning false doesn't prevent it from being focused
15:34 Krock       then don't allow it to be focussed
15:34 rubenwardy  I'm trying to work out hwo
15:36 Krock       or set the rectangle for the drawing area to 0 pixels area and let it overlap (no clip=
15:36 rubenwardy  I can override getElementFromPoint to return null maybe
15:40 rubenwardy  damnit, that function is virtual in Irrlicht dev and non-virtual in Irrlicht stable
15:40 rubenwardy  might have to be the rectangle hack
15:43 rubenwardy  Irrlicht really should have followed release early; release often
15:43 rubenwardy  7 years since the last release is a joke
15:43 rubenwardy  last release that wasn't a patch/backport release, that is
15:45 Krock       maybe it would help to infiltrate/help Irrlicht by at least reviewing the PRs, and convince to an update?
15:45 Krock       keyboard input is broken since ever
15:48 rubenwardy  I've found a disgusting hack
15:49 rubenwardy  it works if you do    setVisible(false);    and then ignore isVisible in the element's draw();
15:49 Krock       yay
15:50 DS-minetest nice idea, but you would still have to use an own class, ie. you can't use irrlicht's image element
15:51 rubenwardy  yes, this works with ItemImage though
15:51 rubenwardy  and I will create an own class for image for this
15:53 rubenwardy  is there a better way to do this?
15:53 rubenwardy  I wonder if we could implement getElementFromPoint and event logic ourself in someway
15:53 rubenwardy  eugh
15:58 DS-minetest there should be a function in the guielement class that returns whether it is at a given point. eg. a circle isn't in the edges of its rect
15:59 DS-minetest and an image in the weird minetest formspec is nowhere
15:59 DS-minetest getElementFromPoint would then not only check the rect but also call this function
16:00 DS-minetest (or just call the function and the function checks whether the point is in the rect)
16:00 rubenwardy  getElementFromPoint is called on each element
16:00 rubenwardy  however, it's not virtual
16:00 rubenwardy  so I can't change it
16:01 rubenwardy  if we pulled Irrlicht GUI into our source code, we could do this
16:01 rubenwardy  it's virtual in latest dev of Irrlicht
16:01 DS-minetest the irrlich gui source should be included into minetest at some point anyway
16:01 rubenwardy  pulling irrlicht into our source code isn't a good idea whilst in a feature freeze
16:01 DS-minetest +
16:02 DS-minetest feature freeze => fix remaining bugs with hacks
16:17 rubenwardy  #9516
16:17 ShadowBot   https://github.com/minetest/minetest/issues/9516 -- Fix non-interactable elements blocking input by rubenwardy
16:20 DS-minetest you could add to that code comments that the draw function ignores this and the element is drawn anyway
16:21 DS-minetest (the codecomments to the calls to setVisible(false))
16:21 DS-minetest oh, and background elements need the not drawing when not isVisible btw.
16:22 rubenwardy  I think I've done that?
16:23 rubenwardy  some elements haven't been checking isVisisble in draw() because it was assumed that it wouldn't be called I guess
16:26 DS-minetest backgrounds use setVisible, eg. here: https://github.com/minetest/minetest/blob/1441281809d9961ff7498df86601db151956b030/src/gui/guiFormSpecMenu.cpp#L1036
16:27 DS-minetest and then here: https://github.com/minetest/minetest/blob/1441281809d9961ff7498df86601db151956b030/src/gui/guiFormSpecMenu.cpp#L3234-L3241
16:28 DS-minetest but this could be fixed by first adding an element to fsmenu that holds all backgrounds as its children and using that element as parent
16:28 rubenwardy  it's also needed by static text
16:29 rubenwardy  haven't we always used a GUI element for labels?
16:29 rubenwardy  so that isn't necessarily a regression
16:32 rubenwardy  alternative: the OnEVent of the element could do setVisible(false) then call getElementFromPoint again
16:32 rubenwardy  slower
16:32 rubenwardy  but slightly less hacky  - is visible will still be usable
16:33 DS-minetest I've used that alternative in invlists
16:34 DS-minetest you are right with the labels, they always used a guielement
16:39 Krock       copying yet another source file from irrlicht hurts my soul
16:39 Krock       it's getting increasingly worse
16:44 rubenwardy  we will need to import all of the GUI code into our code base, and put into a guix namespace
16:44 rubenwardy  then remove all this crap
16:44 rubenwardy  although, I am wonder whether we should give up on fixing formspecs and do another GUI implementation alongside
16:44 rubenwardy  it'll be easier to stop regressions
16:51 DS-minetest here's the change that would have to be done to not break backgrounds when ignoring isVisible in drawing: https://github.com/Desour/minetest/commit/6580abc54c1b8404ff9b3c351b5dd8f3f1884e5d
16:51 rubenwardy  I don't think we actually need to apply this hack to backgrounds
16:51 rubenwardy  because they've always been the backmost things
16:52 DS-minetest yes
16:52 rubenwardy  huh
16:52 rubenwardy  the inventory list hack doesn't work
16:52 rubenwardy  I guess because OnEvent isn't used to change focus
16:52 DS-minetest it doesn't work with hovering
16:52 rubenwardy  perhaps the same hack should be done for focus change gui events
16:54 rubenwardy  how should I fix tooltips?
16:54 rubenwardy  It's not strictly a regression, labels have always stolen hover
16:54 rubenwardy  but it would be nice to fix
16:56 DS-minetest hm, one could store all the elements that should only be visible when drawing in a vector and set them only visible before drawing and set them invisbile directly after
16:57 DS-minetest then one wouldn't have to copy the source of image and label gui element
16:58 rubenwardy  oh that might work
16:59 rubenwardy  how does gui environment get events? Would be nice to do that hack around the OnEvent, rather than around the draw
16:59 DS-minetest idk
17:03 rubenwardy  cutealien claims there's only a few months before 1.9 is released
17:03 rubenwardy  [x]   doubt
17:42 Krock       #9461 #9515 #9514 need a second review
17:42 ShadowBot   https://github.com/minetest/minetest/issues/9461 -- Some minor Lua API documentation clarifications by Wuzzy2
17:42 ShadowBot   https://github.com/minetest/minetest/issues/9515 -- Transform texture UVs with provided tex. matrix (#9481) by dcbrwn
17:42 ShadowBot   https://github.com/minetest/minetest/issues/9514 -- Fix new nodes not being light up by dcbrwn
17:43 Krock       #9411 needs re-reviewing after the recent changes ( rubenwardy ?)
17:43 ShadowBot   https://github.com/minetest/minetest/issues/9411 -- Add starting frame to `animated_image` by v-rob
18:20 Wuzzy       Is it OK to create a node with drawtype="flowingliquid" if the node is not a liquid itself (liquid_type = "none")?
18:20 Wuzzy       I know there's an assert that forbids this, but I wonder if this assert is justified.
19:33 Krock       *shrug* Might need testing to ensure it really doesn't cause side-effects
19:34 Krock       I'd imagine the liquid flow would simply let it vanish