Time  Nick          Message
07:39 ANAND         #8271 is ready for review
07:39 ShadowBot     https://github.com/minetest/minetest/issues/8271 -- Fix extra bubbles being displayed for high breath_max values by ClobberXD
07:40 ANAND         The PR title needs changing, as it's somewhat inaccurate now
07:42 Krock         you can edit it too IIRC
07:43 Krock         now players have s little bit less breath when diving
07:43 Krock         nice print calls btw
07:45 Krock         and the HUD will no longer be removed when the player becomes immortal or the HUD flags change
07:54 ANAND         ouch. Removed the print calls
08:04 ANAND         Fixed. Testing...
08:06 Krock         I'd just add a stupid math.min(2 * scaleToDefault(..), 20) in there
08:08 Krock         til running too fast with an entity attached to your model's side eventually collides
08:11 ANAND         > til running too fast with an entity attached to your model's side eventually collides
08:11 ANAND         Sorry, I didn't catch that
08:12 ANAND         Why math.min? Isn't this a better, more robust implementation?
08:14 ANAND         Btw, 2 * math.min(scaleToDefault(), PLAYER_MAX_BREATH_DEFAULT) is probably what you meant, I guess?
08:15 ANAND         And scaleToDefault() already tunes the value according to PLAYER_MAX_BREATH_DEFAULT
08:21 Krock         will push https://krock-works.uk.to/u/patches/0001-Builtin-Forward-old-return-values.patch in 15 minutes
08:21 Krock         ANAND: yes.. hiding it after some delay is probably the best solution
08:22 Krock         not sure if the C++ code also relies on this additional breath
08:42 Krock         pushing
09:44 ANAND         Krock: Just finished testing (was busy for a bit), #8271 works. But the changes in statbars.lua seem to be somewhat clumsy to me, so let me know if they need to be refactored.
09:44 ShadowBot     https://github.com/minetest/minetest/issues/8271 -- Fix breath statbar scaling; defer breath_bar hiding by one second by ClobberXD
09:59 rubenwardy    I've thought of a way around the app store limitation - by linking Minetest as a library to a MIT licensed iOS app
10:00 rubenwardy    the reason that (L)GPL apps can't be put on the app store is because the license says that you're not allowed to reverse engineer iOS apps on the app store
10:00 rubenwardy    which is in conflict with the (L)GPL
10:09 sfan5         but if that no-reverse engineering covers the entire distributed app, how does it help to have Minetest as a library?
10:10 rubenwardy    that's something I need to check
10:11 rubenwardy    I have a feeling that is in the GPL to make it so people can replace libraries, but maybe it doesn't matter if the app is MIT
10:11 rubenwardy    Apple sucks
11:08 sfan5         rubenwardy: now the button is vertically too big, but it's just personal preference so feel free to ignore my suggestion
11:08 rubenwardy    I can't change the vertical height without redesigning the whole thing to use real_coordinates
11:08 rubenwardy    which I will do in my mainmenu improvement PR that I need to finish
11:08 sfan5         formspecs still suck
11:08 rubenwardy    yeah
11:09 rubenwardy    we're bolting more onto a broken solution
11:09 sfan5         yes
11:09 sfan5         it should be replaced entirely but that's a big task
11:09 rubenwardy    might be worth discussing how much we should be bolting on to it
11:10 rubenwardy    I suggest that we bolt enough on to do the things we really need right now, but not much more
11:10 rubenwardy    ie: it's probably not worth adding support for custom elements or canvases in CSM to formspecs
11:11 rubenwardy    I also suggest not doing too much with formspec styling, as importing IGUIElements and adding styling options is a never ending task
11:11 rubenwardy    Krock: does your approval to #7895 still hold
11:11 ShadowBot     https://github.com/minetest/minetest/issues/7895 -- Add luacheck to check builtin by rubenwardy
11:11 rubenwardy    I need to rebase it
11:12 rubenwardy    I'm also not sure on the technique for that, it's fairly risky given it's done in commits per secion
11:15 Krock         the approval is still valid, but will have a quick look after rebasing
11:15 Krock         what's "risks"?
11:16 Krock         *risky. Is it because you're changing many different files?
11:16 rubenwardy    it's a massive PR which could introduce regressions
11:16 rubenwardy    I wish we had proper unit and integration testing
11:17 Krock         mostly trivial code style changes
11:17 rubenwardy    yeah
11:20 rubenwardy    huh, using posix_spawnp means that xdg-open can't find firefox. I guess somethings wrong with the path?
11:27 sfan5         are you passing envp as NULL?
11:30 rubenwardy    yeah, that'll be why
11:30 rubenwardy    I can't see how to get the environment, it results in a linking error with the tutorial I find
11:31 rubenwardy    ie:        extern char **environ;
11:35 sfan5         that linking error is?
11:36 rubenwardy    undefined reference to `porting::environ'
11:36 rubenwardy    ooohhh
11:36 rubenwardy    I'm an idiot
11:37 rubenwardy    wait, it's also already defined in an included header
11:40 rubenwardy    ok, two things remaining - the call hard codes arg0 to be "/usr/bin/xdg-open",  whereas the program name is just "xdg-open" because it searches it in the path
11:40 rubenwardy    this probably isn't that much of a problem
11:40 rubenwardy    but is unclean
11:41 rubenwardy    second, the regex needs to be fixed to allow localhost and 127.0.01
11:41 rubenwardy    and other IPs
11:41 rubenwardy    and the HTTP standard generally
11:46 Krock         is #8736 supposed to gather performance results on different platforms?
11:46 ShadowBot     https://github.com/minetest/minetest/issues/8736 -- Define and document minimum and recommented requirements
11:47 Krock         if so, then the forums would be way better choice
11:47 rubenwardy    yeah, that would be a good idea
11:47 rubenwardy    what's the best way to collect rendering stats across loads of devices?
11:47 rubenwardy    we need a benchmark suite ha
11:47 Krock         F6 debug stats
11:48 Krock         on low end systems no text should be visible in-game for fast rendering
11:52 sfan5         rubenwardy: you don't need the absolute path in arg0
11:53 rubenwardy    ok
11:53 rubenwardy    cool
11:53 sfan5         also re. regex: just check the string starts with "http://" or "https://" and only contains ascii characters
11:53 sfan5         blacklist ' " $ \ or whatever too if you feel like it
11:54 rubenwardy    and spaces?
11:54 sfan5         i guess
11:54 Krock         %20 can be used if they really want spaces
11:56 rubenwardy    yeah, or +
11:57 rubenwardy    ^https?:\/\/([\w\.-\/\?=\%\+]+)$
11:57 rubenwardy    hmm
11:58 rubenwardy    missing #
11:58 sfan5         i'd also prefer not using a heavy regex engine for a simple validation like this but this is personal preference and may be premature optimization
11:58 rubenwardy    I suppose I could do substring matching
11:58 rubenwardy    but yeah, this isn't a regular thing
11:58 rubenwardy    like, it's on a link click
11:59 rubenwardy    if this becomes regular, then there's something wron
11:59 sfan5         well allowing this to be used from formspecs/csm (with restrictions) would be fine by me
12:00 rubenwardy    yeah
12:00 rubenwardy    that's the eventual plan
12:00 rubenwardy    I'd like to add a dialog like the one steam has, where it shows the URL
12:03 rubenwardy    it would be useful to parse the domain in that case, so you could have   `secure.trusted_domains = minetest.net`
12:03 rubenwardy    or "www.minetest.net"
12:03 rubenwardy    but not important
12:03 rubenwardy    trusted meaning it skips the dialog
12:05 sfan5         just allow the mainmenu to do these things
12:05 rubenwardy    yeah, currently that's the case
12:05 rubenwardy    allowing mods would be later
13:06 Krock         also create a demo map plus predefined setting files to ensure everybody's testing the same thing
13:07 Krock         it makes a huge difference whether to draw an entire city or forest compared to plains
14:35 DS-minetest   would someone like to give #8640 a second approval? :)
14:35 ShadowBot     https://github.com/minetest/minetest/issues/8640 -- Trigger on_place in many situations even if prediction failed by DS-Minetest
14:49 Krock         updated #8282. Only ESC will clear the keys. Re-tested & works as wanted.
14:49 ShadowBot     https://github.com/minetest/minetest/issues/8282 -- Key settings: Clear with escape by SmallJoker
14:55 DS-minetest   does the key already in use work correctly with that PR?
14:55 DS-minetest   (currently it only shows up once)
14:55 Krock         it's kept unmodified
14:55 DS-minetest   ok, then it's good
15:01 p_gimeno      DS-minetest: what are the chances that 8640 breaks mods that expect the old behaviour?
15:06 DS-minetest   good question
15:08 DS-minetest   there might be problems if a mod sets nodes in pt.above in on_place without check
15:09 DS-minetest   but item_place node checks for this: https://github.com/minetest/minetest/blob/47492386ece5b016a7a0ed06d6a44cc7d60adb55/builtin/game/item.lua#L284-L289
15:14 p_gimeno      ok, sounds sensible
18:30 rubenwardy    too many PRs!
18:31 sfan5         do mine first
18:36 rubenwardy    huh, weird how X11 was found twice
18:38 * twoelk      reserves first february2020 weekend for mt related travel option
18:40 Krock         soon one PR more if I succeed with the incremental inventory sending
18:41 rubenwardy    :D
18:50 rubenwardy    oh jesus
18:50 rubenwardy    I was testing
18:50 rubenwardy    but then got compile errors when enabling gles
18:50 rubenwardy    because you need to recompile Irrlicht
18:54 rubenwardy    "I hate this section."      lol, same
18:56 * DS-minetest too
18:57 DS-minetest   even worse are people that want you to fill the section properly ;)
19:04 Krock         I believe for my laptop I also had to modify the irrlicht branch a bit
19:04 Krock         very hacky but it worked.. somehow.
19:32 rubenwardy    from a code inspection, that PR looks good
19:32 rubenwardy    and I tested ENABLE_GLES=0
19:32 rubenwardy    but I don't have the energy to deal with Irrlicht
19:32 rubenwardy    maybe I'll fetch my pi and see if I can do it on that
19:48 Krock         #8742
19:48 ShadowBot     https://github.com/minetest/minetest/issues/8742 -- Inventory: Send dirty lists where appropriate by SmallJoker
19:48 Krock         onemore
19:55 Krock         > + src/unittest/test_world/world.mt
19:55 Krock         now I get why this PR got so many new lines
20:04 Krock         rubenwardy: why are the checkboxes in https://github.com/minetest/minetest/pull/8383#issuecomment-511519170 not ticked?
20:04 Krock         this makes the PR feel incomplete
20:16 Krock         code looks good. tested an old version of that PR a few weeks ago, but will retest tomorrow (?)
22:54 GreenDimond   I went and modified GenericCAO::setChildrenVisible() to always do obj->setVisible(true), but for some reason that I cant figure out the attachment still doesnt draw
22:55 GreenDimond   Am I missing another check somewhere that overrides the visibility of the object according to the camera mode?