Minetest logo

IRC log for #minetest-dev, 2015-01-07

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:00 kilbith no
00:00 kilbith the text exceed way too much their emplacements
00:00 kilbith this is not harmonious at all
00:02 sapier screenshots please ;-)
00:02 VanessaE kilbith: did you pull to latest git from less than about 2 hours ago?
00:02 kilbith haven't compiled today, just took the .deb from 1h ago
00:03 VanessaE idk if that's new enough, probably not
00:04 kilbith well, the text excess is just for particular cases, but i still think that the scale text/gui elements ain't harmonious
00:05 acerspyro joined #minetest-dev
00:05 VanessaE kilbith: pull, compile, set in minetest.conf:  gui_scaling = 0.85  and comment-out any font size settings
00:05 VanessaE see what that does.
00:05 kilbith if you choose a readable enough font size, you'll have a gross gui elements
00:05 VanessaE then play around with the gui scale after that initial test.
00:06 kilbith if you choose a plesant gui elements scaling, the text become hardly readable
00:06 kilbith in either cases, that aint good
00:07 VanessaE sapier: sounds like another vote for setting the default fonts to fixed relative scale...
00:08 sapier I haven't seen a screenshot and I'm not gonna consider "i don't like the change" comments any longer
00:09 sapier I can't do any tuning if ppl don't tell me WHAT is wrong ... this will always be a change no matter what we do
00:11 sapier don't get me wrong I'm eager to fix it but I can't do this without your help!
00:11 VanessaE you might want to wait for him to sign on again :)
00:11 VanessaE I've already given all the input I can, everything further will just be more "I don't like it" :P
00:11 sapier I don't think he's gonna provide the screenshot I already asked twice
00:12 VanessaE he might, he's pretty good about that normally.
00:12 acerspyro lol
00:12 ShadowNinja Wow, the Android build is a lot better that it used to be.
00:12 sapier well it's quite natural we need to tune this to get a acceptable state
00:12 ShadowNinja It actually seems releasable.
00:12 sapier ShadowNinja: we could have had this way before the changes are quite minor
00:13 ShadowNinja I can use a 100ish viewing range.
00:13 sapier well as of performance I didn't change anything
00:13 n4x joined #minetest-dev
00:13 sapier if there's improvement it's withing minetest core
00:14 ShadowNinja sapier: Yes, probably Zeno's improvements.
00:14 sapier what did he change?
00:14 VanessaE what *didn't* he change would gather a shorter response ;)
00:15 sapier well I've only seen the things he broke by now ;-)
00:16 sapier I don't have a significant fps increase on my device ... maybe your memory is biased ShadowNinja ;-)
00:16 VanessaE sapier: zeno's changes have increased performance at VE-Survival spawn by at least 50%.
00:17 sapier so now you have 7fps?
00:17 VanessaE what was 20 or less at 35m is now ~35 fps at ~50 meters
00:17 VanessaE variable, of course, but that seems to be the average
00:17 acerspyro VanessaE: when did his changes happen?
00:17 VanessaE in other areas, it now hits the upper limit of 30 fps/240m but rarely
00:17 sapier did you try on your tablet? I remember you had about 5 fps there when we last tried the android client
00:18 VanessaE acerspyro: over the past few months
00:18 acerspyro k
00:18 VanessaE sapier: oh G*d no, not that poor little thing :)
00:18 VanessaE I was talking about desktop performance :)
00:18 sapier well it's a good benchmark
00:18 acerspyro Why would god be a swear?
00:18 VanessaE acerspyro: offtopic.
00:18 VanessaE sapier: actuall my tab can't get to VE-Survival but on Vanilla it was like 3 fps at best.
00:19 acerspyro Oh you people, such a nonsensical mystery sometimes.
00:19 ShadowNinja sapier: No, it's actually unable with full viewing range now, as long as I don't explore a lot and look back.
00:20 sapier well no matter why, it's good to be better now
00:20 ShadowNinja I noticed that the enable/disable buttons are very touchy though.  One tap and turn then off and back on before you lift your finger.
00:21 sapier if they did use same code as control keys key repeat is about 0.2 s
00:22 sapier which is same as on pc but there you have some mechanical delay till a button is pressed ;-)
00:23 ShadowNinja sapier: Sound volume menu seems to freeze it.
00:24 sapier yes that one isn't switched to formspec by now ... who did reenable it on android? thought it's hidden?
00:26 sapier maybe we should write an issue for it
00:32 n4x joined #minetest-dev
00:32 sapier left #minetest-dev
00:57 shadowzone joined #minetest-dev
01:45 y joined #minetest-dev
01:54 younishd joined #minetest-dev
01:55 younishd joined #minetest-dev
02:07 n4x joined #minetest-dev
02:50 kaeza hmmmm, https://gist.github.com/kaeza/e31abc665441272ec0d8
02:52 hmmmm i'm not gonna lie that's an odd error
02:52 hmmmm could it be that you somehow paired a newer set of builtin scripts with an outdated engine?
02:54 kaeza git pull says up to date
02:54 kaeza oh shit
02:54 kaeza note to self: compile after pulling
02:54 VanessaE lol
02:55 hmmmm yeah that'd do it
02:55 kaeza sorry hmmmm
02:55 hmmmm np
03:06 hmmmm hey guys, in the whole formspec crazyness, have any of you noticed if static GUIs such as the volume changer has text that gets cut off with a higher DPI setting?
03:45 compunerd joined #minetest-dev
04:47 hmmmm okay, I'm boggled
04:47 VanessaE that can't be good :P
04:47 hmmmm how do formspec tabs perform the actual tab switching
04:47 hmmmm I literally am unable to figure this out
04:47 VanessaE they switch out the whole formspec page
04:47 hmmmm where.
04:48 hmmmm i don't see it
04:48 VanessaE that I don't know, but that's how other formspecs do it.  a control button sets a whole new formspec from scratch.,
04:48 VanessaE -m
04:48 VanessaE +m-,
04:51 hmmmm i see where get_formspec() is being called in ui.update()
04:51 hmmmm but i don't see where the tab being switched to gets inserted into ui.childlist
04:52 hmmmm VanessaE, back in 0.4.5 and whatever do you remember the bottom of the transparent main menu increasing opacity in a gradient?
04:53 hmmmm maybe this is an irrlicht 1.8 thing
04:53 VanessaE not in the main menu, no.  but I seem to remember there being a similar effect in some in-game formspecs
04:53 hmmmm hrm
04:53 hmmmm well whatever
04:53 hmmmm I am not wasting time on such BS
04:53 VanessaE (which I mimic in dreambuilder's formspec images)
04:54 hmmmm i already wasted enough
04:55 VanessaE so in that case you'll just rewrite the whole fucking thing ;)
05:09 paramat joined #minetest-dev
05:28 chchjesus joined #minetest-dev
05:51 hmmmm naahhh
05:51 hmmmm so i'm writing some dumb gui code
05:51 hmmmm this is so stupidly easy, it should take me like an hour
05:51 hmmmm i'm going to time myself.  if i can't finish this in 1 hour, i'm messed in the head
06:26 cib0 joined #minetest-dev
06:48 Hunterz joined #minetest-dev
06:58 Player_2 joined #minetest-dev
07:07 gregorycu joined #minetest-dev
07:08 AnotherBrick joined #minetest-dev
07:16 CraigyDavi joined #minetest-dev
07:23 paramat celeron55 and others, this needs to be reverted https://github.com/minetest/minetest/commi​t/1c9f05d792562374046e74ad3eb75988d529b15c step smoothing is now extremely mushy and sluggish, it now takes a whole second for the camera to rise when climbing a step, the camera sinks significantly down towards the stairs. i tested the old and new values on a long 1:1 staircase
07:25 paramat there was nothing wrong with the old value, smoothing has to be subtle while still feeling agile and responsive
07:28 paramat we really need an active audio visual manager with very similar tastes to c55
07:30 Taoki[mobile] joined #minetest-dev
07:32 paramat other stupid changes to Minetest include: fall bobbing when the player has no knees, and camera-tilt on damage which is an unnecessary copy of Minecraft
07:36 paramat i would trust hmmmm to be the new AV manager/decision maker
07:47 paramat here's another example of mindless Minecraft copying https://github.com/minetest/minetest/pull/1953 which i am trying to sort out
08:27 crazyR joined #minetest-dev
08:33 crazyR joined #minetest-dev
08:46 paramat anyway nevermind those side issues, i should focus on the main issue which is the step smoothing, i'll make a pull request later for the original value O/
08:46 paramat left #minetest-dev
08:48 jin_xi joined #minetest-dev
09:02 kilbith joined #minetest-dev
09:02 gregorycu joined #minetest-dev
09:32 FR^2 joined #minetest-dev
09:33 rubenwardy joined #minetest-dev
09:34 rubenwardy We need more moderators on the forum - there have been lots of spam left for ages recently. Bots are now replying to topics with spam.
09:34 rubenwardy Or we need captchas
09:37 ImQ009 joined #minetest-dev
10:02 Amaz joined #minetest-dev
10:11 celeron55 with only two replies this is already starting to prove people have very different requirements for what is publishable quality: https://forum.minetest.net/v​iewtopic.php?f=3&t=10924
10:11 celeron55 not surprising, but doesn't really help
10:32 iqualfragile joined #minetest-dev
10:56 gregorycu I'll reply then
10:57 gregorycu Ohh Android, cbf
10:58 crazyR_ joined #minetest-dev
11:18 sapier joined #minetest-dev
11:18 sapier https://github.com/minetest/minetest/pull/2044 merging this one in a few minutes (memory leak fix)
11:30 kilbith sapier: https://github.com/minetest/minetest/issues/2077
11:31 kilbith sorry for yesterday, it was very late here and i was hurry to lay down in bed
11:31 sapier no problem I'm glad someone finaly provides the information I need
11:32 kilbith it's me in that issue ^
11:34 sapier is this with or without freetype?
11:34 kilbith with
11:34 kilbith as i said in the issue
11:34 kilbith "From the up-to-date Git HEAD (compiled one hour ago with Freetype)"
11:35 sapier except from the mod config tab what exactly is the problem?
11:37 sapier btw they're not adapted to android at all as It didn't even work there correct till yesterday ;-)
11:37 kilbith the mod config is just a minor example, the main issue is the oversizing and the adequation GUI / font scaling
11:38 sapier ok so in short your complain is "fonts are to big" ?
11:38 kilbith if you set something good for GUI, you have a bad result for the fonts, and inversely
11:39 kilbith not just "too big"
11:39 kilbith read better the issue
11:39 sapier but?
11:39 sapier I did that's why I'm asking
11:39 sapier fonts keep exactly SAME ratio to overal form size
11:40 sapier (not counting rounding errors)
11:40 kilbith 1.0 -> extravagant // 0.85 -> GUI is OK but fonts large & skinny // 0.75 -> fonts are OK but GUI too shy
11:41 sapier so you suggest adding a variable scale factor scaling fonts in a different way then the forms?
11:41 kilbith it's not simply saying "that's too big", it's a matter of proportions between the GUI and fonts, and the fonts itself
11:42 sapier to me that sounds same as "fonts are to big compared to menu" which is quite same as "fonts are to big" ;-)
11:42 sapier which is interesting because I'm same oppinion but VanessaE insists on bigger fonts
11:42 kilbith i simply suggest to rollback to the previous scaling way
11:43 sapier there was no scaling before
11:44 kilbith have you seen the 0.4.11 stable screenies ?
11:44 sapier yes but I'm not talking about 0.4.11 where half of scaling was disabled breaking gui on everything not being 72 dpi
11:45 sapier 0.4.11 has a lot of code in it just to drop the result and use something else
11:46 kilbith you thinks it's tolerable to have an initial scaling like that ? https://mediacru.sh/inmOiURnWiM6
11:46 kilbith -s
11:46 kilbith users are not all myopic, eh...
11:47 sapier I could live with it but to be honest I don't really care as long as the result is consistent ... and not crap like the way someone "tried to fix it for 0.4.11
11:48 kilbith ok, so the tabs eat the header's space but that's not a problem...
11:48 sapier if there's a common agreement to reduce the font <-> form aspect I could live with it too
11:49 sapier but I'm quite sure if I do reduce it now tomorrow someone will scream out "it's to small it's to small"
11:50 kilbith no, that's just obvious actually
11:50 sapier if VanessaE tells same I'm gonna change it because she was the one insisting most loud on bigger size ;-)
11:51 kilbith bigger OK, but i guess you had an heavy hand on levelling-up the scaling
11:52 sapier nope it's linear scaling
11:53 sapier if form increases by factor 2 fonts increase by factor 2 ... well as fonts and forms are way different sizes this ain't true for font sizes < 5
11:53 kilbith sorry, i don't meant "levelling-up" for say "vertical"
11:53 kilbith but levelling-up like enlarging overall
11:53 sapier there's no differenc in horizontal and vertical scaling
11:55 sapier those "skinny" font effects are rounding issues if you scale up and down a font. Irrlicht uses bitmap fonts only thus there ain't something like cleartype adjusting the overall result later
11:58 kilbith no way to "bold" them ?
11:58 ImQ009 joined #minetest-dev
12:00 kilbith that'd look more substancial
12:00 sapier well if we "Bold" it it's gonna be bold everywhere"
12:00 sapier I'm not even sure if that conversion code does support it
12:02 sapier no there's no "bold" setting in it
12:02 sapier we could try disabling the antialiasing
12:07 Megal joined #minetest-dev
12:07 sapier no no disabling antialiasing ain't a good idea
12:09 sapier hmm interesting
12:23 sapier well autodetecting dpi results in increase of mainmenu ... maybe we should set default dpi different
12:27 gregorycu I've noticed that the engine treats keys with similar labels the same
12:27 gregorycu For instance, numpad 1 is the same as keyboard 1
12:27 gregorycu Is this a bug or a feature?
12:27 sapier we don't do special keymapping it's most likely irrlicht behaviour
12:28 gregorycu It's us
12:28 gregorycu If you look at keycode.h and operator==
12:28 gregorycu We match on key Char, not the Key value
12:29 gregorycu For keys that have a character
12:29 gregorycu Otherwise we fall back to key value
12:29 sapier true ... never seen that code before
12:30 sapier is there any reason to handle num key's different?
12:30 gregorycu I think this means that people can't bind the num keys how they want
12:30 gregorycu When num-lock is turned on
12:31 sapier yes but on the other hand they can use it for quick select
12:31 gregorycu Actually, I must be mistaken
12:31 sapier if we didn't do it that way  they couldn't
12:31 sapier unless someone adds secondary /tertiary key bindings
12:31 gregorycu Oh ok
12:32 sapier still your question is valid ... guess if someone wants to spend this time it should be changed
12:32 gregorycu So, seemingly by luck, you can actually bind things to the numpad, because of the way operator== was written
12:33 gregorycu But I think this relies on the compilers implementation of std::list
12:33 gregorycu Which is kinda funny
12:34 sapier actually std::list is supposed to behave identical on different compilers ;-)
12:35 gregorycu What I'm suggesting is that operator== is not symettrical
12:35 gregorycu symmetrical
12:35 gregorycu They way it is written here
12:36 gregorycu If you flip the two arguments, you can get a different result :)
12:36 gregorycu Which I think means it doesn't meet the requirement for comparable
12:37 sapier in which situation isn't it symmetric?
12:38 jluc joined #minetest-dev
12:39 gregorycu So, a KeyPress object reduced is the Char, and the Key
12:39 gregorycu Char is the label, Key is the code
12:39 gregorycu With me?
12:40 sapier yes
12:40 gregorycu Wait, I'm mistaken
12:41 gregorycu Sorry
12:42 gregorycu Or am I? Hang on, I'm a little tired, and I'm sick
12:42 sapier no problem :-)
12:46 gregorycu No, I was mistaken, both parts of the or statement are tried, if only only one part of the or statement were tried, I would be correct
12:46 gregorycu If you know know what I mean, it doesn't matter
12:46 gregorycu dont' know
12:46 sapier ok
12:46 sapier I've to leave now, if you find out something different just write it I'm gonna read the log
12:46 sapier left #minetest-dev
12:48 kilbith dammit, we're heavily spammed today :(
12:58 younishd joined #minetest-dev
13:04 ElectronLibre joined #minetest-dev
13:06 ElectronLibre left #minetest-dev
13:07 jluc joined #minetest-dev
13:09 crazyR joined #minetest-dev
13:11 ElectronLibre joined #minetest-dev
13:12 gregorycu joined #minetest-dev
14:11 luizrpgluiz joined #minetest-dev
14:11 luizrpgluiz hi, someone has forecast will come out when the new version of mapgen v8?
14:11 shadowzone joined #minetest-dev
14:26 SudoAptGetPlay joined #minetest-dev
14:31 exio4 joined #minetest-dev
14:36 FR^2 luizrpgluiz: My crystal ball is currently out of order, and the coffee grounds aren't that specific... ;)
14:37 hmmmm joined #minetest-dev
14:41 luizrpgluiz kkkkkkkk
15:12 Player_2 joined #minetest-dev
15:28 luizrpgluiz left #minetest-dev
15:46 roniz joined #minetest-dev
15:52 ElectronLibre joined #minetest-dev
16:07 ElectronLibre joined #minetest-dev
16:17 leat joined #minetest-dev
16:37 ElectronLibre joined #minetest-dev
16:40 twoelk joined #minetest-dev
16:51 Calinou joined #minetest-dev
16:52 kaeza joined #minetest-dev
16:58 Krock joined #minetest-dev
17:15 casimir joined #minetest-dev
17:19 rubenwardy joined #minetest-dev
17:26 acerspyro joined #minetest-dev
17:26 Hunterz joined #minetest-dev
17:35 ElectronLibre joined #minetest-dev
17:39 sapier joined #minetest-dev
17:43 sapier hmmmm tabs are switched in builtin/fstk/tabview.lua L91
17:44 sapier handle_tab_buttons detects if a pressed button is actually a tab selection event and changes the active tab
17:44 sapier if button ain't a tab switch button the button event is passed to the current active tab
17:45 hmmmm right.. which eventually calls switch_to_tab
17:46 hmmmm which to my understanding just calls on_change if it's there, saves settings, and shuffles around the current tab
17:46 sapier yes switch to tab is called on changing the tabs
17:46 sapier no
17:46 hmmmm AHHH
17:46 sapier the tab itself ain't shuffled
17:46 hmmmm it sets last_tab_index there
17:46 sapier it just sets which tab is used
17:46 hmmmm okay i understand how it works now
17:47 sapier event handling and formspec generation are separate
17:47 hmmmm i know
17:48 sapier it's quite primitive ... well compared to things like qt ;)
17:48 sapier and complex compared to html
17:48 hmmmm no I just missed the part where last_tab_index gets set
17:49 sapier ok don't bother to ask
17:49 sapier btw there's an updated version of fstk as pull request fixing some issues left in there
17:52 sapier that version adds support for tab specific data
17:53 sapier but most of the new code is adding in game support
17:55 srifqi joined #minetest-dev
17:58 sapier ShadowNinja is it ok now: https://github.com/minetest/minetest/pull/2051
18:16 crazyR_ joined #minetest-dev
18:34 kaeza joined #minetest-dev
18:42 rubenwardy joined #minetest-dev
18:50 proller joined #minetest-dev
18:52 rubenwardy Is this a bug? https://forum.minetest.net/v​iewtopic.php?f=6&amp;t=10938
18:54 sapier seems to be
18:56 rubenwardy https://github.com/minetest/minet​est/blob/master/src/hud.cpp#L162
18:56 rubenwardy It does take direction into account, but incorrectly.
18:56 sapier I suggest you fix it and we merge the fix ;-)
18:57 rubenwardy I will try
18:57 sapier thanks :-) that's no less we all do ;-)
19:01 rubenwardy bloody hell, the menu is massive
19:03 sapier well we enabled automatic dpi adjustment
19:03 sapier we may need to do some proper baseline adjustment
19:03 sapier ok we enabled for linux ... win32 implementation is still missing
19:09 rubenwardy fixed it
19:09 rubenwardy pushing now
19:09 rubenwardy The person who wrote and documentated HUD will be embarrassed
19:11 rubenwardy #2081
19:12 rubenwardy Please unignore me, ShadowNinja
19:12 rubenwardy https://github.com/minetest/minetest/pull/2081
19:12 sapier that simple? :-)
19:12 rubenwardy Yes.
19:12 rubenwardy Lol
19:12 sapier and noone noticed by now ;-)
19:14 rubenwardy It's a very simple mistake
19:14 sapier things like that happen quite often
19:14 rubenwardy Probably from when it was "pos" and "dir"
19:15 rubenwardy That was fun. I feel like I know the maze a little bit better.
19:15 sapier is there a pos property?
19:15 roniz_ joined #minetest-dev
19:15 rubenwardy It was called "pos" before it was renamed to "position"
19:16 rubenwardy According to this, anyway. https://github.com/rubenwardy/minetest/blob/hud​_direction/src/script/lua_api/l_object.cpp#L48
19:16 sapier better maybe you should handle it same way as for pos, someone may have found out "dir" works and you'd break it on changing
19:17 rubenwardy ok
19:21 rubenwardy Updated
19:23 rubenwardy rebased
19:50 sfan5 sapier: can i merge #2081?
19:50 ShadowBot https://github.com/minetest/minetest/issues/2081 -- Fix direction property of HUD by rubenwardy
19:51 sapier seems ok
19:52 sfan5 merged
19:52 rubenwardy Excellent
20:13 kilbith sfan5: https://github.com/minetest/minetest_game/pulls
20:24 crazyR joined #minetest-dev
20:26 jluc joined #minetest-dev
20:30 monte joined #minetest-dev
20:31 eeew joined #minetest-dev
20:49 prozacgod joined #minetest-dev
21:58 Amaz joined #minetest-dev
21:58 hmmmm for what it's worth, i didn't write all the documentation for hud
22:04 acerspyro joined #minetest-dev
22:06 kaeza joined #minetest-dev
22:08 fz72 joined #minetest-dev
22:10 fz72 joined #minetest-dev
22:12 Amaz left #minetest-dev
22:19 kilbith hmmmm: https://github.com/minetest/minetest/issues/2082
22:19 kilbith and i confirm the bug: https://lut.im/m4r9B9L4/ish6XWJC
22:36 VanessaE kilbith: that's an old bug
22:39 VanessaE the real question is why there are holes and floating lava in the first place
22:54 sapier As you both are here right know VanessaE and kilbith can you please discuss what size the default font is supposed to be ? ;-)
22:54 acerspyro 4
22:56 kilbith ShadowNinja: you don't label my issues/PR on git for a particular reason ?
22:57 ImQ009 joined #minetest-dev
22:58 kilbith if it's not important for you, i can close them
22:58 sapier give him time kilbith
22:59 kilbith strangely the others one have labels except mines
22:59 acerspyro You're special.
22:59 sapier well maybe he has already seen how hot that topic is and doesn't want to burn his fingers ;-)
23:00 kilbith acerspyro, ??
23:00 acerspyro jk
23:01 kilbith sounds like disdain rather
23:01 sapier kilbith: btw we don't have a label for "discussion" anyway
23:02 kilbith discussion ? where ?
23:03 kilbith nothing opened just for discussion
23:06 ShadowNinja kilbith: I missed *one*.  Fixed.
23:12 sapier https://github.com/minetest/minetest/pull/2051 shadow good to merge with those changes?
23:18 VanessaE sapier: imho, default font size everywhere should be around 10-11 point, this seems to be common on other desktop apps, with the option to increase it of course
23:19 kilbith +1  and not skinny
23:19 sapier it's not about a number it's about who big in relation to formspec itself it's supposed to be kilbith doesn't like it as big as it's now
23:19 sapier kilbith: tell irrlicht guys to support freetype I can't fix this
23:20 sapier if for some reason scaling factor happens to cause suboptimal font width to result from the only thing we could do was switching the font in total and hope it's not just gonna switch to a different size
23:21 VanessaE sapier: I think what kilbith means is font hinting needs worked on
23:21 kilbith well, it's to express our wishes through the words in that case
23:21 kilbith it's hard*
23:22 sapier we don't have anyone to fix the font code the only thing changed by font scaling is that all those existing issues now become visible
23:22 VanessaE I get the impression that fonts are "full hinting" in minetest, if it were up to me, that default would be set to "none" or "slight" (or be user-adjustable, or better take it from the desktop environment settings)
23:22 sapier if there's someone having any experience in font code details help is welcome
23:23 kilbith i'll make some examples with Gimp, sapier
23:23 sapier we don't have any support for bold italic or any other of those types
23:23 sapier we only have shadows or now shadows
23:23 VanessaE sapier: it's not a bold/italic thing - it's called hinting
23:23 sapier kilbith: that's as helpfull as nothing
23:24 sapier ttf font is translated to bitmap by a piece of code which is just copied from some forum ... it's always been that way but by now the font size was accidentally at exactly that size where the code resulted in reasonable output
23:24 kilbith that's a demonstrative way for express what we want for the adequation gui/fonts
23:25 sapier ok could help
23:26 VanessaE as for relative size it's simple:  what the default menu had in, say, 0.4.9 before any of the fonts were changed, but increase by around 10% in size.
23:26 sapier argh
23:27 sapier how to increase 12 by 10$
23:27 sapier 10%
23:27 VanessaE well, you could use paypal :P
23:27 VanessaE sorry, bad joke :)
23:27 sapier I don't even have a paypal account ;-P
23:27 VanessaE well approximate it, 12*1.10 = 13.2 so round it to 13.
23:28 VanessaE or even step it to the next whole number cna call it 14.
23:28 VanessaE and*
23:28 sapier which is only 8.3 percent
23:28 VanessaE eh?
23:28 sapier so you're gonna complain about missing 1.7%
23:28 VanessaE 10 percent over 12 is 12*1.10 = 13.2
23:28 VanessaE so round it up to 14
23:28 sapier ok I was at 13
23:28 VanessaE oh right
23:29 sapier 14 is even 16%
23:29 VanessaE I'm just going with he assumption that you're stuck with whole numbers
23:29 kilbith IMO, the GUI scaling is the best at 0.75 : https://cdn.mediacru.sh/9/9gZUXU_MfGlo.png
23:29 sapier well that's about the old size where we did asume 72 dpi while almost everyone has 96
23:31 sapier but that's not the problem ;-) what about fonts?
23:32 VanessaE in kilbith's screenshot, the fonts are actually quite sane imho
23:32 VanessaE just a TAD small for me but tolerable
23:33 kilbith nah, still skinny
23:33 sapier isn't that a regular screenshot?
23:33 kilbith define "regular" ?
23:33 sapier meaning you didn't modify it manuall did you?
23:33 VanessaE kilbith: you mean the fonts' weight is too light.
23:33 kilbith oh no at all
23:34 VanessaE kilbith: which means fonts hinting is st to "full" instead of "none" or "slight"
23:34 VanessaE set*
23:34 VanessaE that's the only way to fix the font weight
23:34 VanessaE is to fix the hinting.
23:35 VanessaE that doesn't require bold, it's just a setting to tweak in wherever the font rendering code is being called I guess
23:35 sapier tell me what you mean with hinting? we don't seem to have a way to change it
23:35 VanessaE sapier: http://en.wikipedia.org/wiki/Font_hinting
23:35 VanessaE I mean this.
23:35 VanessaE this is the problem kilbith is fighting with
23:36 sapier well guys suggest a different ttf font
23:36 sapier we can't configure this codewise
23:36 VanessaE it's not a font issue
23:36 kilbith hinting is just for adjust properly the "shapes" of the caracters, not helps them to gain weight and consistency
23:36 VanessaE kilbith: actually, it does.
23:37 VanessaE it will stop the sudden change in weight by changing size.
23:38 kilbith well, you're prolly right
23:38 sapier ok there's a function for font hinting hidden in our code but for what I see it's enabled anyway
23:39 VanessaE sapier: disable it.
23:39 VanessaE (link to the function?)
23:40 sapier void CGUITTFont::setFontHinting(const bool enable, const bool enable_auto_hinting)
23:40 VanessaE that's the one
23:40 sapier those parameters are set to true by default
23:41 VanessaE set the enable flag to false.
23:41 VanessaE freetype's autohinter sucks anyway :P
23:41 sapier well provide a better one ;-P
23:41 VanessaE heh
23:42 VanessaE can't, it's illegal :P
23:42 VanessaE make it user-configurable for now, see if you can read it from the OS/desktop settings some time later in the future.
23:42 sapier I don't think you want it disabled
23:43 VanessaE how's it differ from what the same setting on my desktop env does?
23:43 VanessaE hm, I talk good engrish ;)
23:44 sapier it's a little bit more blury and at some size you get white boxes instead of a font
23:44 VanessaE the slight blur is normal.
23:44 VanessaE you're seeing the REAL render of the font
23:44 sapier white boxes instead of text ain't
23:44 VanessaE but the boxes?  that's weird.
23:45 sapier well it's antialiased too so not the "real" but you don't want the non antialiased fonts either
23:45 VanessaE well you know what I meant ;)
23:45 VanessaE vertical strokes are not being artifically lined up to screen pixels anymore
23:46 sapier still that change wouldn't help anything
23:46 VanessaE actually it would help his issue if not for those boxes.
23:46 sapier could you guys try non freetype maybe it's better ;-P
23:47 VanessaE because the artificial alignment of pixels causes it to jump from one-pixel-wide strokes to two-pixel wide once it hits a certain size
23:47 VanessaE whereas without hinting, the font just gets a little bigger
23:47 VanessaE that's why I hate hinting.
23:47 sapier no they don't seem to be bigger at all
23:47 VanessaE they do.  I tried it and saw the same effect.
23:48 VanessaE they dont' get bigger, just thicker
23:48 VanessaE so at a certain size, roughly 24 on the old scale, 16 or so on the current scale (I think), the hinting causes the weight to change.
23:48 VanessaE that's what kilbith is complaining about.
23:48 sapier not even to a noticeable degree at our usual size
23:49 VanessaE it's a hard issue to solve; you're right that bitmap fonts might be needed to solve it for him
23:50 sapier freetype fonts are converted to bitmaps for irrlicht too so I don't see how it's supposed to solve the issue
23:50 VanessaE he could always supply his own bitmaps ;)
23:50 sapier we can just try to find another font which looks more thick at default size ... scale it a little bit and you're gonna have same issue again
23:51 sapier if you set the scaling you see exactly this effect
23:51 sapier every few steps the font looks thick compared to menu and then becomes thin again till it gets thick again
23:52 sapier imho we can't "fix" the issue but just get some good looking initial value
23:54 VanessaE idk what to do then
23:54 acerspyro magix
23:54 VanessaE get the size right and let the weight take care of itself; fix whatever's broken in the hinting setting, make that user configurable
23:55 sapier talk to kilbith and find something both of you can live with VanessaE
23:55 kilbith yep, DejaVu Sans font is intrinsically skinny at >14, FreeSans is better
23:55 VanessaE FreeSans is fine by me - I can always select another font myself as well
23:55 VanessaE I'm only concerned with well-defined, consistent *sizes* throughout the program
23:56 VanessaE unless a mod explicitly changes the font size, it should default to and be rendered to some specific point size, whatever that is.  that's all I'm worried about at this point.
23:56 acerspyro monospace
23:56 acerspyro There, fixed
23:57 sapier kilbith: for what I see we have liberation sans
23:58 VanessaE sapier: since when?  I thought the default was Dejavu?
23:58 kilbith sapier: this is not i notice in my client
23:58 kilbith this is the DejaVu one here
23:59 sapier are you sure you're using freetype?
23:59 kilbith 200% surze
23:59 kilbith -z
23:59 sapier because dejavu is non freetype
23:59 sapier and for mono fonts
23:59 kilbith well, i'm not mad, i compiled with freetype
23:59 sapier I don't even have a dejavu ttf file in her

| Channels | #minetest-dev index | Today | | Google Search | Plaintext