Time  Nick        Message
00:39 sapier      vanessae plz check again now formspecs scale again ... if you have time try non freetype mode to it should have scaling fonts now too
00:39 sapier      not as smooth as freetype but still scaling
00:40 VanessaE    I'll try later
00:50 sapier      still if those formspec changes are used mods will have to update their forms
07:43 sapier      VanessaE: is it intentional dreambuilder crashing on current stable?
09:09 Megaf       well, my arm board is faster than my macbook pro to compile minetest
09:10 Megaf       proller: I think that bug is fixed now, but server didnt crash anymore, because of that bug
09:25 sapier      https://github.com/minetest/minetest/pull/1563
09:30 sapier      VanessaE: https://gist.github.com/sapier/f3d6e635c4c8c7d298ce this is something similar to how a fix for unified inventory would look like
15:29 VanessaE    sapier: I was not aware that it was crashing on 0.4.10-stable.  what's the crash?
15:31 sapier      https://gist.github.com/sapier/466021c31965c66a5aab
15:32 VanessaE    nore: ^^^^^^^
15:32 sapier      am I the only one testing mods on stable too? ;-)
15:33 VanessaE    wait a minute, someone else already reported that bug
15:33 VanessaE    I Was under the impression that was already fixed in pipeworks.
15:35 Zefram_Fysh that's not the latest pipeworks code
15:36 VanessaE    I show my local pipeworks repo is up-to-date, and dreambuilder has the latest of that.
15:36 sapier      well I don't pull twice an hour
15:37 VanessaE    I do update dreambuilder as many as 10 times in a day ;)
15:37 sapier      sorry but even for a "rolling release" that's way to much
15:38 VanessaE    well for the secondary build there is always a build ID in the ZIP
15:38 VanessaE    and of course git has its history
15:38 sapier      ok seems to help
15:40 * VanessaE  wonders what prompted sapier to check dreambuilder again :)
15:43 sapier      comparing old and new unified inventory
15:43 VanessaE    ah
15:43 VanessaE    easy way to get it I suppose :)
15:47 sapier      well didn't work ;-)
15:47 VanessaE    heh
15:48 sapier      did you have time to check the new formspec variant with scaling fonts?
15:50 VanessaE    no, I was asleep :)
15:50 sapier      come one sleep ... wasted time ;-)
15:50 VanessaE    haha
15:51 sapier      https://github.com/minetest/minetest/pull/1563 moved the formspec things to a separate pull yet the font engine is of very little use without that addon
19:37 ShadowNinja PilzAdam: You're an admin on the dev wiki, right?  Can you mass-revert this: http://dev.minetest.net/index.php?title=Lua_code_style_guidelines&diff=2490&oldid=2298  Reason: It should be on the talk page and it's not even valid markup.
19:37 ShadowNinja sapier: ^
19:38 sapier      can you just fix the comments instead of complaining about the style
19:38 ShadowNinja (Also full of typos and visual warts)
19:38 PilzAdam    I could
19:38 ShadowNinja sapier: Maybe.
19:38 ShadowNinja sapier: Give me specific issues and possibly examples of how to fix them.
19:38 sapier      unless you want me to do same things you do you shouldn't try to force apply your personal coding style without discussion
19:39 PilzAdam    sapier, could you move your comments to the talk page?
19:39 ShadowNinja sapier: It obviously isn't frorcibly applied, it even says it's just guidelines.
19:40 sapier      no I can't because then noone would read it
19:40 PilzAdam    sapier, add "Draft (see talk page)" to the top
19:40 PilzAdam    in red and bold
19:40 sapier      and it's not that much as it can't be fixed
19:40 PilzAdam    and stop using "--------------------------------" and "############################" in wikis
19:40 ShadowNinja ''Comment sapier: Why not doxygen?''  --  Because doxygen is designed for C++ code, I've seen how you use it in mobf and it's horribly ugly.
19:40 HLuaBot     Unknown command "Comment". Try "help".
19:40 sapier      and luadoc doesn't even work at all
19:40 sapier      so what's the difference?
19:41 ShadowNinja sapier: Try ldoc then, it uses a compatible format.
19:41 sapier      doxygen with lua adapter works way better then luadoc atm ... maybe ludaoc evolves there
19:41 sapier      so why force some doc tool in coding style while it's not even marture?
19:42 ShadowNinja sapier: Show me how doxygen Lua comments would look.
19:42 sapier      see mobf
19:43 sapier      http://animalsmod.bplaced.net/doxygen/ that's result
19:44 PilzAdam    should I move the guidelines to ShadowNinja's userspace first?
19:44 sapier      And most of my other comments are about unspecific guidelines that can't even be followed if you try.
19:45 sapier      why not just discuss about them and get a result ... at least my comments aren't that much
19:46 ShadowNinja sapier: http://paste.ubuntu.com/8082466/
19:46 sapier      yes that's mixed luadoc/doxygen
19:46 ShadowNinja LuaDoc format is waaay shorter, and simpler.  It also makes use of context.
19:46 sapier      luadoc for eclipse doxygen for usefull output
19:46 sapier      still luadoc can't create a usefull documentation
19:47 ShadowNinja sapier: It can, for example http://stevedonovan.github.io/ldoc/
19:47 sapier      USEFULL
19:47 ShadowNinja sapier: Or http://stevedonovan.github.io/Penlight/api/index.html
19:48 ShadowNinja Very usefull.
19:48 sapier      any you consider this messed up style usefull?
19:49 sapier      I might be oldfashioned but I like graphics showing me dependencys
19:49 ShadowNinja sapier: Of course.
19:49 sapier      because I usually need way more time for reading 20 pages of text compared to looking at a single half page graphic
19:49 ShadowNinja sapier: We don't need fancy graphics and JS and <video> and what have you.
19:50 ShadowNinja Puero que exaherado eres!  ;-)
19:50 sapier      Oh so ShadowNinja is now Minetest
19:51 sapier      nice to know
19:51 sapier      any other opinions about sticking to a premarture incomplete doc definition?
19:52 sapier      I see instead of taking place in discussion you prefere to get silent again
19:52 ShadowNinja Comment sapier: Unspecific definition potentially resulting in shitload of discussions -- Then think of a better description.
19:52 sapier      make it a suggestion isntead of a rule, quite simple
19:53 sapier      we have same suggestion for c++ coding style why make it a must have for lua?
19:54 ShadowNinja PilzAdam: So, can you revert that now?  I'll move sapier's thoughts if he won't.
19:54 sapier      so you're ignoring valid comments ... nice to know
19:55 sapier      well guess I can adapt to that style of contribution
19:58 ShadowNinja Comment sapier: Way to unspecific and potential source of lots of discussions ... what's ok 5 spaces 10 15?  --  Common sense is assumed, we aren't all computers.  Use your best judgement.
19:58 sapier      that's fine for a suggestion but not for a rule
19:59 sapier      You'll be the first to complain about "but it's 1 space to much there"
19:59 ShadowNinja Comment sapier:  nil evaluating to false on boolean comparison doesn't mean they're equal.  --  Obviously.
20:00 sapier      so why do you make rules how to do it?
20:00 ShadowNinja Comment sapier: Paranthesis always improve readability a lot  --  No, too many of them make code harder to read.
20:00 ShadowNinja sapier: Because it's more readable.
20:00 sapier      yes (((((((((((((((((((((())))))))))))))))))))))))))))))) will cause this
20:01 sapier      but a single set will never make a difference
20:01 sapier      as of making it worse
20:03 sapier      and you should consider other's comments too as most ppl most likely haven't even noticed those rules
20:04 sapier      I will accept beeing voted agains my suggestions I just don't like rules to be declared without even discussing about them
20:15 sapier      ShadowNinja: You're not consider any comment right?
20:16 ShadowNinja 00:09, 19 August 2014‎ Sapier (Talk | contribs)‎ . . (66 bytes) (+66)‎ . . (Created page with "Noone will read this page so why even write something in here? ...")
20:16 ShadowNinja ^ Not helpfull.  And I moved your comments.
20:18 ShadowNinja The one I consider most valid is "Why are spaces around equals signs in short tables allowed to be omitted? They help readability."  I made them optional because that's what everyone does currently, and it helps keep short tables short.
20:19 sapier      you only allowed them for single line tables how long do you expect variable names and values to become there?
20:19 ShadowNinja And the -------... lines look ugly to me.  Two speces are plenty for me to easily distinguish functions.
20:20 sapier      you did allow two new lines for separating blocks within functions too so where's the obvious difference?
20:20 ShadowNinja sapier: The most common case is groups, so keys are 4-16 chars or so and values are 1 char.
20:20 sapier      there's almost no optical difference between some  block change and another function
20:21 ShadowNinja sapier: I didn't say that, although I didn't say otherwise.  Perhaps I should.
20:21 sapier      and yes if you read it precise you will see it ... but that's what I'm talking about a line of --- will enable you to see function end without reading
20:22 sapier      btw you forgot two important comments
20:23 sapier      and you don't need to link to a discussion page for me as I'll not read it for sure ... I already have way to many things to read
20:24 ShadowNinja Comment sapier: Inconsistent example at beginning there's shown to use spaces around binary operators (as told some lines below!!!) -- Elaborate.
20:24 ShadowNinja And the first comment needs to be redone.
20:25 ShadowNinja "Explanation of what makes a good comment needs to be more specific." or so.
20:25 sapier      There's no really good explanation for a "good comment" as this is highly subjective
20:26 sapier      one of my professors once said removing all code you should be able to reconstruct code ... well that's one extreme position
20:27 ShadowNinja sapier: And == nil is fine where it's needed.  But if nil and false are equivalent, or you will only ever receive one of those values, then without == is easier to read, and keeps things compatible if a function starts returning nil.
20:28 sapier      and what's wrong if someone does different except of you not beeing able to complain about it?
20:28 ShadowNinja sapier: And you MUST make good comments, it's not optional.  It's just that the definition of a good comment is fizzy, you're expected to use some common sense.
20:29 sapier      it's not wrong it's not less readable  ... and it's not gonna fill up your disk for sure
20:29 sapier      well if you believe "good" comments to be a must I'm just not gonna make any comments at all because I'll not discuss about what good and bad comment is
20:30 sapier      And I believe everyone once having to discuss about this may tend to do same next time
20:31 ShadowNinja sapier: Give me a break.  Are you trying to give me a hard time?  You're perfectly capable of creating helpfull comments without someone having to spell out to you exactly what a good comment is.
20:32 sapier      sorry ShadowNinja but I have experienced more then once how picky you are about coding style ... thus I read them the way I expect you to interpret them
20:35 ShadowNinja sapier: Just do your best.  I'll try not to be too bad about them.
20:35 sapier      I know your best ;-P
20:36 sfan5       sapier: why does the android version suck so much?
20:37 sapier      because everyone complains unspecific instead of making suggestions what exactly could be done better
20:37 ShadowNinja sapier: Multi-singleplayer-world support.
20:37 ShadowNinja sapier: Like SurvivalCraft has.
20:38 ShadowNinja (Another Android game in our genre)
20:38 sapier      not enough room on screen but you can switch to pc menu quite easy ... yet it's most likely not gonna be very usable
20:38 sfan5       1) the menu font is ugly 2) MOD is not an acronym 3) putting the up button between left and right is the worst idea ever 4) the jump and descend things look compressed
20:38 sfan5       and the font is too big
20:39 sfan5       I can't even see the pause menu completly
20:39 sapier      sfan5: how many android devices do you have?
20:39 sfan5       s/completly/not obscured by text/
20:39 sfan5       1
20:39 sfan5       with my Raspi 2
20:39 sfan5       but I'm not going to test it there
20:40 sapier      ok I have 5 here and this is best middle way between all of those devices I didn't find another way where you could at least use it on all devices
20:40 ShadowNinja sfan5: I think the jump/descend buttons are supposed to look 3D.
20:40 sfan5       sapier: can you add a menu entry to change the font size?
20:40 ShadowNinja (Like docky-style 3D)
20:40 sfan5       sapier: and hide the _game mods from the configure menu by default, I'm pretty sure it confuses people
20:41 ShadowNinja sapier: Does the configure menu support nested modpacks yet?
20:41 sapier      not yet sfan5 as font sizes don't scale independent ... but there's already a pull request making it possible
20:41 sapier      it's exactly same as on pc so no
20:42 sfan5       ShadowNinja: I don't think it does
20:42 sfan5       and is there still the modpack-with-the-name-of-a-mod bug?
20:42 sapier      for what I know that one was fixed quite some time ago
20:43 ShadowNinja sapier: Alright, well can you fix that soon?  I have my mods set up like mods/mods/technic-related/technic/technic/init.lua at the deepest level, so the menu's unusable.
20:43 sapier      nested modpacks?
20:43 sapier      no
20:43 sapier      that's a major rewrite in configure code
20:43 sfan5       sapier: so.. will you fix the issues?
20:43 sapier      which issues?
20:43 ShadowNinja sapier: Why?  Formspec table added support for that, it was the main reason it was added in my understanding.
20:44 sfan5       wat
20:44 sfan5       2) MOD is not an acronym 3) putting the up button between left and right is the worst idea ever 4) the jump and descend things look compressed
20:44 sfan5       at _least_ 2 requires no debating
20:44 sapier      2 yes 3 no (cause it renders some of my devices unusable)  4) intentional to save room for minetest
20:45 ShadowNinja Also, rename mods to plugins.  :-)  </unrelated>
20:45 sapier      ShadowNinja: no for sure not
20:45 sapier      mods an plugins have different meaning
20:45 ShadowNinja sapier: Exactly.  And what we're using now aren't really mods.
20:45 sapier      a mod is as name says a modification of base game while plugin is something additional
20:46 sapier      no it's what you'd like them to be to be able to load and unload them
20:46 sapier      but that's not what they're designed to be
20:46 ShadowNinja sapier: Yes.  Mods don't actually change Minetest, they just add content to Minetest.
20:46 sapier      you're wrong
20:46 sapier      some mods even replace core api
20:48 ShadowNinja sapier: Oh?  How do they change the Minetest core then?  They can decorate methods, but they can't, eg, add real VoxelAreaEntities, that needs to be done in the core and they can only touch the provided API.
20:48 sapier      sfan5: how are your design capabilies?
20:49 sapier      minetest core consists of c++ as well as builtin code additionally mods can overlay core.* functions effectively beeing able to change almost everything
20:50 sapier      sfan5: because buttons are supposed to be some initial things to make ui work, I know they're not very nice but I'm not a designer
20:52 sapier      sfan5: 2 is fixed
20:52 ShadowNinja sapier: No, it can access the API and wrap arround it.  It can't actully change the core.  If it could you would be able to add things like VAEs to Minetest (real ones, not giant hacky entity cloud things that only work for a few drawtypes) like MineCraft mods can do, because MineCraft mods are real mods, they actually patch the core.
20:54 ShadowNinja sapier: Have you tried SurvivalCraft?  It's controls might be worth looking at.
20:54 sapier      I will not change this name because you want some reason for breaking old mods ... if you want this start a discussion get supporters and let majority decide ... at current state you're not even close to convinving me a mod isn't a mod but a pligin
20:55 ShadowNinja sapier: You wouldn't be the one to change this, and it can be done compatibly of course.
20:55 sapier      I don't think you can do it in a compatible way
20:56 sapier      Just one simple example if you prefix all mod variables mods depending on those variables are broken, if you don't you're unable to remove them later
20:56 ShadowNinja sapier: core.get_mod* = core.get_plugin*  if (fs::isFolder("mods")) { fs::move("mods", "plugins"); }
20:57 sapier      I was talking about the mod unloading
20:57 ShadowNinja sapier: Oh, I'm not.
20:57 ShadowNinja sapier: I'm talking about renaming mods to plugins, because that's a better name for them.
20:58 ShadowNinja Reloadability is nice, but unrelated.
20:58 sapier      renaming the folder will just be a shitload of discussions incompatibilitys and things like that because you missed wherever that damn name was used
20:58 sapier      and as usually you wont fix it
20:59 ShadowNinja There would be a lot of changes, indeed.  On the forum, wiki, dev wiki, main site, and, of course, the code.
20:59 sapier      btw I'll mention the player unloading issue you caused by unloading the player files till you finally fix it
20:59 sapier      because I will not fix it
21:00 ShadowNinja sapier: What issue?
21:00 VanessaE    the only thing renaming them to "plugins" will be any good for is resolving the confusing english syntax of talking about a server moderator and a server modification/"plugin" in the same discussion.
21:00 VanessaE    frankly I think that's a waste of time otherwise.
21:00 sapier      ShadowNinja: you made players to be non persistent now resulting in player object unloading causing player to be unloaded
21:01 sapier      sadly this happens too in case of a object can't be added to a block because of max objects
21:01 ShadowNinja It also makes it clear that they aren't like MC mods (don't require core changes and client-side installations).
21:01 VanessaE    and THAT ^^^^^ needs fixed.
21:01 sapier      there's a workaround pull request but I'm not sure it really fixes the issue or does even cause more problems
21:02 ShadowNinja sapier: No, it worked like that before.  Before it set the peer_id to 0, which was interpreted as a disconnected player.
21:02 sapier      because for what I understand it just results in the object not beeing deleted causing a dangling object not beeing saved in any mapblock
21:03 sapier      ShadowNinja: I don't know exactly how it is triggered but it's triggered since player can be unloaded
21:03 ShadowNinja The player and it's SAO are weirdly tied together.
21:03 sapier      I know that's another reason why I refuse to fix it
21:04 ShadowNinja The player SAO destroys it's player when it's unloaded, you told me where to put some code in the connection to clean it up in a nicer way, but I don't know if I added that.
21:04 sapier      still now a player is deleted in some situations without beeing disconnected
21:05 ShadowNinja Oh, BTW, did I miss anything important?  I've been away for the last week.
21:05 sapier      https://github.com/minetest/minetest/pull/1557
21:05 sapier      farming is broken
21:06 VanessaE    something else in minetest_game is broken as well
21:06 VanessaE    10+ second lags for no discernible reason
21:06 VanessaE    ever since _next was merged in
21:06 VanessaE    unless the internet is just breaking.
21:07 sapier      https://github.com/minetest/minetest/pull/1561 this one combined with https://github.com/minetest/minetest/pull/1563 might fix sfan5's font scaling issue but as it fixes formspecs positioning mechanism old formspecs containing workarounds are slightly messed up ... all formspecs contain workarounds
21:07 sapier      that something most likely is farming
21:08 sapier      seems like all abs are done in same server step
21:08 VanessaE    ew.
21:08 VanessaE    that would explain the multi-second lags you saw in plantlife also:  it uses growing ABMs for the mapgen'dtrees to spread out their load
21:08 sapier      ahh and realbadangel_ fixed faces shading ... (again ;-))
21:09 sapier      VanessaE: that's just a guess ... but as usually there's almost no abm load by farming except of those huge spikes It'd be a educated guess
21:19 sapier      ShadowNinja: what do you think about the mod profiling thingy?
21:19 VanessaE    sounds reasonable to me.
21:20 kahrl       sapier, ShadowNinja: no, 1550 doesn't fix the "player is deleted if there are too many objects in mapblock" issue
21:20 kahrl       https://github.com/minetest/minetest/pull/1550
21:20 kahrl       it just prevents mods such as worldedit from inadvertently deleting player SAOs
21:21 sapier      ok so we don't even have a workaround right now
21:21 VanessaE    which *does* work.
21:21 kahrl       right
21:22 kahrl       I looked into the other bug, but I couldn't figure out why the code wanted to make player SAOs static at all
21:22 kahrl       (I also only could reproduce it once, then never again for the next hundred tries)
21:22 sapier      aren't all objects saved in some sort of fixed interval?
21:23 sapier      ok maybe a dynamic interval not sure about that
21:23 kahrl       sapier: they shouldn't be if isStaticAllowed() returns false
21:23 sapier      hmm
21:23 sapier      guess that's gonna be some work to be fixed
21:24 sapier      another thing opinions about fixing the formspec positioning
21:24 sapier      in order to get scaling fonts we need to do that
21:24 sapier      because some formspec elements base uppon font size
21:27 Zefram_Fysh it would be nice in the long run to be able to explicitly incorporate font size into form element positioning, rather than only specifying by inventory slots
21:28 sapier      you can't specify it by inventory slots too
21:28 Zefram_Fysh for example, I want the top of a label to be 1 inventory slot below top to form, and then the top of the next label to be 1 inventory slot plus 1 font height below top of form
21:29 sapier      I already thought about adding that feature but each element and (potentially) whole drawing could woul've to be modified
21:30 sapier      Zefram_Fysh: for this usecase use a textarea ;-)
21:30 Zefram_Fysh this would need each coordinate of position and size to be specified in two components, an inventory slot component and a font height component
21:31 sapier      while I see possible usecases imho that'd be overkill for a simple form language
21:31 sapier      we don't wanna reimplement latex
21:31 Zefram_Fysh or the coordinates could be kept simple if we decide on a fixed ratio of font height to inventory slot height
21:31 sapier      actually my patch does exactly this
21:31 Zefram_Fysh a fixed ratio would make layout way easier to get right
21:32 sapier      yet there's some arithmetic and font definition error
21:32 sapier      e.g. there ain't a font size 13 for non freetype fonts
21:32 sapier      so even if calculation results in 13 size 14 will be used
21:34 Zefram_Fysh that's not too bad.  if it's fixed within some bounds, it's reasonably easy to make sure a form layout works with the whole range of ratios
21:34 sapier      as I said try https://github.com/minetest/minetest/pull/1563 it's supposed to behave that way
21:35 sapier      I might have missed some elements so tell me if you find them ... I tried not to change the inventories themselfs
21:36 Zefram_Fysh the difficulties I have with forms at the moment, for technic and unified_inventory, are about trying to align text and non-text elements, and to determine how much space to leave for text.  these difficulties arise specifically because of the extreme variablility of text size relative to inventory slots
21:36 sapier      well if you did one time look at what I showed you you'd know that I fixed that issue too
21:37 sapier      fixing all of this is reason for breaking so much things ;-)
21:37 Zefram_Fysh for example, I'd like to align a text label with the text on a button next to it, but the specified position determines the location of the bottom of the label and the middle of the button, so variable heights change the alignment
21:38 Zefram_Fysh relax, I'm not saying you've failed to address these
21:38 Zefram_Fysh I'm trying to provide exposition of what the practical issues are
21:39 sapier      goddamn  open http://imgur.com/qwMqupb read the text "Some test above 1000" ... look at the end of it to see it's exactly aligned to button caption
21:39 Zefram_Fysh BUT fixing these things doesn't require changing the way form element positions are specified
21:40 sapier      sorry but you want the positionong to be fixed but tell me you don't need to fix the positioning
21:40 Zefram_Fysh yes, you've got some aligned elements there
21:41 sapier      https://github.com/sapier/minetest/blob/formspec_pos_cleanup2/builtin/mainmenu/tab_test.lua is the code for that tab ... case you're not to frightened to click a github link
21:42 Zefram_Fysh I want to be able to align elements.  I *don't* need that alignment to be achieved by giving the same numerical Y coordinate to both the label[] and button[] elements.  it's acceptable for the alignment to require some offset between those Y coordinates, as long as that offset is predictable
21:43 Zefram_Fysh the problem with the situation before your patch isn't that it requires an offset between those coordinates, but that the offset that's effectively required varies according to the user's window size, of which I have no knowledge
21:43 sapier      Zefram_Fysh: I wont discuss this again if you believe this to be possible do it ... and don't only keep talking about like you did for the mapblock update thingy
21:44 sapier      the problem before is that offsets base on different parially interlinked scaling factors
21:44 Zefram_Fysh with the prediction preservation, if that's what you mean, I did produce a patch.  and still intend to produce a revised version
21:45 sapier      as you said "intend"
21:45 Zefram_Fysh as you've asked, sure, I'll work on doing a variant of your scaling patch
21:46 sapier      I will not accept a patch just post applying a fourth scaling factor uppon all elements
21:47 Zefram_Fysh I'm not clear what you mean by "just post applying a fourth scaling factor"
21:47 sapier      unless you sign a contract to support that code until end of your life ;-)
21:47 sapier      of course scaling can be "fixed" by adding a new scaling factor  and just post aply that one prior finally creating all form elements
21:50 Zefram_Fysh I'm still not clear what you mean by it or what would be the problem with it.  chances are what you're trying to rule out is something silly that I'm not going to be anywhere near
21:52 sapier      ShadowNinja: stop rewriting my comments on discussion page claiming me to have written them
21:54 sapier      "what would be wrong with it?" are you kidding? there already are position calculations beeing 2 lines long having 3 variable values in them and you're asking what would be wronb by just applying another variable value?
21:56 Zefram_Fysh I don't imagine introducing another configuration parameter, if that's what you mean by "another variable value"
21:57 sapier      sorry but I don't think we'll ever understand each other just write your suggestion and let me see what you really intend
21:57 Zefram_Fysh sure
21:58 sapier      but I still believe fixing the positioning would be worth it as you'd be able to really write a form instead of experimenting how it might look correct
22:01 Zefram_Fysh I largely agree with that.  we could certainly do with nicer ways to specify positioning.  I only disagree with tying that to font scaling and with breaking compatibility when adding it
22:01 sapier      Zefram_Fysh: you did realize that some elements do base uppon font size ... yes?
22:02 Zefram_Fysh yes, I've looked quite a bit into the form layout code
22:02 sapier      ok, buttons as well as field height ONLY depends on font size
22:03 Zefram_Fysh indeed
22:03 sapier      so next step slot height does not depend on font height at all
22:03 sapier      still with me?
22:04 sapier      Zefram_Fysh: ?
22:04 Zefram_Fysh in current code they're unrelated, yes.  but the point of your change is to link font height to slot height, isn't it?
22:04 sapier      no
22:05 Zefram_Fysh no?
22:05 sapier      there's two things
22:05 Zefram_Fysh somewhere above I said it'd be useful to have a fixed ratio of font height to slot height, and you said that's exactly what your patch does.  did I misunderstand you?
22:06 sapier      there's more
22:06 sapier      the default font size only bases uppon gui_scaling_factor and dpi
22:06 sapier      last one is now fixed (but can be configured by user)
22:07 sapier      what my last patch adds is enabling formspec menu to decide for a own font size to use for the form itself (changing the default font would mess up all other fonts in minetest)
22:07 sapier      all other texts in minetest
22:08 sapier      in theory you could use a different font per element (in practice it'd be a lot of work to make it really work)
22:09 Zefram_Fysh font per element would be rather a lot of layout complexity that we don't really need
22:10 sapier      that's why I didn't spend time on it
22:10 sapier      yet font means font size too
22:11 sapier      you don't get different size by changing the "size" but by using a different font
22:11 Zefram_Fysh right
22:11 VanessaE    I could just imagine cramming x font strings into a formspec element :)
22:12 sapier      I'd gues most would be adding a font size to formspecs but unless we did settle the current issue I'll not waste any time in this ;-)
22:13 Zefram_Fysh crap, your font_engine branch segvs repeatably for me.  I can't get into the game to test anything
22:14 sapier      latest one?
22:15 sapier      freetype or non freetype?
22:15 Zefram_Fysh b7a96e8
22:15 sapier      gui_scaling or not?
22:16 Zefram_Fysh my minetest.conf has gui_scaling=1.0
22:16 sapier      that should be correct
22:16 sapier      can you give me a backtrace?
22:16 Zefram_Fysh CMakeCache.txt:ENABLE_FREETYPE:BOOL=OFF
22:16 sapier      ahhh
22:16 sapier      non freetype mode
22:17 Zefram_Fysh formspec_pos_cleanup2 segvs even earlier, before showing initial menu
22:17 sapier      yes you're missing fonts
22:17 Zefram_Fysh OK, let's see about installing some
22:17 sapier      #font_path = <full path to minetest dir>/minetest/bin/../fonts/lucida_sans
22:17 sapier      #mono_font_path = <full path to minetest dir>/minetest/bin/../fonts/mono_dejavu_sans
22:18 sapier      add those line to settings
22:18 sapier      well remove the # ... I'm fixing the default values for those settings
22:18 Zefram_Fysh to minetest.conf?
22:18 sapier      yes
22:19 sapier      you're really not using freetype?
22:19 Zefram_Fysh still segvs
22:19 sapier      does it find the fonts?
22:19 Zefram_Fysh 23:19:10: ERROR[main]: WARNING: Font file was not found. Using irrlicht default font.
22:19 Zefram_Fysh apparently not
22:19 sapier      that's bad
22:20 sapier      let me try with non freetype build I just tried with freetype and freetype disabled by setting
22:20 Zefram_Fysh the lines you gave me look incompatible with those in minetest.conf.example
22:20 Zefram_Fysh minetest.conf.example:#font_path = fonts/liberationsans.ttf
22:21 Zefram_Fysh still get the same results with "font_path = fonts/liberationsans.ttf"
22:21 sapier      ttf --> freetype
22:22 Zefram_Fysh I'm running from the build dir, so the relative "fonts/" should be correct
22:22 Zefram_Fysh ah, I see
22:22 sapier      and I replaced the old fonts by new ones adding fonts with different sizes for non freetype mode
22:23 sapier      without freetype you didn't have a chance to get a different font size
22:23 Zefram_Fysh I do have libfreetype installed.  wonder why the configuration didn't pick it up
22:23 sapier      it's not enabled automaticaly
22:23 sapier      yet I discovered freetype font loading code to be quite unstable too ... it doesn't check for any error and just crashes
22:24 sapier      e.g. if someone wants a font with size 0 or passes a png instead of ttf file
22:24 Zefram_Fysh just what we need
22:24 sapier      same does irrlicht if you pass a ttf file instead of png/xml one
22:24 sapier      so font config errors usually cause a crash to happen
22:25 Zefram_Fysh OK, rebuilding with freetype enabled
22:26 Zefram_Fysh oh, sorry, never gave you a backtrace.  do you still want one?
22:27 sapier      I guess it's been a misconfiguration of fonts
22:27 sapier      btw you need to remove those lines in order to get freetype work
22:27 Zefram_Fysh right
22:28 sapier      I just pushed a commit that is supposed to fix the non freetype build
22:29 sapier      any luck with freetype?
22:30 Zefram_Fysh still compiling
22:31 sapier      don't try to set dpi to 1000 it's gonna crash ;-) freetype fonts that big aren't handled by our freetype code
22:32 Zefram_Fysh ok, that works.  thanks
22:32 sapier      if you find issues in there like mispositioned elements or crashes tell me ... it's not 100% stable by now
22:33 sapier      well it's me to thank you found the non freetype bug