Time Nick Message 00:05 iqualfragile no, actual outdated links, to the forum to be more specific 00:24 ShadowNinja Outdated, but they still work due to my compat code. I'll fix them... 00:31 ShadowNinja Done. 12:30 sfan5 ShadowNinja: I can merge the big-endian fixing patch, right? 13:47 iqualfragile question: should i migrate the mmdb to use the new forum user database as authentification backend or migrate the old accounts into the native django acount format? 13:48 sapier hmmm what about mixed mode? 13:49 sapier wait native django would be a separate db ? 13:52 iqualfragile sapier: table, seperate table 13:52 iqualfragile but it could as well be put in a seperate database, yes 13:52 sapier so you'd need a separate password compared to same pw on using the forum auth? 13:52 iqualfragile sapier: yes 13:53 sapier then I'm for for forum auth 13:53 sapier forcing users to remember another password wont help mmdb acceptance 15:14 RealBadAngel hi sapier 15:15 RealBadAngel whos about to push the changes we made yesterday? i mean no wrap and height instead of ascender? 15:53 RealBadAngel proller, how about contributing back with kv storage? 15:54 proller why? 15:54 RealBadAngel because you are takin everythin? give somethin in exchange :) 15:55 RealBadAngel and mainly because im askin you 15:56 proller its not my code 16:19 RealBadAngel you have modified it numerous times already 16:20 RealBadAngel but thats not the point 16:21 RealBadAngel we can go hard way if you want 16:21 RealBadAngel i will forbid to use shaders code in freemier 16:21 proller how? 16:21 RealBadAngel simply, by adding explicit rule 16:22 sapier RealBadAngel: I tried yesterday but could you post the final changes, you mentioned some additional locations where to use height 16:22 sapier RealBadAngel: sad to tell you, but you can't 16:22 RealBadAngel rly? 16:22 sapier if you add a rule like that we can't use it in minetest either 16:22 RealBadAngel its not fuckin fair 16:23 proller 8-) 16:23 proller no, 8-P 16:23 RealBadAngel hes takin everythin he wants to 16:23 sapier well that's the downside of lgpl 16:24 RealBadAngel and im just askin for kv 16:24 sapier even if he did want to do he couldn't 16:25 RealBadAngel hmmm, is that possible to take ones code and relicense it? 16:25 RealBadAngel under different one? 16:25 sapier no the only one who can do relicensing is copyright owner 16:25 sapier and copyright stays at author if not explicitly transfered 16:25 RealBadAngel so how the fuck is he relicensing mt? 16:26 sapier he can't 16:26 RealBadAngel freeminer is another license 16:26 sapier the code from minetest included in freeminer is still lgpl but anything added is gpl 16:27 RealBadAngel how can you tell which line is under which license> 16:27 sapier actually thats the problem ;-) 16:28 sapier as there's no (usable) way to do you'll have to use the most limiting one and for freeminer it's gpl 16:28 sapier of course you could git blame every file and track down each parts origin 16:29 RealBadAngel weird 16:29 RealBadAngel proller why cant you just share the code? 16:31 proller why me? 16:31 proller its not my code 16:32 sapier freeminer is already dead end as of minetests perspective amount of work to get code from there is most likely same as writing that code ourselfs 16:33 proller also you cant write same code if you looket at freeminer code ;) 16:33 sapier I don't even look there 16:33 sapier useless work 16:37 sapier RealBadAngel: what's state of your shader fixes are they completed yet? 16:39 sapier np 16:39 sapier oops wrong text field .. I'm gonna merge https://github.com/minetest/minetest/pull/1395 in a few minutes 17:00 ShadowNinja sfan5: That's what I asked you. ;-) I saw discussion about it but wasn't sure what the conclusion was. 17:14 LemonLake Been experimenting with html-less css stylesheets 17:14 LemonLake by that I mean using no html for the styling :p I used the mt mod recorder as a practice, hows it look? http://i.imgur.com/XpTwEGU.jpg?1 17:14 ShadowNinja http://sprunge.us/QiEd?diff <-- Comments? 17:15 LemonLake wait, I'll get a non-jpg screen 17:15 ShadowNinja Also, any objections to setting preload_item_visuals off by default? It substantially lowers load time and doesn't affect inventory opening time very much. 17:15 LemonLake Waaiiit 17:15 LemonLake Soz guys, wrong chan lol :P you kept sending msgs so it auto-put it here 17:18 ShadowNinja For reference: It takes about 3 seconds to join my server, and about 1 second to open the inventory (UI, so lots of items) without preload_item_visuals, and about 40 seconds to join with about 0.2 seonds to open the inventory with it enabled. 17:48 sapier I agree to disabling preload item visuals ... it's due for way to long ;-) 17:54 Calinou disable by default, preload only if it's useful (to not annoy players during gameplay) 17:54 Calinou when items preload, they'll often do when the player is in inventory, not fighting a mob or such 17:55 Calinou anyway I'm disappointed of the issue triaging: https://github.com/minetest/minetest/issues/1362 → this is very easy to do, so how is this low priority? I think priority should take difficulty into account 17:55 Calinou easier to do = less incentive to not do it 17:55 Calinou but thanks for adding it to milestone anyway 17:56 sapier as far as I remember I haven't added that label guess it's due to possible discussions that could arise about this ... or did you manage to get consens to this on? 17:56 sapier +e 17:57 ShadowNinja Calinou: A algorithm hasn't been decided on, and there are more important things to do. 17:58 ShadowNinja And it shouldn't be on the milestone, we want to finish 0.4.10 soon. 17:59 sapier I guess that's not gonna be a issue to delay 0.4.10 ... there are other way more critical things ;-) 18:01 Calinou anything else is better than current system 18:01 Calinou the current one is totally wrong and there are at least two tables that fix it 18:01 Calinou sapier, RealBadAngel said he tested it, but didn't comment anything in particular 18:03 ShadowNinja Why is this on the milestone? The system has received many changes so the bug may no longer exist: https://github.com/minetest/minetest/issues/793 18:04 sapier droping it is a way to fix it for next milestone too 18:06 ShadowNinja No more milestone issues unless they're important bugs. 18:06 sapier https://github.com/minetest/minetest/pull/1343 is this worth to be added to 0.4.10 or shall it wait for next version? 18:06 sapier android port is important for 0.4.10 18:06 sapier it's time to release it 18:07 ShadowNinja sapier: Definitely for later since the formspecs should be redone. 18:07 Calinou keep in mind, redoing formspec likely breaks all mods 18:07 sapier I can't hear that damn excuse any longer 18:07 ShadowNinja Calinou: Not with a converter. 18:08 Calinou then redo them; I'm around to fix my mods anyway 18:08 Calinou but not everyone is 18:08 ShadowNinja We need a solid specification first though. Should I create an issue? 18:08 sapier next one claiming improving formspec isn't an option without providing a replacement is gonna pay me a drink 18:09 sapier non-alcoholic ... otherwise I'll die from alcohol 18:10 sapier yet the fstk ingame addons can wait till after 0.4.10 that's not the topic here 18:11 sapier #1359 #1370 and the final android port pull are crucial for android porting please test them in order to get them merged 18:12 sapier ShadowNinja: have a look if there isn't already something like "formspec must die" 18:13 sapier imho we don't need a spec ... the only sane way is using a existing gui description language 18:16 sapier https://github.com/sapier/minetest/commit/8bb87b57fbc4632103b03da36f606dc83d505f5e hmmmm can you have a look at this one? for what I found out merging those two flags doesn't do any harm in typical case and will only add some performance overhead in quite uncommon silly cases 18:29 sapier https://forum.minetest.net/viewtopic.php?pid=138276 can some forum mod please cleanup the non declaration posts in there? 18:31 ShadowNinja https://github.com/minetest/minetest/issues/1399 18:34 sapier already seen and commented ;-P 18:41 ShadowNinja sapier: Moved a bunch to trash. 18:50 ShadowNinja sapier: https://github.com/minetest/minetest/issues/1398 should be a pull request. 18:52 sapier will be done once it's usefull 18:52 sapier right now I'd have to create a new pull once a day because it's based on other pulls 18:53 sapier and rebasing fails terribly for that one 18:53 sapier I'll add a link to latest code 19:02 ShadowNinja sapier: I can't parse this: Additionally this translation I mentioned is acually what specifies the API not the low level format as anything not translated is by definition not part of it. 19:03 ShadowNinja sapier: Also, you can obviously do the same checks, eg "if not element.text then return "text us required" end" 19:03 ShadowNinja (In C++) 19:04 sapier you can always do additional work but what's the point of replacing it if nothing improves? 19:05 RealBadAngel Sapier, before 0.4.10 i do not plan any further work on shaders 19:05 ShadowNinja I don't understand what you mean by translation. And the serialization method is an irrelivant implementation detail. 19:05 ShadowNinja sapier: What? 19:06 ShadowNinja sapier: Even adding keyed fields to the original format would be immensly helpfull, but it's still ugly, hard to read, and hard to manipulate. 19:06 RealBadAngel next step is introduce multiple rendering passes, rendering buffers and rendering to textures 19:07 sapier ShadowNinja: imho the improvement by adding tables is far from big enough to justify the work required to translate it to actual gui elements 19:07 proller goood ;) 19:08 RealBadAngel that gonna allow reflections, refractions, and all the post process effects 19:08 ShadowNinja sapier: You're already doing the same work for the current format. 19:08 sapier be carefull about rendering to textures rba there are quite a lot of android devices out there not supporting it and still some (intel) gpus having issues too 19:08 ShadowNinja My version has auto-positioning though, which requires a little more work. 19:09 sapier ShadowNinja: the main difference is I actually specify the api you suggest providing just a format not giving any api wise help to use it ... fstk does this implicit by providing the functions 19:09 ShadowNinja sapier: I don't think shaders not working on Android is a concern, due to their low power. 19:09 RealBadAngel sapier, that is definitely not intended to use with pseudo computers ;) 19:10 RealBadAngel we are talkin here about high end visual fx 19:10 ShadowNinja sapier: Ah, yes, some DOM/FSTK-like API could be added. 19:10 sapier actually I'd prefere some existing language 19:10 ShadowNinja sapier: But standard Lua functions already let you deal with this easily. 19:10 ShadowNinja like table.* 19:10 RealBadAngel i agree with SN 19:11 RealBadAngel tables are what we need 19:11 ShadowNinja And you can do indexing with standard Lua indexes instead of havilg to parse or generate it. 19:12 sapier problem with what you suggest is that it's not helpfull at all without a lot of additional documentation and code checking those tables 19:12 ShadowNinja Eg, formspec[4].tabs.list.elements[1] = "New text" 19:12 ShadowNinja s/list/List/ 19:13 sapier who does the "tabs" decoding? 19:13 sapier who checks for consitency 19:13 sapier how to tell user what's inconsistent 19:13 sapier yes that'a all solvable ... but it's a lot of work 19:14 sapier and what happens if someone wants two use two independent tab header elements in a single form for example 19:15 sapier if you use tabs you give away the implicit order the current formspecs are written thus you can't tell any longer "last one is to be the one displayed" 19:16 sapier because within a tab there ain't a order ... well unless you define any tab to not have a name but a index only 19:22 RealBadAngel insted of finding possible holes in another solution try to say why it will be far better than current one ;) 19:24 sapier there are so much holes in it it's most likely not gonna be better at all after those are fixed 19:24 sapier in worst case you can do this in a way it's even worse 19:24 sapier and not considering them at the beginning is the road to that worst case scenario 19:25 sapier lua tables give full freedom to do anything ... but don't believe you don't have to pay the price for it ... and price for freedom usually is compexity 19:27 sapier https://gist.github.com/sapier/cc2b30a55c7afe4e2572 makes ipv6 error a message only 19:28 sapier any comments? 19:30 jin_xi pls consider something like this: html/css for ui via external library. i dont think we need to invent and develop another ui system 19:31 sapier that'd be my preferred way of replacing formspec too 19:32 RealBadAngel tables are not thing to invent 19:32 jin_xi tables are tables 19:32 jin_xi we dont need tables 19:32 RealBadAngel its a natural solution 19:32 RealBadAngel we do have them 19:32 sapier tables aren't the solution just the raw dataformat ... they're about 10% of your suggestion 19:32 RealBadAngel easy to handle in lua, no need to add here anythin 19:33 sapier I guess I could talk to a wall too 19:33 jin_xi its going to be another implicit- and underspecified language 19:33 jin_xi and end up like common lisp once fixed 19:40 sapier https://gist.github.com/sapier/ac00d80a1fc8b00f2c3d shows why tables are not a api 19:41 sapier all of those variants will be capable to specify a gui but usually only one of them will be translatable by our parser 19:43 RealBadAngel its like you have cut formspec string defining a button and stick it again in random order and complain why it is so fucked up ;) 19:43 RealBadAngel ofc we will need some rules 19:44 sapier come one don't you wanna understand that THOSE RULES are same crazy shit as formspecs are? 19:44 sapier and those rules are the actual api not the tables 19:45 sapier it'd be no big deal to translate formspec to tables because formspecs are way more specific 19:45 RealBadAngel but theres one thing talkin pro tables, theyre native in lua 19:45 RealBadAngel so manipulating them is extremaly easy 19:46 RealBadAngel when we will introduce something like html/css we will end up with difficulty manipulating the elements 19:46 sapier formspec is almost a typ0 grammar 19:47 sapier not exactly true since tables have been added 19:47 RealBadAngel formspec is just playing with lotsa glue 19:47 sapier your tables will require even more glue 19:47 RealBadAngel and then theres no option to edit/change, just do it again from scratch 19:47 RealBadAngel no 19:47 sapier of course 19:48 sapier because your table still may break on any edit 19:48 RealBadAngel break because of what? 19:48 sapier pos = nil 19:48 sapier boom broken 19:49 RealBadAngel why broken? 19:49 sapier tables give way more flexibility ... but at cost of way more complexity too 19:49 sapier because a button with pos nil results in broken formspec 19:49 sapier -formspec + gui 19:50 RealBadAngel so parser just pisses on such element and skips it 19:50 sapier and it's not the tables defining the gui but those rules 19:50 RealBadAngel i dont see any problems there 19:50 sapier then don't do it for a button but for a tab 19:51 sapier I know if someone doesn't want to see the problem he wont see it ... it's always been same 19:51 jin_xi the plan is to write a parser for lua tables? 19:51 sapier but don't expect a full featured type1, maybe even type2 grammar to be less parsable then a almost type0 one 19:51 RealBadAngel "parser" rather 19:52 RealBadAngel iterate over table elements and fill in existing structures on c++ side 19:52 sapier actually meant to be less effort to parse 19:52 sapier iterate a tree? 19:52 sapier level by level? branch by branch? 19:53 sapier we've got lua you can do a prototype implementation there and see how much work it is 19:54 RealBadAngel much less than with any other solution 19:54 sapier I use tables in mobf so don't believe I haven't thought about it for fstk ... I decided not to use them because of being not specific enough and thus requireing way to much docs and check functions 19:57 sapier still you're free to write a prototype in lua, straight forward implementation without checks and error handling will be a matter of days ... later one will require weeks 19:59 sapier especially if you don't wanna have a use look for ages to find a missing parameter 20:00 sapier e.g. if he did write tooItip instead of tooltip 20:00 sfan5 relevant: something I made last year http://sprunge.us/XMhi?lua 20:01 sapier looks almost like the prototype I mentioned, a year ago? and still no error handling ? ;-) 20:02 sfan5 there is error handling 20:03 sapier how is it supposed to work? 20:03 sfan5 tables goes in, formspec string comes out 20:03 sapier oh I see if it's not known it's an error 20:04 sapier does it handle missing parameters too? 20:04 sfan5 sure 20:04 sfan5 (it should) 20:05 sapier and how are optionals handled? 20:06 sfan5 see lines 93-107 20:07 sapier different lookup tables 20:08 sapier ok so what'd be the difference to current formspecs ... modifying a form would be more easy 20:08 sapier as your top level table is indexed it's most likely keep it's order too 20:09 proller sapier, you need make your fork with keeping strings in formspec 20:09 proller sapiertest ! 20:10 sapier I don't wanna keep formspecs if there's a reasonable good replacement but I don't wanna have formspecs replaced by formspecs2 20:11 sapier sfan5 this most likely will improve writing of tables too ... but wont help for tabs 20:13 sapier maybe we should first find out what a formspec replacement shall be capable to do 20:19 RealBadAngel easy of modyfing an element, removing it etc 20:19 RealBadAngel without need to rewrite whole form 20:19 sapier https://gist.github.com/sapier/8648a89011bfd0897f76 any additional requirements? 20:20 RealBadAngel "detectable" is not needed 20:20 RealBadAngel table can hold extra mods data for example 20:20 sapier no 20:20 sapier if you want this add some addon_params field 20:21 RealBadAngel check for predefined fields 20:21 RealBadAngel if type == "button" etc 20:21 RealBadAngel this way parameters are flexible 20:21 RealBadAngel you check if theyre provided or nil 20:21 sapier additional data has to be separated from form data if you don't do it you never can tell if it's correct or not 20:22 sapier that's implementation detail rba 20:22 RealBadAngel but explainin your questions 20:22 RealBadAngel table from its definition can hold anything 20:22 sapier a table is a format not an api 20:23 RealBadAngel its up to table reader what it looks for 20:23 sapier lets do formspecs in bytes 20:23 sapier a byte can hold anything 20:23 sapier absolutely free to read and modify 20:24 sapier ... sorry forgot to enable the "ironic" tag ;-) 20:24 RealBadAngel i got it ;) 20:25 RealBadAngel btw, can i push this wordwrap and height thing? 20:25 sapier it's still two true/false changes and two height changes? 20:25 sapier or did you have to change other things too? 20:26 RealBadAngel two true->false, one ascender->height 20:26 sapier thought it's been two ascender-> height? 20:26 RealBadAngel second was not so right, caused text in checkboxes not being in line with boxes 20:27 RealBadAngel but a little bit lower 20:28 sapier that's bad 20:28 RealBadAngel not really 20:29 RealBadAngel second was really meant as height of a character not line 20:29 sapier I see 20:29 sapier I already tested those 3 changes yesterday and it did look fine for me 20:29 RealBadAngel so pushing 20:32 sapier https://gist.github.com/sapier/8648a89011bfd0897f76 any objections? if not I'm gonna push it next 20:32 sapier fixes two issues 20:36 RealBadAngel ahem, check it ;) 20:36 proller sapier, your fork must have xml formspec 20:36 RealBadAngel i mean the link ;) 20:37 RealBadAngel unless youre going to push now formspecs replacement 20:37 sapier lol ... I meant this one: https://gist.github.com/sapier/cc2b30a55c7afe4e2572 20:37 RealBadAngel damn youre fast ;) 20:37 sapier I know I'm fast but not that fast ;-) 20:40 RealBadAngel idk what above is for ;) 20:40 RealBadAngel btw, since tooltips are in, we have lotsa work to do now 20:41 sapier right now if you run minetest (debug build) on a machine without ipv6 support it immediatly exits with a very cryptic message not telling you anything about why it exited 20:41 sapier I don't have work, you added it you're gonna add the descriptions ;-P 20:42 RealBadAngel i do have my own rdy 20:42 RealBadAngel but all of them will need translations too :) 20:43 sapier guess what ... you're gonna have a lot of work to do to get your inventory ready for android 20:43 sapier it's just tooo big 20:44 sapier ok I'm pushing that ipv6 fix now 20:47 sapier does anyone know how to fix this https://github.com/minetest/minetest/issues/1389 ?