Time Nick Message 03:31 Kacey hi all 05:30 Sylvertech Hey what are you thoughts about the Oculus Rift? 05:34 hmmmm no idea, but now you've got me wondering if it's minetest core development-related 05:38 Sylvertech Yes it is. 05:39 Sylvertech Related, I mean. 05:39 Sylvertech What I had in mind was rather simple. 05:39 Sylvertech If a person's right hand is freed from the mouse and can move the camera with his head, 05:39 Sylvertech then he can have it on the keyboard along the other one. 05:40 Sylvertech The left hand would still be used for WASD movement and SPACE jumping and such, 05:40 Sylvertech and the right hand can be used for dual wielding items. 05:41 Sylvertech " K ", " O ", " P ", and " ' ". 05:42 Sylvertech Or " L " and " : " for the two middle fingers. 05:42 hmmmm sounds.... interesting. 05:42 Sylvertech The control configuration really depends on one's preferences, but my point is that now one has about double the fingers to use. 05:43 Sylvertech Imagine the implications for PVP when you're juggling a sword and an axe or something. 05:43 Sylvertech Different attack speeds. 05:44 Sylvertech Anyway, if one were to also consider Mount and Blade: Warband's control scheme, with four different directions of melee combat, 05:44 Sylvertech the result would be pleasant. 05:44 Sylvertech So? 05:52 hmmmm i don't really have any opinions on that.. maybe others do 05:52 hmmmm sounds complicated though 05:54 VanessaE personally, 05:54 VanessaE I think taking your right hand off the mouse is a bad idea. 05:54 VanessaE never make the user shift hands from keyboard to mouse 06:04 Sylvertech So did anyone get what I said about the Oculus? 06:09 VanessaE my only comment is that you should never require a user to shift their hands between keyboard and mouse 06:09 VanessaE (re: your right-hand-keys idea) 06:11 Sylvertech I was thinking about a person doing away with the mouse completly. 06:11 Sylvertech If the Rift is available, that is. 06:12 Sylvertech Soem people might not have it, which is why I suggested it be a SP feature, or perhaps only on certain servers for those who can afford it. 06:12 VanessaE no way 06:13 VanessaE keyboard+mouse is the most efficient way to play this sort of game 06:13 VanessaE this isn't a turn-based RPG here 06:13 Sylvertech Ok you're right. 06:13 VanessaE it bears more similarity to a fps if anything 06:13 Sylvertech But I just like the Oculus Rift. 06:14 Sylvertech Anyway this is of secondary importance. 06:14 Sylvertech No need to fret. 06:15 Sylvertech Sorry if this might be rude, but you have heard about the Oculus Rift, right? 06:15 VanessaE nope. don't really care too much about it though. 06:15 VanessaE this channel is not for that kind of discussion 06:15 Sylvertech Sorry then. 06:16 Sylvertech Isn't this considered a game feature, however? 06:16 Sylvertech Anyway toodle-oo. 10:45 RealBadAngel https://github.com/minetest/minetest/pull/516 10:46 RealBadAngel ive changed editbox in formspec 10:47 RealBadAngel it was annoyin when inputing data or search strings that pressing enter causes to close the formspec 10:48 RealBadAngel instead it just loose focus. when nothin is focused then enter can close the formspec 10:48 PilzAdam so you have to hit enter twice in signs? 10:48 RealBadAngel yes 10:48 RealBadAngel or use button 16:15 realbadangel_ hi all 16:17 RealBadAngel hmmmm, posted a fix for editbox in formspecs 16:18 RealBadAngel https://github.com/minetest/minetest/pull/516 17:11 hmmmm fix for editbox in formspecs? 17:11 hmmmm what was broken? 17:13 VanessaE hmmmm: in some cases, pressing enter after typing in some kind of input should merely cause the input line to lose focus, rather than actually close the window. Unified Inventory's search function, for example. 17:14 VanessaE can't say I agree about how it changes sign formspecs though 17:14 VanessaE (sorry RBA) 17:14 PilzAdam I agree to the change 17:14 VanessaE s/window/formspec/ 17:14 PilzAdam pressing enter twice in signs isnt very hard 17:14 VanessaE No, not at all 17:15 VanessaE some people will be confused by the change, but it's not like you go around editing sign text continuously. 17:15 VanessaE so that's okay if that's how it has to be 17:16 hmmmm so you need to press enter twice in order to close a formspec when you had focus on a textbox? 17:16 VanessaE hmmmm: that's the plan. 17:16 hmmmm what.. why 17:16 VanessaE well 17:16 hmmmm that doesn't.. 17:17 VanessaE in a sign, it's probably less sensible than enter-to-commit-and-close, but for example in Unified Inventory, focus currently defaults to the search field. When one hits enter, the logical assumption is that this will execute a search - but instead it just closes the formspec. 17:17 hmmmm okay so i understand the problem 17:17 hmmmm but why does this have to be the change 17:18 hmmmm i assume pressing enter with the formspec in focus closes it 17:18 hmmmm while pressing enter within the textedit's focus makes it lose focus 17:18 VanessaE yes, that's the plan, but as it stands now, enter always closes the formspec. 17:18 hmmmm when you press enter while focused on the textbox, does the formspec itself receive the enter event? 17:18 VanessaE I don't think so, no. 17:19 hmmmm this was originally designed with the intention of only accepting some input from some formspec 17:20 hmmmm you don't have control over the actual code that gets executed though 17:20 hmmmm so wait, explain to me how the formspec works. how do you specify the elements, what do you get to specify exactly 17:20 hmmmm how do you interact with this all, through lua? 17:21 PilzAdam https://github.com/minetest/minetest/blob/master/doc/lua_api.txt#L686 17:21 PilzAdam you basically list all the elements in a string 17:21 hmmmm okay, that defines the interface, i knew that much already 17:21 hmmmm but what actually gets the text when you enter it 17:22 PilzAdam and when pressing a button this callback is called: https://github.com/minetest/minetest/blob/master/doc/lua_api.txt#L1515 17:22 hmmmm okay 17:22 hmmmm so then why is quitMenu() called after accepting the input? 17:22 PilzAdam s/button pressed/button pressed or hitting enter in a field 17:22 hmmmm quitMenu() should be called by the acceptInput callback 17:23 hmmmm the callback decides what to do after somebody entered text 17:23 hmmmm it only makes sense to do this, instead of just changing behavior and having people press enter ....twice.... to do whatever, for some odd reason 17:24 PilzAdam currently Lua cant close formspecs 17:24 VanessaE didn't I suggest a separate function/callback/whatever before? 17:24 hmmmm i'm sorry, i want to be an agreeable person but i can't agree with this. it's too odd and it doesn't make any sense to require the user to do such an ass backwards thing like press enter twice. they'll think THAT is a new bug 17:24 VanessaE (well, maybe to RBA in private, we talked about this issue some weeks ago, without any solution at the time) 17:24 hmmmm then why don't you add a Lua API to close formspecs? 17:25 hmmmm that would be the solution; not this 17:25 hmmmm maybe it's even easier 17:25 PilzAdam if someone does this, an on_close() callback would be nice too 17:25 hmmmm what does on_receive_fields return? 17:26 PilzAdam nothing 17:26 hmmmm well what the hell? 17:26 PilzAdam maybe return true -> close formspec 17:26 hmmmm shouldn't it return a boolean of whether or not to close the formspec 17:26 hmmmm there's your answer 17:26 hmmmm modify the api to have it return true if close, false if stay open 17:26 hmmmm any other 'solution' to this is just weird and unacceptable 17:27 hmmmm sometimes i wonder what minetest would be like if everybody just said yes to everything 17:29 VanessaE heh 17:29 RealBadAngel hmm, why dont you look at this from this point of view 17:30 RealBadAngel editboxes are unusable by now 17:30 RealBadAngel because if you press enter on any of them 17:30 VanessaE to be fair, we've seen what MT almost became when everyone said 'no'... so a few yes's isn't so bad 17:30 RealBadAngel whole formspec closes 17:31 RealBadAngel there shall be exit editbox used for signs (like exit button) 17:31 RealBadAngel and this one shall be used with signs 17:32 RealBadAngel calling fucked up editbox behaviour a feature is a nonsense 17:33 hmmmm realbadangel, if we changed on_receive_field to return whether or not it closes, this becomes a non-issue 17:33 RealBadAngel i can add another type of editbox with is_exit flag set to use in simple inputs 17:33 hmmmm just read the above solution 17:34 RealBadAngel you dont get it. i do have 10 edit boxes. pressing enter on any closes formspec 17:34 RealBadAngel let it be one of them command line of my computer 17:34 RealBadAngel shall i avoid pressing enter in computer? 17:35 hmmmm well then why would you all of them return true to close on enter? 17:35 hmmmm have all of them* 17:35 hmmmm then it would be the mod maker's fault 17:35 RealBadAngel i just want them not to close the whole formspec 17:36 RealBadAngel because theyre not supposed to be doing so 17:36 RealBadAngel pressing enter shall send stored data 17:36 RealBadAngel not close the application 17:38 hmmmm yes... and what i proposed would solve that. 17:38 hmmmm re-read it 17:41 RealBadAngel first of all i dont like enter being able to close formspec but nvm 17:41 RealBadAngel toggle I shall do it (or ESC) 17:42 RealBadAngel "so then why is quitMenu() called after accepting the input?" this is hardcoded behaviour for enter 17:43 hmmmm i was just going back and trying to make sense of this, i was wondering why it was designed like that in the first place 17:44 hmmmm the original design is stupid, yes 17:44 hmmmm we're fixing that though 17:44 RealBadAngel it was designed only for signs 17:44 VanessaE my guess is that the idea that enter should do anythign but accept-and-close simply never occurred. 17:45 RealBadAngel there are two cases for enter (if not focused on editbox) quits 17:45 RealBadAngel if focused on editbox accepts input and quits 17:46 RealBadAngel if anybody is thrilled by having to press enter twice there shall be another kind of editbox with is_exit = true flag set 17:46 RealBadAngel like for exit buttons 17:47 RealBadAngel so it will accept input and quit 17:48 RealBadAngel and such one shall be used for sign like formspecs 17:53 RealBadAngel and with my change callback function is called 17:54 RealBadAngel fields.editboxname holds then input 18:11 hmmmm what's wrong with what i said.. 18:22 RealBadAngel im lost with this lol. what do you propose exactly 18:22 RealBadAngel ? 18:27 thexyz proller: http://forum.minetest.net/viewtopic.php?pid=73621#p73621 18:27 thexyz probably related to your changes 18:29 proller celeron55 wrote about one curl flag for multithread 18:30 thexyz hm.. 18:30 thexyz 22:29:18: ACTION[ServerThread]: announcing to not.exists 18:30 thexyz *** longjmp causes uninitialized stack frame ***: ./bin/minetestserver terminated 18:30 thexyz i thought i fixed all those flags 18:31 thexyz yes, definetly 18:32 proller wow 18:32 thexyz all three curl_easy_perform(curl); are preceded with curl_easy_setopt(curl, CURLOPT_NOSIGNAL, 1); 18:34 thexyz oh crap 18:35 thexyz i forgot to rebuild minetest that time 18:35 RealBadAngel shouldn't it return a boolean of whether or not to close the formspec 18:35 RealBadAngel there's your answer 18:35 RealBadAngel modify the api to have it return true if close, false if stay open 18:36 RealBadAngel how i could do that if theres hardcoded enter->quit behaviour there?? 18:36 VanessaE hmmmm: status report on emergethread: it's still lagging by a few seconds from time to time, but the lockups seem to be pretty much history. Nice work. 18:37 thexyz oh 18:37 thexyz so lockups don't occur anymore? 18:37 VanessaE thexyz: nope. not so far as I've seen 18:37 thexyz great 18:38 VanessaE what's the unit for the 'uptime' field? minutes? 18:38 thexyz where? 18:38 RealBadAngel not the seconds? 18:38 VanessaE in the /status display 18:38 thexyz seconds 18:38 VanessaE ok. 18:39 VanessaE hm, uptime's only ~6900. musta crashed earlier *shrug*. 18:39 VanessaE still, I haven't seen emergethread hard-lock at all since that big update. 18:40 thexyz VanessaE: any logs? 18:40 VanessaE didn't look. 18:40 VanessaE probably some random mod 18:40 thexyz can you upload debug.txt somewhere? 18:42 VanessaE hm. *looks at log*.. whatever prompted the last restart didn't get logged. Oh well, it's not related to anything here. 18:42 VanessaE it musta segfaulted or something. 18:43 thexyz that's bad 18:43 VanessaE if you're interested: http://pastebin.ubuntu.com/5582992/ 18:44 thexyz bad 18:44 VanessaE I'll run it through gdb later on and grab a backtrace if it dies again. 18:50 hmmmm RealBadAngel, there's obviously no way around this without modifying the core 18:50 VanessaE ok. it's running via gdb now (and it's a debug build) 18:50 hmmmm the idea is to change the API so that it returns a boolean 18:51 hmmmm so you do if (acceptInput()) quitMenu(); 18:51 VanessaE I agree with hmmmm's idea, btw. 18:52 RealBadAngel so you want to acceptInput wait for Lua callback to return a value? 18:53 hmmmm yeah 18:53 RealBadAngel ok, i will try this way 18:58 RealBadAngel just a thought, doesnt that means need to rewrite all mods that used callbacks to make them return something? 19:00 hmmmm yes 19:04 RealBadAngel and all this because of 2 times enter needed? 19:05 hmmmm absolutely 19:05 RealBadAngel then i rather will like to see here edit box with default quit behavior 19:06 RealBadAngel like other formspec elements 19:06 RealBadAngel button_exit[,;,;;