Time Nick Message 00:01 ShadowNinja RealBadAngel: But I don't think mesecons will, as it won't support advanced features. 00:01 ShadowNinja sapier: Does Android have operator precedence mixed up? 00:11 ShadowNinja (Because your Android tweaks branch adds extra parenthesis) 00:12 ShadowNinja sapier: You name ConnectionSend/ReceiveThread "ConnectionSend/Receive", but name EmergeThread "EmergeThread" 00:14 sapier ShadowNinja: length of thread names is limited appending the "Thread" wouldn't change anything as it'd be ignored 00:15 sapier Shadow as far as I know operators are defined by language standard what are you pointing at? 00:16 ShadowNinja sapier: Shouldn't you use a regular iterator here? http://pastebin.ubuntu.com/7082168/ (Or children.clear(), I'm not sure exactiy what it does) 00:16 ShadowNinja Either that loop will never terminate or clear() should be used. 00:17 sapier nope that loop works that way clear doesn't do a remove 00:17 sapier it's not std 00:17 ShadowNinja sapier: You're changing &x[y] to &(x[y]) 00:18 ShadowNinja Oh, isn't there a better way though? 00:18 sapier if this is the location I think about there wasn't a loop at all 00:18 sapier it's been done completely different before .... way more complicated 00:19 ShadowNinja (*children.getLast())->remove() doesn't actually remove it from the list though, it follows a pointer and calls a menber function of a entry in the vector. 00:19 sapier IRC the remove itself does remove it from it's parent 00:19 ShadowNinja There was a loop, two loops in fact. And yes it was more complicated. 00:20 ShadowNinja Better check that, that seems just about impossible to do. 00:20 sapier irrlicht types ... 00:21 sapier that code was suggested multiple times ... of course that may be wrong too but noone complained about it ... and it works way better then the previous version which resulted in use after free 00:21 sapier grabing isn't done correct in current master, depends on os memory cleanup if this fails quick, rarely or never 00:21 ShadowNinja sapier: This is significantly more complicated though: https://github.com/minetest/minetest/pull/1168/files#diff-65f34680878a6bd86f3a59ebc0c06c6dR1130 00:23 sapier well that particular line may not be required it's just been there to ensure irrlicht data is built from data NOT beeing deleted right after it 00:23 ShadowNinja Something like that would be needed with std::move in C++11 (more efficient). But that's just copied. 00:24 ShadowNinja sapier: spec hasn't been destructed yet. 00:24 sapier1 well it's gui and it's copied converted whatever multiple times back and forward another memcopy wont make a big difference 00:25 ShadowNinja sapier: And why did you move the push_back up? It will need to be down there if and when we switch to C++11 and use std::move there. 00:25 sapier1 because push_back creates a copy and I wanted to pass the c_str from the copied value not from the temporary object 00:26 ShadowNinja sapier: (--m_fields.end())-> is much more complicated than spec. though. 00:26 sapier1 yes but spec IS the temporary object ;-) 00:26 ShadowNinja And end() can be a special iterator, -- doesn't have to be valid on it. 00:26 ShadowNinja sapier: And it's still there at that call. 00:27 sapier1 as I said this is not exactly necessary it's just gonna reduce risk of memory corruption in case of strange compiler behaviour 00:27 ShadowNinja sapier: Was there a reason to change that chunk? 00:27 sapier1 ShadowNinja: we just pushed a element in so end will always point to that element unless something very very very bad happened 00:28 ShadowNinja sapier1: Any C++ compiler is guaranteed to work with the old way. 00:29 sapier1 As I said I didn't want to pass a pointer to a temporary to some function I can't control ... irrlicht guys said they do a copy of that data but I'm not exactly sure about this 00:29 ShadowNinja sapier1: I mean end() could be equivalent to 2^32-1 or similar. AFAIK the C++ standard doesn't guarantee that end() is the element after the size()'th element. 00:30 sapier1 end() is bidirectional iterator it can't be something else 00:30 ShadowNinja sapier1: The data MUST be valid the old way. If you move the puch_back up and are using C++11 with std::move then it might not be valid. 00:31 sapier1 we're not using c++11 so we don't need to discuss about this now 00:31 ShadowNinja There's simply no reason to change it. The old way is simpler and cleaner, and the new way provides no advantages. (and has a (very) slight performance hit) 00:32 sapier1 I don't really care about that particular change unless I'm gonna get use after free error again 00:32 sapier1 android is quite picky about memory handling and causes a lot more of trouble on invalid/risky use of temporarys 00:34 sapier1 those crapy irrlicht types stop valgrind from detecting the errors so it's hard to track them down 00:34 ShadowNinja That use isn't risky. The FieldSpec won't be freed until you call ->free() or the like or you exit the block. 00:34 sapier1 fieldspec will be called immediatly on exit of this scope 00:34 sapier1 -called + freed 00:34 sapier1 it's a stack variable 00:35 ShadowNinja https://github.com/minetest/minetest/pull/1168/files#diff-65f34680878a6bd86f3a59ebc0c06c6dR1606 this should use an iterator. 00:35 ShadowNinja Exactly. 00:36 sapier1 wouldn't match rest of code in this location and I wont change all of it 00:36 sapier1 and stack variables are deleted as soon as scope is left so the c_str() would be invalid immediatly 00:37 ShadowNinja sapier1: Yep. 00:37 ShadowNinja But you're still in the same scope. 00:37 sapier1 irrlicht shouldn't store it anyway ... but I don't trust them ;-) 00:37 sapier1 no I'm not 00:38 sapier1 that function is exited very very quick 00:38 ShadowNinja https://github.com/minetest/minetest/pull/1168/files#diff-7ec3c68a81efff79b6ca22ac1f1eabbaR1477 this is inconsistent. 00:39 sapier1 did you hear about "goto fail" ? ;-) 00:39 sapier1 oh the if above ... didn't see it 00:40 ShadowNinja https://github.com/minetest/minetest/pull/1168/files#diff-857580bc1b624f434924371da1390f86R40 this *should* be 00:40 ShadowNinja sapier1: I see no goto there. 00:41 ShadowNinja The rest seems good. 00:41 sapier1 apples bug wouldn't have caused problems if they did use {} ;-) 00:41 * ShadowNinja always uses brackets. 00:41 sapier1 isdigit is a irrlicht function too without std:: it's gonna fail 00:42 sapier1 and as far as I know the short non .h names are already deprecated in c++ again 00:42 ShadowNinja sapier1: Hmmm, that sounds like something irrlicht should change. Do you have references for the .h re-addition? 00:43 sapier1 but even if they aren't isdigit is defined in ctype.h 00:43 ShadowNinja (Not that we shouldn't work arround it now) 00:43 sapier1 it's not a workaround it's just more precise so this isn't wrong anyway 00:44 ShadowNinja Yes. It seems fine. But last I checked .h was depreciated. 00:44 sapier1 and no I don't thing irrlicht is gonna remove that function ;-) 00:44 sapier1 http://irrlicht.sourceforge.net/docu/coreutil_8h.html 00:46 sapier1 .h is used for c headers while the non .h names are used for c++ headers ... hmm as this is std it might be a c++ header true 00:46 sapier1 no not exactly its ctype ... >c< 00:49 ShadowNinja sapier1: The C++ header is , the C header is (double 'c's) 00:49 ShadowNinja ctype = Character Type. cctype = C Character Type. 00:50 sapier1 yes realized this to so this is definitively a c-header 00:51 sapier1 I don't see any use in renaming headers just for sake of renaming them ... the header is ctype.h just because c++ can read it using cctype too is no reason to use it for me 00:52 sapier1 that's one of those changes in old c++ which made c++ unreadable 00:53 ShadowNinja .h headers have been depreciated for future optimization. 00:53 sapier1 where? 00:55 sapier1 ok found the reference ... sometimes I wonder who writes references like that ... guess someone wanted to do something different again 00:56 sapier1 still they'll never abolish those headers and for sake of readybility I suggest keeping .h naming for c headers 01:02 sapier1 guess this should be another point in coding style ... unless you want to replace ALL c headers ... including stdint stdio ... 03:55 hmmmm [08:35 PM] https://github.com/minetest/minetest/pull/1168/files#diff-65f34680878a6bd86f3a59ebc0c06c6dR1606 this should use an iterator. 03:55 hmmmm instead of critquing minor style preferences of other peoples' pull requests, why don't you code something of your own? 03:55 hmmmm do you really have that much free time? 03:56 hmmmm and besides, I discourage the use of C++ specific things 03:56 hmmmm no, STOP doing things the "C++ way" 03:56 hmmmm you say this "should" be an iterator - should? why? 03:56 hmmmm what if it's not? 12:51 celeron55 00:56:56 <+ShadowNinja> Hmmm, getBlockAsInteger returns a long long, aka s64, shouldn't it be u64? 12:51 celeron55 what 12:52 celeron55 i'm not going to bother to even check, but if you seriously changed it like that: congratulations, you just broke everyone's worlds 12:53 celeron55 so i checked 12:53 celeron55 it looks like that; hopefully nobody pulls a new version until that is fixed 12:54 celeron55 also if somebody is on #minetest, tell everyone to not use the latest commit 12:57 celeron55 it seems like a right time to take this into actual use again: https://github.com/celeron55/minetest-worldtest (maybe ShadowNinja is sorry enough to bother updating it to the current version) 12:58 celeron55 ... 12:58 celeron55 it seems like it actually wasn't broken 12:59 celeron55 well good luck then 13:00 sfan5 database-dummy does it wrong then 13:00 sfan5 it uses an std::map 13:01 celeron55 any of that works on pure luck then 13:01 celeron55 as far as i know, the sqlite database has negative and positive ids 13:03 celeron55 it does; just checked 13:11 BlockMen can i push https://github.com/BlockMen/minetest/commit/03297acbf4efba9a19a7f31950cb6f2bc3c65ec8 13:12 BlockMen and https://github.com/BlockMen/minetest/commit/e4d1970abfd6206aa8780cb6cafd2efc46a2666f 13:33 sfan5 BlockMen: both seem fine to mege 13:33 sfan5 merge* 13:35 BlockMen sfan5, good. i gonna push in a few minutes then 13:51 BlockMen done 13:59 iqualfragile /home/minetest/minetest/src/util/serialize.cpp:432:25: warning: format ‘%llu’ expects argument of type ‘long long unsigned int’, but argument 4 has type ‘u64 {aka long unsigned int}’ [-Wformat] 13:59 iqualfragile &similar others 14:23 ShadowNinja celeron55: It shouldn't actually cause any issues because 16+16+16=48 bits of the 64 bit bit number. And I didn't change the types of anything, just their names. So if this has caused issues it's been arround for a while. 14:23 ShadowNinja -bit 14:28 ShadowNinja iqualfragile: Hmmm... 14:29 ShadowNinja iqualfragile: How do you reproduce it? 14:29 iqualfragile ühm… just compile? 14:29 ShadowNinja iqualfragile: I don't get that warning. 14:31 ShadowNinja iqualfragile: Try this: http://ix.io/b4m 14:33 ShadowNinja %llu expects long long unsigned, which is usually 64-bit. But I suppose your system has a 64-bit long unsigned. 14:34 ShadowNinja That warning should be harmless, although a long long unsigned cast would remove it. 14:34 celeron55 you shouldn't use the regular %things for fixed-length variables like u64 14:34 celeron55 there are constants in the standard library for those cases 14:35 celeron55 http://stackoverflow.com/questions/6299083/cross-platform-printing-of-64-bit-integers-with-printf 14:35 celeron55 you may understand based on this why the code prefers C++ streams instead of format strings 14:53 ShadowNinja How's this? http://ix.io/b4r 15:06 sfan5 is that serializeStructToString even used anywhere? 15:07 ShadowNinja sfan5: The Settings writer I imagine. 15:36 troller https://play.google.com/store/apps/details?id=funny.games.worldcraft 16:04 ShadowNinja Seems gettext doesn't like C++11... http://pastebin.ubuntu.com/7085366/ 16:05 sfan5 wat 16:06 sfan5 did you make sure to put extern 'C' { somewhere? 16:06 sfan5 (or rather did you make sure it is there) 16:06 ShadowNinja I tried that. And the header already does that via __BEGIN_DECLS and __END_DECLS. 17:36 celeron55 oh god lol 17:36 celeron55 i'm reading the reviews on funny.games.worldcraft 17:37 celeron55 this is top humor 17:39 sapier I like this one http://www.heise.de/download/special-die-lustigsten-fehlermeldungen-aller-zeiten-161398.html 17:39 troller any reviews on play from stupid 12-years old girls 17:40 troller sapier, do not change spaces and code in one commit!! 17:41 Jordach celeron55, links :P 17:44 ShadowNinja Jordach: http://pastebin.ubuntu.com/7085949/ :-D 17:44 celeron55 that one was probably the best 17:44 Jordach where's \n ShadowNinja 17:45 ShadowNinja Jordach: That's one of the issues. 17:45 ShadowNinja (But comments may not support them) 17:45 Jordach celeron55, who needs Comedy Central when Failstarz are doing it for us 17:45 celeron55 that pastebin sucks for non-wrapped text; this works well for those http://pastebin.com/YUMuyKyB 17:46 ShadowNinja pastebin.com is evil. 17:46 Jordach lies 17:46 celeron55 yes, but it works 17:47 ShadowNinja Ads, Spamfilters, Captcha, Adds whitespace, Slow, Ugly, No comment/fork/annotate, Breaks copy/paste, Blocked for some people, etc. See !pastebin.com 17:47 celeron55 but it wraps .txt pastes properly 17:48 ShadowNinja I consider not wraping a feature. But in that particular case it's helpfull. 17:51 sapier1 did anyone see novatux recently? 17:53 VanessaE sapier1: sure, he shows up in #mintest-technic frequently 17:53 VanessaE +e 17:54 VanessaE in fact he just signed off of there 30 mins ago 17:54 VanessaE er almost an hour actually 17:54 sapier1 hmm guess I need to find out myself how to make targeting by touch work again after the camera offset changes 17:54 VanessaE hell he was just in HERE too 17:54 VanessaE "nore" 17:56 sapier1 why can't people just use one name? :-) 17:56 VanessaE you're one to talk ;-) 17:56 VanessaE or is that -1 to talk. 17:57 sapier1 ok ok :-) 17:57 sapier1 but it's just a suffix ;-) 18:48 ShadowNinja We've been working on switching the forum to phpBB. xyz has a test forum set up at phpbb.minetest.net. Are there any comments on this switch? If there are no serious issues we plan to switch to it soon. 18:49 celeron55 it still doesn't have a proper distinct style 18:50 ShadowNinja celeron55: Yes, we've been waiting on that for a while. 18:51 ShadowNinja celeron55: I found this, which Redcrab uses: http://demo.phpbb3styles.net/MinecraftBB+-+Dark 18:51 VanessaE ew. 19:01 BlockMen oh shit, the layout. 19:01 BlockMen sorry guys, completly forgot that :( 19:07 rubenwardy I prefer punbb to phpbb, IMO/ 19:07 rubenwardy However, that is primarily because of the style 19:08 sfan5 rubenwardy: we should switch to fluxbb then 19:09 PenguinDad minetest-classic uses fluxbb for their forums IIRC 19:10 ShadowNinja PunBB is O.K. user-side, but phpBB is much better mod/admin-side. 19:10 rubenwardy If you gave phpbb the exact same style as in design, I would be happy 19:11 ShadowNinja BlockMen: ^ 19:11 ShadowNinja :-) 19:11 ShadowNinja It could be made even better though, I think. 19:11 rubenwardy Does phpbb have a special thing for mobiles (/ cell phones, mobile phones) 19:11 rubenwardy I agree 19:12 BlockMen so new style = old style? 19:12 ShadowNinja thats how far the theme is before i forgot it http://i.imgur.com/sriY5sw.png 19:12 rubenwardy new style = feature of old style 19:12 ShadowNinja I prefer that, it matches the rest of the site much better. 19:12 rubenwardy For example, the stickies do not sand out 19:13 rubenwardy ShadowNinja: nice 19:13 rubenwardy However, what if the website changes again? 19:17 BlockMen the person that changes the website has to change the forum theme too then 19:18 VanessaE ...as soon as you finish making it ;) 19:18 VanessaE personally, without the default dirt border/background, I think it'll be too jarring 19:19 VanessaE people won't recognize it as the official forum 19:19 VanessaE there will be a LOT of "wtf?" 19:19 VanessaE that and the blue elements on the page. the rest is fungible 19:19 rubenwardy You could give a week of warning. Display messages like "the forum will be changing platforms in a week. What does this mean?" 19:20 rubenwardy "what does this mean?" being a link to a forum post 19:22 BlockMen VanessaE, i agree that it looks kinda clean (too clean) 19:22 VanessaE *nod* 19:22 BlockMen but first aim was to make a similar theme 19:27 ShadowNinja rubenwardy: A news post is planned. 19:44 celeron55 sfan5: fluxbb is shit 19:45 celeron55 it doesn't even have a proper extension system 19:47 celeron55 i don't think the forum should look similar to the "landing website" 19:47 BlockMen celeron55, you dont like corporate design? 19:48 BlockMen and landing website sounds wrong... 19:48 celeron55 if not for anything else, then only because it's a forum and needs a wider width and smaller header and so on to be usable 19:48 celeron55 corporate design? sounds boring 19:49 sfan5 celeron55: I haven't looked into fluxbb that much, if it has no extension system it won't be considered 19:50 celeron55 sfan5: also it stores passwords without salts and that cannot even be modified practically at all 19:50 sfan5 what? 19:50 sfan5 that is security 101 19:50 celeron55 it's modding system consists of their own patch format (not even regular diffs) and a thing that applies them and backups things 19:50 celeron55 its* 19:51 sfan5 nobody sane would not salt hashes 19:51 BlockMen boring > hippie (everything other design) 19:51 sfan5 I guess FluxBB isn't an option then 19:52 celeron55 i don't understand why some people consider fluxbb to be any kind of imporovement on punbb 19:52 celeron55 punbb is like fluxbb except that it does more things right 19:52 ShadowNinja ...and is dead. 19:53 celeron55 fluxbb 2 might be usable, but they seem to be a long way until they have anything releaseable 20:01 celeron55 (and i would doubt their competence anyway) 20:30 RealBadAngel anybody tested #1117 since yesterday? 20:30 ShadowBot https://github.com/minetest/minetest/issues/1117 -- Normal maps generation on the fly. by RealBadAngel 20:39 ShadowNinja RealBadAngel: These ifdefs can be combined: https://github.com/minetest/minetest/pull/1117/files#diff-389b43527c3443e3887d3e56c07ee6e8R41 20:42 ShadowNinja RealBadAngel: elses should be "} else {". 20:43 ShadowNinja Space before {, and after } if there is something directly after it (like else or while). 20:46 RealBadAngel ofc they can be combined, idk why i left it that way 20:46 RealBadAngel i could swear i already merged them 20:47 RealBadAngel huh, its already done, but not pushed :) 20:56 RealBadAngel ShadowNinja, where you have found such elses? 21:01 RealBadAngel nvm, found it 21:01 ShadowNinja In the C++ code. 21:01 artur99 hi 21:02 ShadowNinja Hello artur99. 21:03 ShadowNinja What do you have to discuss? 21:07 RealBadAngel ShadowNinja, apart from style, have you tried it? 21:07 ShadowNinja RealBadAngel: Nope. 21:09 ShadowNinja I'll push this soon if there are no objections. http://ix.io/b4r 21:11 artur99 hey, what is ix.io ? 21:12 ShadowNinja ix: command line pastebin. -- ix.io 21:13 PilzAdam ShadowNinja, you could append /diff to the url 21:13 ShadowNinja What do you think about moving the C++ mapper under the Minetest organization? 21:13 ShadowNinja PilzAdam: Hmmm, didn't know about that. 21:26 RealBadAngel so, any other objections on #1117? 21:26 ShadowBot https://github.com/minetest/minetest/issues/1117 -- Normal maps generation on the fly. by RealBadAngel 21:27 ShadowNinja PilzAdam, RealBadAngel: Comments on the patch or mapper? 21:28 ShadowNinja celeron55, sfan5: ^ 21:29 sfan5 fine as long as it works as expected 21:29 ShadowNinja sfan5: And the mapper? 21:29 sfan5 that seems like a good idea 21:30 * sfan5 goes to sleep now 21:31 RealBadAngel im ok with mapper being included 22:27 sapier http://animalsmod.comuf.com/downloads/Minetest-debug.apk added sound support 22:35 iqualfragile what about #958? 22:35 ShadowBot https://github.com/minetest/minetest/issues/958 -- Add Player list, viewable by holding TAB key by sfan5 22:36 sapier as far as I remember there's a unsettled issue about how this is supposed to behave 22:37 iqualfragile options would be? 22:38 sapier as far as I know the tab key only shows the users around player but not all online players 22:38 sapier at least if server is configured this wa 22:38 sapier y 22:39 sapier I don't think the playerlist is usefull if you can't rely on it 22:40 sapier we should find a better way to support it maybe by some sort of action key events 22:41 sapier this way other mods could use those keys to do different things with it too ... and mods have access to full player list 22:47 VanessaE generic key events would be most welcome 22:53 sapier ~ tell nore I need a little bit of help to get touch digging work with those offset changes again, for some reason it doesn't work any more. guess it's a minor issue but I don't see it. 22:53 ShadowBot sapier: O.K. 23:00 sapier guys it's time to start discussion about merging android ports to master, mine is feature complete now. It may lack some specific features of other branches but I don't know of any major differences. The one and most critical thing is controls. We need to decide what way to go. 23:01 RealBadAngel sapier, have you tried Discovery on android? 23:01 RealBadAngel it has imho cool controls 23:02 sapier I had a look for their screenshots but they don't show a lot of their conrtols ... And my tablet doesn't have play store access so I can't try live :-/ 23:02 RealBadAngel lemme look for apk then 23:02 sapier everyone has cool controls all are better for some thing and less good for something else. That's why we need to make a decision 23:03 RealBadAngel its game similar to ours 23:03 sapier yes and even our ports have at least 3 slightly different controls ;) 23:04 RealBadAngel just try it 23:04 RealBadAngel it is way better than separate keys for movement 23:04 sapier 22 mb .. I wonder what they put in there 23:05 RealBadAngel no idea, ive tried Lite version from store 23:05 sapier btw ;-) it'd have been good to tell about it PRIOR implementing everything ;-) 23:06 RealBadAngel i saw that game a few days ago 23:07 RealBadAngel and you will admit that with it movement is no longer pain in the ass 23:07 sapier I don't think movement is a problem with my controls either ;-) 23:09 sapier hmm my mouse got frozen again ... linux is starting to get as buggy as windows was some years ago 23:10 sapier how to jump? 23:11 RealBadAngel theres no jump i think 23:13 sapier hmm moving forward backward feels almost exactly as with my style of control 23:13 sapier main difference is the variable movement speed 23:15 sapier the view feels interesting ... yet a little bit limited 23:17 sapier good night ;-)