Time Nick Message 00:00 est31 if pressure is too high, a merge appears 00:00 est31 we see that at the example of cyanogen 00:00 est31 or, more popular, libreoffice 00:01 hmmmm maybe this will solve itself 00:01 hmmmm maybe. 00:01 hmmmm in the meantime i'll be OK since i use freebsd 00:01 exio4 est31: merge? you mean fork? 00:02 est31 yes 00:02 est31 ofc 02:31 hmmmm https://github.com/kwolekr/minetest/commit/6c5a6967ecc162befc44f1adf15bcc9d11658ff7 02:31 hmmmm yo 02:32 Zeno` yo 02:34 Zeno` what's been happening? 02:34 hmmmm I screwed up transparency for a couple things 02:34 hmmmm moved ObjDefManager out to its own file 02:34 Zeno` :D 02:34 hmmmm talked about systemd 02:34 hmmmm talked shit about systemd 02:35 Zeno` meh, not interested in sys/systemd debates hehe 02:36 Zeno` I use sysvi anywyay 03:02 est31 interesting, very interesting. 03:03 est31 2709 is indeed a crash 03:04 est31 man this stuff is weird 03:05 est31 likely some memory corruption going on 03:05 hmmmm memory breakpoints 03:05 est31 l_craft.cpp:408 03:05 hmmmm that helped me find the bug where I held onto a const char * of a std::string that fell out of scope 03:06 hmmmm well at least you can readily reproduce it 03:06 est31 yup 03:06 est31 its really something going on there 03:06 est31 I mean 03:07 est31 if you add default to depends.txt there is a bug, if you remove it, there is none 03:07 Zeno` it's all in the hand 03:07 Zeno` hand is "" 03:07 est31 lol 03:12 est31 man, I can think of dozens of ways that thing can go wrong 03:13 est31 at least I have answered a couple of my questions: why only when default is enabled? because before it isn't loaded 03:13 est31 why a crash? strtok_r returns a null pointer if no tokens can be found 03:14 hmmmm oh.. 03:14 hmmmm you're not using mingw by any chance, are you...? 03:16 est31 me? 03:17 est31 lol no 03:17 est31 never getting close to win anymore 03:18 hmmmm oh okay. 03:18 hmmmm on MinGW, mystrtok_r() is used for strtok_r() which i wrote 03:18 hmmmm there haven't been any known problems with it, but I never trust my own code 100% 03:19 hmmmm and I did test it quite a bit 03:19 est31 I'll replace that with more proper code 03:24 hmmmm ?? replace mystrtok_r?? 03:24 ShadowBot hmmmm: Error: The command "replace" is available in the Format and Topic plugins. Please specify the plugin whose command you wish to call by using its name as a command before "replace". 03:24 hmmmm what's wrong with it 03:25 est31 it can fail for these cases for example 03:25 hmmmm which cases 03:25 est31 when you call it with "" 03:26 hmmmm strtok("", "asdf", &last) is supposed to return NULL, right? 03:27 est31 yes 03:27 hmmmm how does mystrtok_r() not return NULL in that case? 03:28 est31 it does return NULL 03:28 est31 and its not your fault 03:28 hmmmm Oh 03:28 est31 I mean its not a difference in behaviour between mystrtok_r() and the native linux /unix call 03:28 hmmmm what you're saying is that strtok_r's semantics are not appropriate for your usage 03:28 est31 its more that its used wrongly 03:28 est31 yea 03:28 hmmmm okay 03:28 hmmmm you scared me there 03:29 hmmmm I tested the crap out of that function when i wrote it 03:37 Zeno` doesn't crash for me :/ 03:37 est31 only for 0.4.12 03:37 Zeno` oh 03:40 Zeno` we need a new label then :p 03:41 Zeno` stable only or some crap 03:42 est31 I'm still interested now because I wonder what happened to it in master 03:43 Zeno` it probably got changed when the whole get_recipes was changes 03:43 Zeno` changed* 03:43 Zeno` i.e. there was a whole heap of changes gregory_cu made 03:43 est31 ah yes 03:43 est31 lemme see 03:46 est31 yea thanks to that speedup 03:46 est31 I dont think we need a label for this, perhaps just close it? 03:46 Zeno` should we revert the speedup because the commit fixes a bug as well? 03:46 est31 yes please :) 03:46 Zeno` commit should only do one thing 03:46 Zeno` lol 03:47 Zeno` well it probably should be closed if it's fixed 03:49 est31 yup done 03:51 est31 man, that commit has been thanks to me asking for a speedup, and I totally forget it 03:51 est31 not good 03:51 est31 *and* I've done a speedup inside that same file 03:54 * sofar goes to test HEAD 03:59 est31 bye 04:01 Zeno` there is more than one problem there 04:02 Zeno` but the one that is causing the segfault can be demonstrated with this test case: http://codepad.org/za2ndlPv 04:03 Zeno` err, he is gone :( 04:03 Zeno` line 12 is the culprit anyway 04:03 Zeno` well, rather if (NULL == s) is 04:08 Zeno` so strtok_r is vindicated! (for once) 04:31 hmmmm merging https://github.com/kwolekr/minetest/commit/6c5a6967ecc162befc44f1adf15bcc9d11658ff7 in 15 minutes... 04:50 Zeno` OBJDEF_MAX_ITEMS (1 << 18) ?? 04:51 hmmmm yeah the limit is 2^18 objects in an ObjDefManager 04:53 Zeno` ok 04:53 Zeno` and UID? 04:53 hmmmm a small randomizer to ensure that handles cannot be recycled 04:53 Zeno` 127 is the limit? 04:53 hmmmm those things have been there for a while :p 04:54 Zeno` hmm ok, I'll ignore them then. The rest looks straightforward and good 05:01 Zeno` https://github.com/minetest/minetest/blob/master/src/unittest/test_filepath.cpp#L90 05:01 Zeno` ^-- any volunteers to debug that unit test? 05:02 hmmmm ah you want to tackle the bug on windows 05:02 Zeno` I'd rather not but I guess someone has to :( 05:02 hmmmm I have my own list of things todo 05:02 hmmmm things that have been bugging me for a very long time 05:02 Zeno` it worries me that the unit test is not very understandable lol 05:03 Zeno` I suppose I'll do it 05:03 hmmmm yeah, the test is much more complicated than the code it tests. 05:03 Zeno` *nod* exactly the words I used 05:03 hmmmm i suggest the correct way to fix this is to blow it up and code it over again 05:07 Zeno` of course hehe 05:07 Zeno` any opportunity to blow something up should never be missed 05:08 Zeno` ok I'll do it :( 05:11 hmmmm jesus christ 05:12 hmmmm i'm not too picky about code style but all i ask is that the same person stick to the same code style 05:12 hmmmm not keep changing it around within the same function 05:12 hmmmm that just tells me they're being lazy 05:12 hmmmm rather, careless 05:12 Zeno` or have no idea what they're doing and just copy/pasting 06:24 Hunterz trying compile latest master but, get some errors... http://pastebin.com/H33uuVZb Its Linux x86_64. 06:33 hmmmm ahh 06:34 hmmmm looks like glibc doesn't include cstring when string is included 06:40 hmmmm try again now, hunterz 06:41 Hunterz oki 06:48 Hunterz fixed, thanks 15:28 hmmmm https://github.com/kwolekr/minetest/commit/f17e5bc1eb24ae5aaf9ad026330c83fb4638e3c2 16:39 rubenwardy Reviews please :D #2575 16:39 ShadowBot https://github.com/minetest/minetest/issues/2575 -- Add texture overriding by rubenwardy 16:40 VanessaE oooooo 16:41 VanessaE I *like* that 16:42 VanessaE another thing I'd love to see is a way for a texture pack to provide an alternate model file for a given node -- if the node uses mesh drawtype that is. Of course at that point it would be a "resource" pack or something. 16:45 VanessaE bbl 18:19 est31 hmmmm, looks good: https://github.com/kwolekr/minetest/commit/f17e5bc1eb24ae5aaf9ad026330c83fb4638e3c2 18:19 est31 just one question 18:19 est31 perhaps better to have an additional typedef, OrderedMap 18:19 est31 or so 18:20 est31 and then replace all maps that can't be made unordered map when we add c++11 18:20 est31 or the other way round ofc 18:20 est31 I've two minor remarks to the commit 18:20 est31 in github 18:20 rubenwardy Any feedback on #2575 ? 18:20 ShadowBot https://github.com/minetest/minetest/issues/2575 -- Add texture overriding by rubenwardy 18:20 hmmmm shoot 18:21 hmmmm yeah I missed a space or two 18:22 hmmmm we'll make the OrderedMap distinction when C++11 happens though... I mean what's the purpose in doing additional work right now to 'prepare' for something when it doesn't reduce the amount of work later when you need to change it 18:22 est31 why wouldn't it reduce the amount of work? 18:22 hmmmm if I do add the OrderedMap typedef right now and change all StringMaps where ordering is necessary, what work did I change? I just moved it to right now instead of later 18:23 hmmmm and it increases the scope of the current commit 18:23 est31 ok then perhaps not 18:23 est31 was just an idea 18:23 hmmmm I feel strongly about keeping commits focused 18:24 hmmmm but it's a hard thing to do 18:24 hmmmm in any case, all those instances of StringMap don't have any reason to be ordered. I guess it's a decent side effect of having lists of things spit out in alphabetical order though 18:24 hmmmm I doubt anybody relies on this behavior 18:25 hmmmm what's your opinion on 2575? 18:25 est31 I've +1ed it already, didnt test though 18:26 hmmmm yeah I didn't test it either of course 18:26 hmmmm but how about the things that cannot be fixed after being committed, like the syntax 18:28 est31 its a simple solution to a simple problem 18:29 hmmmm in the future we may want to add explicitly +x -x etc. as face selectors 18:29 est31 ah 18:29 est31 yea perhaps thats even better 18:30 hmmmm yeah est, about trailing newlines in git 18:31 est31 ? 18:31 hmmmm a newline at the end of the file won't explicitly be displayed, but if there was no trailing newline it'd show a big red error icon about it 18:31 est31 yes 18:31 hmmmm in git diff/show as well as github 18:31 est31 I was confused by kate, because kate does show an additional line 18:32 hmmmm yeah wow 18:32 hmmmm 2575 is a clean, neat piece of code that does exactly what it's supposed to do 18:32 hmmmm i like it 18:32 est31 my thought too 18:33 est31 perhaps the logging message can be made more unique 18:33 est31 << " : Unknown node " << splitted[0] << " on line " 18:34 est31 to have grepability of error messages 18:34 hmmmm it prefixes it with the override filepath 18:36 est31 I'll merge it and improve the error messages ok? 18:39 rubenwardy Would you like me to change "nodes" to "nodename", or are you planning on doing that on merge? 18:39 est31 I'm doing that 18:39 rubenwardy Awesome 18:43 est31 ok some other issues I've seen 18:43 est31 filepath better a const string 18:44 est31 and if a line is too long then the following lines are indented by one tab, not two 18:44 est31 only exception is for loops, or other structures where it would look weird 18:44 est31 there two tabs 18:45 rubenwardy That's only there: https://github.com/minetest/minetest/pull/2575/files#diff-70868aa6d6b96c0c1623c761500d23c4R713 18:45 rubenwardy Wait. The MT codestyle says two tabs 18:45 est31 here too https://github.com/minetest/minetest/pull/2575/files#diff-70868aa6d6b96c0c1623c761500d23c4R732 18:45 rubenwardy When line breaks 18:45 rubenwardy That's two tabs, which is correct 18:46 est31 can you cite? 18:49 rubenwardy Damn, that's for if statements 18:50 est31 if its not inside the guidelines, then it perhaps should be added 18:50 rubenwardy Yeah, there's nothing about tabs on breaking lines 18:51 rubenwardy Do what you think best code style wise, doesn't matter much 18:53 rubenwardy gtg# 18:53 est31 bye 18:59 * est31 wrote conse instead of const 18:59 est31 off course that wont compile 19:00 est31 hmmmm, can I push https://github.com/est31/minetest/commit/46f06c20312a9b149a61a256711cf8b951e9e23c 19:24 hmmmm sure 19:24 hmmmm in any case I don't mind too much about extraneous indentations 19:24 hmmmm the official minetest code style: don't change your code style mid-function. don't do anything unreadable or cryptic