Time Nick Message 00:42 gregorycu After I finish this perf improvement, I'm going to move onto functional improvements 00:42 gregorycu You know, the fun stuff 01:58 paramat hi hmmmm, if you have time i have 2 PRs ready for merge, #2267 (i added more explanation for you), and #2175 (please could you check the relevant game PR https://github.com/minetest/minetest_game/pull/412 to review what i have done with blob ore?). Also, some screenshots for you https://forum.minetest.net/viewtopic.php?f=3&t=11188 01:58 ShadowBot https://github.com/minetest/minetest/issues/2267 -- Mgv7: Add lower limit of zero to mountain height noise by paramat 01:58 ShadowBot https://github.com/minetest/minetest/issues/2175 -- Mgv5: Remove blobgen, add large caves. De-link terrain level from water level. by paramat 04:51 est31 Zeno`, #2301 ? 04:51 ShadowBot https://github.com/minetest/minetest/issues/2301 -- Fix crash on passing false as value in table to table.copy(t) by est31 04:56 Zeno` In C or C++ superfluous parenthesis are one on my pet peeves, and it now seems apparent that the same emotion is evoked when I see them in Lua as well :) 05:02 Zeno` Can you remove them, est31? 05:15 est31 Zeno`, in this case, they make understanding the function easier, but I can remove them if you wish. 05:16 * Zeno` looks again 05:16 Zeno` ok, leave them as is. Merging 05:17 est31 fine :) 05:23 Zeno` I suppose that should be committed to the freeze branch as well. Doing that now 05:31 Zeno` hmm 05:32 Zeno` What's the best way to cherry pick commits from another branch? 05:32 sofar git cherry-pick? 05:32 est31 ^ 05:33 Zeno` it will recognise the hash id from another branch? 05:33 sofar yah 05:33 est31 yes 05:33 Zeno` cool, thank you 05:33 est31 tried it yesterday and was impressed by it too :) 05:33 est31 did an extra git log as I didnt believe it 05:34 * Zeno` wonders why he even asked that now. How else would cherry pick be useful, lol 05:40 Zeno` I intend to add the following the the feature freeze branch: http://dpaste.com/39FZQRY Any comments? 05:41 Zeno` hmmmm? 05:42 Zeno` err 05:42 Zeno` wrong list 05:44 Zeno` This is harder to keep track of than I expected heh 05:45 Zeno` It's the right list except Fix getttext on MSVC is already there 05:49 Zeno` http://dpaste.com/2SBDZCZ 05:51 Zeno` meh, forget it. I've been looking at the wrong git repo 08:17 nrzkt &IRC 08:17 nrzkt oops :) sorry 08:30 nrzkt hmmmmm can your review #2175 please ? 08:30 ShadowBot https://github.com/minetest/minetest/issues/2175 -- Mgv5: Remove blobgen, add large caves. De-link terrain level from water level. by paramat 08:48 est31 nrzkt, #2308 08:48 ShadowBot https://github.com/minetest/minetest/issues/2308 -- Server should send other players inventories when its needed 08:50 nrzkt i see, but i can't doing it now, i need to finish the network rework before. The TCP design is finished but i need to fix some little bugs and finish to handle sessions properly 08:50 est31 ok can I do it? 08:52 nrzkt propose a patch, but it need to be rebased at when patch 2/4 will be applied because there are changes into packet writing and handlers. If the datas are same between the two packets but only the player change, adding a field for player into the original packet is better 08:53 est31 data are same 08:55 est31 then I'll start my work basing on that. 08:56 nrzkt perfect. For string serialize, please use the (u16)length + string insertion into packet 08:56 sfan5 est31: that android page should still be correct, but read what's written at the top of the page 08:56 nrzkt when patch 2/4 will be rebased you will use *pkt << string to replace it 09:02 est31 nrzkt, how can I distinguish client side? Use protocol version? 09:04 nrzkt distinguish what ? 09:04 est31 whether the data is inventory or player name + inventory 09:04 nrzkt use playername + inventory every time 09:05 nrzkt and if playername == client player name it's the client else it's others 09:05 est31 yes, but new client needs to be abled to connect to old server 09:05 nrzkt no 09:06 nrzkt with > 0.4.12 protocol will be break, totally 09:06 nrzkt no compat layer for old clients and old servers, we fix all protocol problems and change the networking layer 4 mode 09:07 nrzkt 0.4.12 is the last version which keep compat for 0.4 branch 09:08 est31 so there are other changes? what for servers that use git master? 09:08 nrzkt they shouldn't 09:08 est31 because it will cause problems for old clients and new servers too 09:08 nrzkt they must follow 0.4-stable 09:09 est31 VanessaE is doing this too. 09:09 nrzkt master is for developpers, it's the devel environment, not the stable patchfix environment, for stable, follow feature-freeze when it is, or stable-0.4 09:09 VanessaE most servers that use git follow master 09:09 VanessaE you need to create a 0.5.x branch for this, nrzkt 09:09 nrzkt no 09:09 VanessaE if you're gonna break compat, you can't do it in master, imho 09:09 nrzkt it's not the guidelines 09:10 VanessaE actually, it sorta is. the decision was made some time ago that the "break everything" release would be named 0.5.0. the idea of putting it on a separate branch hadn't been considered though 09:10 nrzkt if servers are doing this, we must found how to change this, because it's bad, nobody must follow master branch for a stable environment, everything can happen 09:10 nrzkt the stability is on stable-0.4 09:11 nrzkt it's named for that :) 09:11 VanessaE yeah but "stable 0.4" might not necessarily be all that stable either. 09:11 nrzkt on git guidelines master is the devel branch, where the developpers works actively 09:11 VanessaE see also, 0.4.11 09:11 VanessaE (which as you know is rather broken) 09:11 nrzkt it's a problem with old release process 09:11 jin_xi what happened to 0.4.12? was that not supposed to be a bugfix release to be out soon? 09:11 VanessaE jin_xi: it's supposed to be 09:11 VanessaE idk what the hold-up is though 09:12 jin_xi because 0.4.11 is not very good and will break maps and crash 09:12 nrzkt jin_xi, hmmmm must finish him but doesn't take time to do it 09:12 kilbith must finish it* 09:12 jin_xi must finish what? there are other people who can build and a release with no binaries is not unheard of in for minetest 09:13 nrzkt ofc, but i don't know what we wait... 09:13 est31 sb should post preliminary builds in forums of https://github.com/minetest/minetest/tree/freeze-0.4.12 09:14 est31 as a release candidate replacement 09:14 jin_xi or release 0.4.11.5, but last release is really nothing you'd want to leak out 09:16 * est31 waits until the "git master breaks things" debate is resolved. 09:16 nrzkt in fact i could do the linux and FreeBSD build, but i don't about windows, i need to look at this. If i could create the release engineering process on jenkins it will be simpler for all 09:17 est31 if it is decided to make master breaking, this should be communicated. 09:17 nrzkt ofc 09:17 jin_xi all im saying is that 0.4.12 was only debated as a bugfix release, bug got fixed, no release. mt still bugged for many in a major way 09:18 nrzkt i agree. 09:19 jin_xi its no fun when your fancy map gets filled with crashing nodes 09:21 nrzkt i have a problem also. Is this #2300 ? 09:21 ShadowBot https://github.com/minetest/minetest/issues/2300 -- SharedBuffer ::drop() [T = unsigned char]: Assertion (*refcount) > 0 failed. 09:22 est31 there was a mapgen bug in 0.4.11 creating weird random nodes 09:22 jin_xi lol try a lua mapgen with 0.4.11 09:24 VanessaE nrzkt: he means weird blocks like this: http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2001122015%20-%2008%3a49%3a29%20PM.png 09:24 VanessaE :) 09:24 nrzkt oh, i see 09:24 est31 is this kde 2?? 09:25 jin_xi VanessaE: try this with no mods installed, you will get furnaces and crash. 09:25 VanessaE (which sometimes contained nodes with garbage metadata that would cause mods that relied on it to crash) 09:25 VanessaE jin_xi: case in point 09:25 est31 :D 09:25 VanessaE est31: xfce with a custom theme. 09:26 nrzkt ok, please give me a mod which cause the crash 09:27 nrzkt and i'll try to reproduce the problem 09:27 VanessaE nrzkt: the problem has since been resolved. 09:27 nrzkt if it's resolved why 0.4.12 is not launched ? 09:27 VanessaE issue 2300 is something entirely different. 09:27 VanessaE idk for sure, probably because no one is able/willing to produce good windows builds 09:28 nrzkt hmmm i see 09:28 nrzkt by good you mean working ? 09:29 VanessaE generally working, doesn't crash, contains all the third-party libs players will want (sound, luajit, levbeldb, etc) 09:29 VanessaE (for example, MichaelEh can't seem to run anything newer than 0.4.10) 09:44 est31 good night 10:03 Krock daboom! using zlib 1.2.8 now 10:22 BlockMen VanessaE, wut? 10:22 BlockMen "because no one is able/willing to produce good windows builds" 10:26 Zeno` nrzkt, what is jenkins? sfan5 are you able to prepare a release? All, what's happening with 0.4.12? :) 10:26 Zeno` sfan5, prepare windows build I mean 10:28 Krock VanessaE, a wut? from my site aswell :P 10:30 nrzkt jenkins is a continuous integration system, like travis. I use it at work to build profesionnal application for our web platform (java + tomcat) and deploy the new releases 10:30 nrzkt jenkins permit to automate it if we want by passing the integration script 10:31 nrzkt look: http://jenkins.unix-experience.fr/job/minetest/ it uses my forked repo to make builds 10:31 Krock is it better than travis? 10:32 nrzkt i don't know if it's better than travis, because i don't use travis, but jenkins is very very used. FreeBSD uses it for FreeBSD -current and -stable builds and releases 10:33 Zeno` does either travis or jenkins make the builds they make available for download? 10:33 nrzkt it could be done if we have our own instance 10:33 nrzkt look here 10:33 nrzkt https://jenkins.freebsd.org/ 10:33 nrzkt the port tree and the cross compile is done by the instance :) 10:34 Zeno` If everything can be automated I imagine releases would be easier 10:34 nrzkt you can also delegate compilation to slave servers 10:34 nrzkt ofc, i could help 10:35 nrzkt i can do the freebsd integration and Linux integration on my own jenkins, i need to find if this is possible for windows with mingw because i use freebsd and it may be hard to cross compile the minetest build, but i could try 10:36 Zeno` my cross-compile environment is FUBAR heh 10:36 Zeno` I dunno what I did but it was working one day and not the next 10:37 Zeno` I guess I just don't understand the MT release process properly; e.g. who provides the builds 10:37 Zeno` and what builds are required 10:38 BlockMen Zeno`: http://dev.minetest.net/Releasing_Minetest 10:39 Zeno` So we need a build from both you and sfan5? 10:39 Zeno` (assuming everything is ready to go of course) 10:39 Zeno` ^^ BlockMen 10:39 Zeno` brb 10:41 BlockMen yes. but 0.4.11 release shows, that someone else should be able to provide msvc builds aswell (32+64bit, all 3rd party libs) 10:42 Zeno` yeah 10:42 Zeno` I guess I'm now pushing for 0.4.12 to be released because it fixes some pretty nasty bugs in 0.4.11 10:43 BlockMen just say when its ready, then i start building 10:43 Zeno` There is one outstanding issue (possible regression) but nobody seems able to track down the cause or why it happens and there is (AFAIK) only one user affected (Wayward_One) 10:43 BlockMen which one? 10:44 Zeno` Let me find it... 10:44 Zeno` https://github.com/minetest/minetest/issues/2145 10:45 Zeno` I've looked at his callgrind results and so has hmmmm. I've also produced many profiles of my own. At this point we still don't have much idea (AFAIK) and we've been looking at it for almost 30 days 10:46 Zeno` There was one commit that should have improved things, but I don't think it did 10:47 Zeno` https://github.com/minetest/minetest/commit/9a0dd47057cad18522e43d159a06c856dfdeeef4 <-- that was thought to be the main culprit 10:48 Zeno` i.e. getDisplayDensity() was closing/opening X on every call 10:50 Krock Releasing a new version is not optimal. The VERSION_EXTRA variable is not erased on a stable build, that causes wrong strings in the C(++) codes 10:54 Zeno` If that's the case the -freeze branch should just be deleted 10:54 Zeno` But I'm not 100% sure what you're saying... 10:55 BlockMen Zeno` hmm...maybe its something in _game? 10:56 BlockMen but for now i would say we should release, since he seems to be the only one 11:02 Zeno` well, here goes nothing 11:03 Zeno` pushing cherry picked commits to freeze-0.4.12 branch :3 11:04 * Zeno` hovers his finger over the enter button LOL 11:05 Zeno` done 11:05 * Zeno` waits for explosion 11:06 Zeno` hmmm nothing *seemed* to explode 11:06 Zeno` can someone check the freeze-0.4.12 commit history for me? :3 11:07 BlockMen check what exactly? 11:09 Zeno` I dunno... does it look right? lol 11:09 Zeno` building now 11:11 BlockMen if you mean if its in right timeline, no 11:11 BlockMen is the freeze branch ment to be kept? 11:11 Zeno` I don't think the timeline matters all that much (does it)... as long as all the fixes are in there 11:12 Zeno` I think celeron55 said it would be deleted (my opinion though is that it should be kept) 11:12 BlockMen when it should be kept, then it should have the same timeline IMO 11:12 BlockMen if its deleted, who cares :D 11:13 nrzkt 'yes, after release FF branch is merged into stable-0.4 and deleted 11:13 Zeno` ok, I think we should release 0.4.12 11:13 nrzkt okay 11:13 Zeno` we've done as much as possible I think 11:14 BlockMen well, then he ff branch needs to be corrected 11:14 BlockMen *the 11:14 Zeno` the branch builds fine. I guess we might need to allow a day for testing? 11:15 Zeno` I dunno 11:15 Zeno` does it need testing? 11:15 Zeno` it's kind of already been tested in master 11:16 Zeno` the part/join flood is starting to annoy me, hehe 11:17 Zeno` let's get this baby out of the way :P xD 11:20 Krock Is a Clang build error common on a change like #2310 ? 11:20 ShadowBot https://github.com/minetest/minetest/issues/2310 -- Update zlib support by SmallJoker 11:21 Zeno` Depends if travis is working or not 11:21 Krock 1/4 builds failed 11:22 BlockMen zeno why didnt you include https://github.com/minetest/minetest/commit/2630f45fdf9888038a5f0e08319dd0d4d1ce5e82 ? 11:22 BlockMen or https://github.com/minetest/minetest/commit/c7a78037809a2f9c431441f7ef91f84eb8744fff 11:22 Zeno` BlockMen, only because it was not a fix. I was led to believe that was a feature 11:23 Krock Why not take current master for the freeze? much easier 11:23 BlockMen the second has "fixes" in msg ;) 11:23 Zeno` Krock, because some features that have been added haven't had enough time for testing 11:23 BlockMen i would say those two should also go in 11:24 Zeno` c7a78037 should probably be cherry picked, yes 11:24 BlockMen Zeno`, im updating the freeze branch right now bcause of timeline 11:24 Zeno` if Update czech translation is backported is there a script that needs to be run? 11:24 Zeno` Blockmen, ok 11:25 Zeno` I don't know anything about how the translations work 11:25 Zeno` well, I know *how* they work, but I don't know if there are steps involved :) 11:26 BlockMen no special script the normal using of gettext when making a clean build 11:27 Zeno` ok, cool 11:30 BlockMen Zeno`, fine that way? https://github.com/BlockMen/minetest/commits/freeze-0.4.12 11:31 Zeno` as far as I can see... 11:32 Zeno` Krock, I'm starting the zlib build again 11:34 BlockMen should i override the freeze branch then? 11:34 Krock ^ 11:34 Zeno` BLockmen, probably yeah 11:35 BlockMen celeron55 ^^ 11:35 Zeno` good idea ;) 11:36 Zeno` Krock, passed this time 11:37 Krock hmm interesting 11:37 Krock I thought a machine acts the same way all the time and now it works the 2nd time 11:41 nrzkt http://jenkins.unix-experience.fr/job/minetest-windows/27/console 11:41 nrzkt build is working on my instance for win32 with -master branch 11:42 Zeno` is the timeline (of commits) in a branch important? 11:42 Zeno` Krock, it failed because of a download error 11:43 Krock okay 11:43 Zeno` So I guess the remote server for whatever it was downloading was down temporarily 11:45 Zeno` celeron55, I am recommending that 0.4.12 is ok to release. (pending BlockMen's re-ordering of the timeline... see above conversation) 11:46 celeron55 timeline re-ordering? 11:47 Zeno` yeah... *shrug*, I dunno 11:48 celeron55 sounds like something that shouldn't be done 11:48 * Zeno` pokes BlockMen 11:48 celeron55 (altough i have no idea what that means) 11:49 BlockMen it means that commits in freeze-branch are in same order like they were pushed 11:49 Zeno` Is that important? 11:49 PilzAdam no 11:49 Zeno` The end result is the same 11:50 celeron55 well whatever, as long as the resulting code stays the same 11:50 celeron55 and nobody force-pushes anything to master or stable-0.4 11:50 nrzkt i think the 0.4.11 time has gone. 0.4.12-freeze waiting a non incoming event and is ready for release engineering process 11:51 BlockMen Zeno`, then you can decide. you can cherrypick those 3 commits or i will forcepush the freeze branch 11:52 PilzAdam this whole freeze branch thing is already messy, no need to try to clean that up 11:52 BlockMen PilzAdam, already did :P 11:53 PilzAdam BlockMen, I mean it will get messy after you merge the second freeze branch to stable-0.4 11:53 PilzAdam the commit order will be messed up by that 11:53 BlockMen hmm, true 11:54 PilzAdam and now you already annoyed anyone who checked out the freeze branch already 11:54 BlockMen wut? i didnt push anything 11:55 celeron55 the alternative way to do an asynchronous release like this is to do the freeze in master, but continue development in a separate branch which is rebased onto master when the freeze is done 11:55 PilzAdam but Zeno` force pushed? 11:55 Zeno` I think the problem was that nobody (or very few of us) committed fixes to the freeze branch after it was created. IMO this is just a lesson and the current commit order doesn't matter much 11:55 Zeno` I didn't force push 11:55 Zeno` It was a normal push 11:55 Zeno` cherry-pick and push 11:56 PilzAdam celeron55, I think we get a merge conflict when we try to merge freeze-0.4.13 into stable-0.4, since we cherry-picked commits onto freeze.0.4.12 11:56 PilzAdam Zeno`, ah, okay then 11:56 celeron55 yes, it can cause merge conflicts, but git is quite clever about those and they won't necessarily happen if the commits weren't rebased too much related to each other 11:56 BlockMen celeron55, i think thats the better way 11:57 PilzAdam commits that should go into freeze branches should pushed there, and the merged into master (yes, with merge commits) 11:57 PilzAdam that way we get no merge conflicts 11:57 celeron55 i mean, git will often just see that "oh this is the same commit" and not bother with it 11:57 Zeno` yeah, that's what I mean. But this first experience with the freeze branch didn't turn out that way :) 11:58 celeron55 PilzAdam: umm... that's not what i meant 11:58 PilzAdam celeron55, yea, thats what _I_ meant 11:58 PilzAdam freeze branches suck anyway, we should use dev branches and freeze master 11:59 Zeno` why are we changing the subject? AFAICS the freeze branch is fine and we should be talking about release :P 11:59 Zeno` we can talk about how to refine the process after that 11:59 gregorycu_ nrzkt: The problem is my opening braces are not on the same line? 12:00 nrzkt ofc gregorycu_ 12:00 gregorycu_ Right 12:01 gregorycu This seems very important, and I'll fix this right away 12:01 Zeno` gregorycu_, the idea is that as things are changed the style should be updated; so that one day the style in the entire codebase will be consistent 12:01 Zeno` he probably cannot hear me anyway :D 12:04 Zeno` btw, all, if I stuffed up the freeze branch let me know so I can avoid it in the future (please) 12:05 Zeno` well, let us all know as well I guess 12:06 gregorycu I'll (somewhat) happily fix the style for any lines that I've modified, but I'm not going through the code and fixing this up 12:06 nrzkt no, just modify the blocks where you are working on 12:08 BlockMen Zeno`, freeze branch is still missing af7f369d28dcd1f944fa396ab6b79198866cc39f, 2630f45fdf9888038a5f0e08319dd0d4d1ce5e82, c7a78037809a2f9c431441f7ef91f84eb8744fff 12:08 Zeno` BlockMen, can you commit them or would you like me to? 12:08 BlockMen k, will do 12:16 BlockMen done. concerning release: do we publish some rc builds first 12:16 BlockMen or do we release now? 12:17 Zeno` dunno. I've never been involved with this before 12:18 Zeno` :( 12:18 Zeno` Some guidance would be great hehe 12:18 BlockMen rc builds only make sense when we do NOT release today. so question is: is there something stoping us from releasing now? 12:18 BlockMen PilzAdam? ^ 12:18 Zeno` Only that single issue that I mentioned (afaik) 12:19 Zeno` Which I don't think can be fixed in a reasonable timeframe 12:19 Zeno` Unless someone has a revelation of course :) 12:19 nrzkt RC build could be good for testing the build before public publication 12:20 Zeno` I'm fine with an RC if it can be announced and made public 12:20 kilbith the changelog is not even published... 12:20 Zeno` RC makes sense 12:22 * BlockMen is uploading rc builds and will publish on forums 12:30 BlockMen https://forum.minetest.net/viewtopic.php?f=18&t=11227 14:31 Zeno` #2309 looks like Java :( 14:31 ShadowBot https://github.com/minetest/minetest/issues/2309 -- Fix superflous shader setting updates by gregorycu 14:59 BlockMen #2314 14:59 ShadowBot https://github.com/minetest/minetest/issues/2314 -- Fix font_size by BlockMen 15:07 BlockMen _game-maintainers: what about game#429? 15:07 ShadowBot https://github.com/minetest/minetest_game/issues/429 -- Add reverse recipes for hoes v4 by MT-Modder 15:24 Krock ^ looks ok 15:26 Krock BlockMen, and defaultsetting.cpp? 15:26 Krock I think the example file and the internal defaults should be the same 15:27 BlockMen https://github.com/minetest/minetest/blob/master/src/defaultsettings.cpp#L197 15:27 Zeno` BlockMen, https://github.com/minetest/minetest/issues/2315#issuecomment-74378796 15:27 Krock okay, I revert my comment 15:27 Zeno` BlockMen, you're malicious! 15:28 BlockMen lol :D 15:28 BlockMen stupid chrome.. 15:28 Krock BlockMen, https://github.com/minetest/minetest/blob/master/src/defaultsettings.cpp#L198 https://github.com/BlockMen/minetest/commit/ae747139#diff-2fb5b44f6775f04cbd1e45bc2dfeb8b8R263 15:30 BlockMen Krock, fine. here, have a cookie ;P 15:30 Krock thanks :D 15:33 BlockMen Zeno`, can i merge #2314? 15:33 ShadowBot https://github.com/minetest/minetest/issues/2314 -- Fix font_size by BlockMen 15:35 Krock Comments of experts required on #2310 15:35 ShadowBot https://github.com/minetest/minetest/issues/2310 -- Update zlib support by SmallJoker 16:04 nrzkt sfan5 can you give me the developper privilege on IRC please ? :) 16:04 sfan5 sure 16:04 nrzkt thanks 16:05 nrzkt for the forum it's also you ? 16:05 sfan5 no, i don't administer the forum 16:05 nrzkt okay, thanks :) 16:05 sfan5 -ChanServ- 14 nerzhul_fr +AVt (dev) [modified 8s ago] 16:07 nrzkt i managed to compile windows 32b on a jenkins, then for release engineering it's good :) 16:11 BlockMen nrzkt, forum user "nrz"? 16:11 nrzkt right 16:12 nrzkt sfan5, if you change your versions in your toolchain, please tell me, i'm using jenkins to integrate your script 16:13 BlockMen nrzkt, changed 16:13 sfan5 nrzkt: you mean the versions in buildwin{32,64}.sh? 16:13 nrzkt thanks BlockMen. Yes sfan5 16:16 sfan5 nrzkt: I'll try to remember to tell you when I update those 16:16 nrzkt sfan5: you don't use the same version for win32 and win64, example oggversion 1.2.1 for 32b and 1.3.1 for 64b 16:17 sfan5 I haven't seen a reason to update libogg yet 16:17 sfan5 also the 64-bit build script was made later than the 32-bit one 16:17 nrzkt ok and for freetype, 2.3.5 and 2.5.3 ? 16:17 sfan5 basically all version differences are because the 32-bit script has been there longer 19:01 casimir There was an issue on github with the server shutting down because the memory was used up or something. I can't find it again. 19:21 BlockMen comments on #2316 plz 19:21 ShadowBot https://github.com/minetest/minetest/issues/2316 -- Fix wielded item by BlockMen 20:28 est31 has this "master breaks network" issue been resolved? what was the result= 20:28 est31 ?* 20:30 est31 nrzkt, are there already other breaking changes in master? 21:21 nrzkt est31 what would you mean ? there are no break in current master 21:22 est31 will patch 2/4 break? 21:30 nrzkt no 21:30 est31 so my change would be the firs one that breaks? 21:35 nrzkt yes, but i think we will merge it avec patch 3/4, because patch 3/4 is protocol rework 21:36 gregorycu nrzkt: For #2225, you are talking about brackets again, right? 21:36 ShadowBot https://github.com/minetest/minetest/issues/2225 -- Fix for getCraftRecipes implementation is very slow by gregorycu 21:36 est31 so I can make a pull request to your repo? 21:36 nrzkt right gregorycu 21:37 nrzkt it's useless, use minetest repository please, we will rebase after patch 3/4 integration (your rebase will be very little) 21:39 est31 ok 21:43 est31 so 3/4 is breaking protocol? 21:44 gregorycu If we are breaking protocol, are we making sure we get all the PRs in which also need a break in protocol? 21:45 nrzkt we must study it, but master is not for production, master is the devel branch, so breaks can occurs until feature freeze 21:47 gregorycu The thing is, we get great testing done from people using master branch 21:47 gregorycu We want to continue to leverage those people 21:47 gregorycu Can we get the best of both worlds? 21:48 VanessaE ~tell BlockMen "(and Krock) I meant official release builds." 21:49 ShadowBot VanessaE: O.K. 21:49 * VanessaE reads backscroll.. 21:52 est31 The central server list will need to distinguish between older (0.4.12 -branch) and newer servers (5.0-branch) depending on the client version. 21:52 nrzkt i think 21:53 nrzkt maybe have a version field on master server, registered by >= 0.5 to show the compat, and if old tell old compat. 21:54 est31 there is already a version field 21:54 est31 only the client needs to say "hey, this server is too new for me" 21:54 nrzkt okay, i didn't know, (i haven't reviewed all the code :p) 21:54 est31 np :) 21:55 est31 I think it would be better though if the master server did the distinction. 21:55 est31 after all, old clients will still display servers >=0.5 21:55 gregorycu There is no reason why the master server couldn't have a parametised query 21:55 nrzkt yes, because at the end of the networking rework old client couldn't connect to new servers because it will be TCP instead of UDP and init packet will be reworked to a better creation design 21:56 gregorycu We are seriously moving to TCP? 21:56 nrzkt i'm testing it and it works very very well 21:56 gregorycu This is a bad idea 21:57 nrzkt the network layer is so simple. If we need no check packet i think we could use UDP for non secure packets 21:58 nrzkt don't stay in old school games. UDP for FPS: OK, UDP for RPG ? useless. World of warcraft mainly uses TCP sockets and the packets sent between client and servers are ~1000 differents and there are many and amny packets sent unlike minetest 21:59 nrzkt the TCP also permit to guarantee sessions and we don't have to code a secure layer to evade the packet injection problem that Minetest has which permit to inject datas into an existing session with a source address spoof... 21:59 Calinou “is Minetest a RPG” that's a nice troll question :P 21:59 gregorycu I'm not going to argue with you, merely because I want you to put a +1 on my PR 21:59 nrzkt minetest is more like a RPG than a FPS :p, on network side i mean 21:59 nrzkt xD 22:00 nrzkt in fact your PR are on a unknown layer of MT for me, but they seems cool (except the coding style :p => http://dev.minetest.net/Code_style_guidelines 22:01 gregorycu Except for braces 22:01 gregorycu My PRs are very cool 22:02 gregorycu One day they'll get merged, and people will be able to benefit from them 22:02 gregorycu Like a bugfix for a 4 year old bug 22:03 nrzkt if you find bug on shaders, yeah cool, personnaly i found bug on network layer at each time i replace a thing. The last network layer patch will fix many design problems 22:03 kilbith they won't get merged at all since you decide to ignore Zeno 22:03 gregorycu Is this Zeno's project? 22:03 gregorycu I didn't know that 22:03 gregorycu I hope he never gets hit by a bus, otherwise the project is dead 22:03 nrzkt in fact Zeno` needs to discuss with you about those PR because he is the most competent and available on the parts where you work 22:03 kilbith other devs are agree with it 22:04 gregorycu All Zeno did was downgrade the priority 22:04 est31 I dont think you should exclude ppl just because they dont like one core dev. 22:04 nrzkt before yes, because we was on feature freeze without branch. But now master is detached and your PR needs attention 22:05 kilbith Zeno is mainly responsible for optimizing stuff 22:05 nrzkt but don't ignore him, i think you could have a good collaboration for the project 22:05 est31 yea, best is you both resolve this 22:05 hmmmm oh god wtf is happening to minetest 22:06 gregorycu hmmmm: Don't ask, don't tell 22:06 hmmmm i haven't been around for about a week 22:06 nrzkt someone knows why this f******g cmake ignore the CMAKE_LINKER option to change /usr/local/bin/ld by another program ? xD 22:09 gregorycu nrzkt: I have updated the code for #2225 22:09 ShadowBot https://github.com/minetest/minetest/issues/2225 -- Fix for getCraftRecipes implementation is very slow by gregorycu 22:09 VanessaE there, caught up. 22:11 nrzkt gregorycu, warn, functions braces mustn't be like if/for/while braces, they are on a new line, reread the dev guidelines i linked 22:12 nrzkt est31 this PR works well? it seems you tested it 22:13 gregorycu Oh, so functions on nerline, other things not 22:13 nrzkt yes 22:14 nrzkt #2312 seems good for me, Tesseract, you agree 22:14 nrzkt ? 22:14 ShadowBot https://github.com/minetest/minetest/issues/2312 -- Update Spanish translations by ngosang 22:17 gregorycu nrzkt: How's that? 22:20 nrzkt gregorycu: sorry ? 22:20 gregorycu I updated the PR, with brackets on newlines for functions, but on the same line for other things 22:20 est31_ brackets on newlines for functions?? 22:21 est31_ thought always same line... 22:21 est31_ http://dev.minetest.net/Code_style_guidelines 22:21 est31_ if this is up to date 22:22 gregorycu Yeah, I'm not changing that code again 22:22 JakubVanek (Linux kernel style) no, functions are the only exception to this rule 22:22 gregorycu If it's not correct, I'm just going to close it 22:22 gregorycu I have better things to do 22:25 est31_ ok then i was wrong 22:25 nrzkt function braces on new lines. Others at the line's end 22:26 nrzkt #2278 bug fixed in master. i haven't used a PR, the fix is trivial 22:26 ShadowBot https://github.com/minetest/minetest/issues/2278 -- Rare main menu error when starting Minetest 22:31 nrzkt est31 #2232 is already triggered on only in certain cases? 22:31 ShadowBot https://github.com/minetest/minetest/issues/2232 -- Invalid read of size 8 22:53 gregorycu Minetest runs a lot better through a profiler 22:54 gregorycu Almost twice better 23:02 gregorycu lol, I had breakpoints with counters 23:04 nrzkt sfan5 can you review #2317 please ? 23:04 ShadowBot https://github.com/minetest/minetest/issues/2317 -- Add an option to disallow guest connection (disabled by default) by nerzhul 23:06 nrzkt or hmmmm ? or Tesseract 23:08 VanessaE I dunno, is it really all that much faster to handle it in-core? 23:09 VanessaE I mean, perceptually speaking 23:09 nrzkt yes 23:09 nrzkt we only need to compare the 5 first char to define it's a guest. Why do you need to do a difficult code ? :p 23:09 kilbith plus not all admins have the reflex to install the NoGuest mod 23:09 VanessaE well the no-guest mod does that with a simple regex 23:09 nrzkt if you agree please add +1 on the PR 23:10 kilbith it's not my point, not all admins are aware of that mod 23:10 VanessaE I'm neither for nor against it 23:10 kilbith which is essential, IMO 23:11 nrzkt ofc VanessaE, i could use a regex too, but why use a regex instead of comparing only the 5 first char ? it's a little bit faster. And the core PR is handled on _INIT packet, mods cannot, they only are active a little bit after: on_prejoinplayer 23:11 nrzkt in fact there is no need for a mod there, it's useless lua code. core can do it easily, mod don't need to have an existence. This also permit to reduce memory and CPU overhead of using a mod at connection 23:12 gregorycu What is a guest? 23:12 VanessaE nrzkt: well no reason to use a regex in the case of "guest" but in the case of the no-guests mod it's useful, because it's just one more pattern to match among several others 23:12 kilbith all serious servers have NoGuest installed, so why not put it hardcoded 23:12 VanessaE for example that mod also checks for all-numeric names 23:13 VanessaE and names containing "Admin" and l33tsp34k variants 23:13 nrzkt else, this is not a guest for me. I only remove the guests here :) 23:13 gregorycu What is a guest? 23:13 VanessaE gregorycu: "Guest1234" 23:13 nrzkt it's a player called by default by minetest GuestXXXX 23:13 VanessaE the default username the client assigns you when you sign onto a server without giving a name 23:14 gregorycu This should not be part of core 23:14 n4x gregorycu: a wart in minetest 23:14 VanessaE ("Admin" is what buildcraft et al. use in place of "singleplayer" and somehow that also replaced Guestxxxx) 23:14 nrzkt i think a mod is great for custom patterns, right. But guest is part of minetest core, then it must be handled by core 23:14 gregorycu What if I name myself Guest? 23:14 VanessaE then you're screwed :) 23:14 nrzkt if you enable the option you will be rejected by server 23:15 VanessaE imho "Guestxxxx" needs to be ripped out of clients, and disallowed by servers. 23:15 kilbith this is a stupid nick anyway 23:15 gregorycu I agree with Vanessa 23:15 VanessaE whether the disallow is done in-core or in a mod, I don't really care all that much 23:15 gregorycu I almost agree with Vanessa 23:15 nrzkt i agree for it. But why not keep the Guest name for new players if they want and let the admin disallow them if they want ? 23:16 gregorycu It doesn't feel right 23:16 nrzkt i know this is useful for many of us, but somes can let them. Guest-aware admins enable the option, the other does anything :) 23:17 VanessaE nrzkt: because it creates player account clutter, inflates player counts on a server, and doesn't serve to encourage players to return to a server. 23:17 nrzkt with the option in core player isn't created 23:17 gregorycu If we want this concept, it should be a concept 23:17 VanessaE plus, if the server requires a password, that locks out "guestxxxx" the first time that xxxx gets used. 23:17 gregorycu It should be some half-arsed thing 23:17 gregorycu shouldn't 23:17 nrzkt and with the mod i think no, if it's bind on on_prejoinplayer 23:18 VanessaE nrzkt: when using the mod, the same is true - player accounts are not created when the player is denied access. 23:18 nrzkt ok, then he uses the correct event 23:18 gregorycu Maybe we should have the concept of "guest" but it's a flag on the connection 23:18 gregorycu Where the client doens't need a password, they are assigned a name by the server 23:19 VanessaE gregorycu: that's a thought. 23:19 gregorycu I have them, from time to time, sometimes they are good 23:19 VanessaE with the proviso that nrzkt's idea still be put into use 23:20 gregorycu If it's a flag on the connection, then I would fully support being able to deny access in core 23:20 VanessaE on my servers, "Guest" accounts almost never ended up being legit users. just griefers. 23:20 gregorycu A default name isn't a strong enough concept for something to exist in core, in my eyes 23:21 gregorycu Of course, if you're concerned about griefers, using the name is a pretty shitty way to try and stop them 23:22 VanessaE well, with the variant of the mod that I use, players have to pass a few basic filters to create an account (not "guest", not "admin", not too many numbers in their name, can't seem like gibberish, can't be too short). That was enough to stop most of the griefers, probably because they're too lazy to bother :) 23:23 gregorycu Yeah, but do you do that to stop griefers? 23:24 gregorycu Or to facilitate a better playing environment? 23:24 VanessaE naw, I have rollback, area protection, and moderators/admins I trust also 23:24 gregorycu (Or both really) 23:24 VanessaE that particular bit is more of a community-building effort 23:24 gregorycu Yeah 23:24 VanessaE you don't easily build a community on a server if everyone is "Anonymous Coward". 23:24 VanessaE oh btw, line 1472 of your patch, maybe you should write to the user, "Guest *usernames* are not allowed" (otherwise it looks like just any kind of guest is disallowed) 23:25 gregorycu 'Deny access to users whose name start with "Guest" ' 23:25 VanessaE below that, the message you send to the user 23:26 nrzkt ok, then i add the keyword usernames. And i also add the following tip, suggested by kilbith: please chose a nickname 23:26 VanessaE s/you send/nrzkt sends/ ;) 23:26 gregorycu I have bug fixes for 4 year old bugs, and we are looking at implementing something that can already be done by mods, and questionably should not be done by core 23:27 VanessaE nrzkt: +1 23:27 VanessaE gregorycu: you're right, the fixes for those old bugs need applied too 23:28 nrzkt sorry gregorycu, but repeating something whereas i tell you i'm not the most competent for your PR (not for C++ but for core component itself) doesn't let me merge your commit :( 23:29 nrzkt what about: "Guests nicknames are not allowed on this server. Please choose a valid nickname." 23:29 VanessaE s/Guests/Guest/ but let's table this for now. 23:29 gregorycu Nothing personal nrzkt, it's just a microcosm for things in general 23:30 kilbith or just "Please choose a valid nickname." 23:30 nrzkt no kilbith because we can have a different reason for rejecting name, here we are precise 23:30 gregorycu You should probably at least define what a valid nickname is 23:31 kilbith player are not aware of "Guest" attribution before entering in a server 23:31 nrzkt in fact here we define guest as a non valid, not all the valid username list 23:31 gregorycu What if someone names themselves "Guestapo" ? 23:32 VanessaE then they need to learn to spell 23:32 VanessaE :) 23:32 gregorycu They'd want to know what is wrong with their name 23:32 VanessaE (and if they spell it right, "Gestapo", they need to piss off entirely :P) 23:32 gregorycu Apart from, you know, being offensive to Jews, and the english language 23:33 est31_ godwin strikes again 23:33 nrzkt i updated the PR with the last optimisations, thanks for your ideas. I will talk with other core devs for this tomorrow to see if the patch was okay and will be merged. Good nich all 23:33 est31_ gn 23:33 VanessaE night nrzkt 23:34 gregorycu :/ 23:34 gregorycu I just don't like it