Time Nick Message 00:00 Fixer minetest_game was left out as usual, 10 PRs 00:02 rubenwardy Lol 00:02 rubenwardy That ain't Minetest 00:04 Cornelia What? :P 00:16 paramat 10/ 05:17 nerzhul rubenwardy, i need to finish all my pr to close 5 more xD 07:05 nerzhul sfan5: i have a nearly ready clang-tidy integration on the CI, i just enabled one warning as an error, the remaining thing is to fix the return code of the processus 07:05 nerzhul https://travis-ci.org/minetest/minetest/jobs/360163640 09:52 nerzhul Hello guys. #6295 is now ready for a review/merge (merge after adding the commit which fix the clang-tidy POC) 09:53 ShadowBot https://github.com/minetest/minetest/issues/6295 -- LINT: add clang-tidy step by nerzhul 09:53 nerzhul a step to increase our code quality 12:08 Krock nerzhul, I'm not sure how I could give a helpful feedback there. The build system is not something I know too well 12:10 nerzhul you can comment the tool itself 12:10 nerzhul there is nothing to review on the build tool iself :p 12:30 nerzhul Krock: did you agree with a such tool to increase easily our C++ PR code quality ? (and regular codebase too) 12:31 nerzhul Clion integrate it, we can also add a .clang-tidy file (or maybe i did it some ages ago) permitting to sync all IDE to clang-tidy. QTcreator 4.6 has added clang-tidy support too 12:31 nerzhul generally i use it to fix many code mistakes around my modifications 12:31 Krock PR code quality increase is generally welcome as long our code styles are used correctly 12:31 nerzhul clang-tidy is not for code style but yes :) 12:31 nerzhul and we have an increasing number of files perfectly compliant between clang-format and our code style :) 12:34 nerzhul the current output is huge heh, not all cases needs to be solved but at least we have a nice list to fix :p 12:34 nerzhul i need to find what's the problem with 7179 now... it compiles like a charm on my clang but not on travis... 13:00 nerzhul #6295 is now okay for a merge, Krock, sfan5, rubenwardy are you okay to introducing a such tool ? 13:00 ShadowBot https://github.com/minetest/minetest/issues/6295 -- LINT: add clang-tidy step by nerzhul 13:01 sfan5 depends on how it performs (usefulness of suggestions) 13:01 nerzhul i just enabled the push_back -> emplace_back check in blocking mode currently, we can enable more in another PR discussion 13:01 nerzhul it's warn on performance-* and error (CI blocked) for push_back -> empalce_back currently 13:01 nerzhul warning should be read from times to time to be recued when needed and function should be promoted to error when we should always use it 14:53 nerzhul how can i sell the devel team the fact to have clang tidy in CI ? hmmm, for me it's a time gain on PR review, mistake problem autofixer for PR owners :) 15:20 nerzhul will merge #7179 after travis pass 15:20 ShadowBot https://github.com/minetest/minetest/issues/7179 -- Client eventmanager refactor by nerzhul 16:33 nerzhul merged. Travis network is buggy today 17:16 rubenwardy nerzhul: how hard would it be to add support for virtual hosts? 17:17 rubenwardy say foo.com and bar.com both point to the same IP 17:17 rubenwardy basically having a program to route packets based on requested host 17:18 rubenwardy on connection the client should send the requested host which the router reads and then adds a routing rule until they disconnect 17:18 nerzhul on what ? on MT ? 17:19 rubenwardy 99% of servers would ignore that packet as they don't do virtual hosting 17:19 rubenwardy yeah, MT 17:19 nerzhul currently it's not possible 17:19 rubenwardy well yes 17:19 nerzhul and it needs to break the whole low layer protocol to add a new field 17:21 rubenwardy couldn't it just be part of the client->server hello packet? 17:21 rubenwardy like, added to the end 17:21 rubenwardy then if it's possible to differentiate between different clients coming on the same IP, couldn't you add routing rules to redirect for a particular session? 17:21 nerzhul what do you want to achieve ? 17:21 rubenwardy I dunno, that's a rather philosophical question 17:21 nerzhul if you come to MT session layer it's too late to "route" as it's a layer 4 problem :) 17:22 nerzhul you are on layer 7 17:22 rubenwardy well, the routing wouldn't be with MT 17:22 rubenwardy ah, I guss 17:22 nerzhul but some virtualhosting especially on HTTP is done on layer 7, but that mean you need an additional software for that 17:23 rubenwardy well, it's hardly a thing that's needed 17:23 rubenwardy just interested in how it would be done 17:23 rubenwardy would rather work on UPnP / LAN discovery first 17:32 nerzhul to discover MT server ? just use a library if needed 17:32 nerzhul we should stop to develop some things and let better coders do it :) 17:33 troller it was implemented in freeminer years ago ;) 17:34 rubenwardy How? 17:34 rubenwardy Nerzhul: well yeah 17:34 rubenwardy Thinking of researching bonjour 17:34 rubenwardy And others ofc 17:35 troller in simplest way as always 17:35 rubenwardy Lol 17:35 rubenwardy Is the routing part of Minetest? How do you give the host name? 17:35 rubenwardy *freeminer 17:36 troller wat? why need routing? 17:36 rubenwardy Because you have multiple servers on one IPm 17:36 rubenwardy Oh wait 17:37 rubenwardy You could just respond with a port 17:37 rubenwardy So 30000 would just be a redirect to another port 17:37 rubenwardy That's a lot simpler 20:13 Krock nerzhul, I think unit tests can be added in a separate PR for he entity collision PR. That piece is lying around for too long already 20:14 Krock meanwhile I picked up the player inventory callback idea. if I get time tomorrow I'll open a PR for it 21:56 nerzhul Krock each time we said that nothing happens 21:56 nerzhul and it's not a huge piece of code 22:05 longerstaff13 permission to be present :) 22:31 sofar not needed