Time  Nick        Message
00:08 celeron55   VargaD: if you want, you can set xyz as the commit author in various ways by using git
00:09 celeron55   (he should answer by himself though)
00:10 celeron55   (but that'll be after sleep i guess)
00:10 VargaD      thanks, I didn't know that. I'm not a git expert :)
00:28 thexyz      I'm not sure, there are copyrights already in the new files
00:29 thexyz      and the code is so damn simple I honestly don't care that much
07:39 VargaD      'morning
13:32 VargaD      Are there any release schedule? When so you release next version?
13:33 PilzAdam    VargaD, we dont have release schedules
13:33 VargaD      How do you decide to release a new version?
13:34 PilzAdam    when many core devs think that its time to release it
13:35 VargaD      I see, thanks
14:08 sapier      actually release time is pure chance ;-)
14:12 VargaD      according to github there are 46 open bug
14:15 sapier      just ignore them ;-) (at least some of em)
14:15 sapier      or the better option, fix it
14:38 celeron55   VargaD: any core dev can set and oversee a release; the only one who really cares to do that is hmmmm so you might wait to ask him (or request it)
14:40 VargaD      thanks celeron55, I just wanted to understand minetest development better
14:41 celeron55   i guess there could be a better description somewhere...
14:43 sapier      true :-) Imho release procedure is chaotic :-)
14:44 celeron55   i've always called this the "minetest chaos development model"
14:44 sapier      quite suitable description
14:44 celeron55   it started pretty soon right when the first contributors came and i haven't figured out how to make it non-chaotic since then 8)
14:45 sapier      I don't think that'd give better results, by now we at least tried to get worst bugs out prior releasing
14:46 sapier      btw celeron hope you don't develop reliable communications protocols nowadays ;-)
14:46 celeron55   sapier: it was my first-ever attempt in making one; i'm surprised you could even bother to fix it
14:47 sapier      I think I got it :-)
14:47 sapier      but if I want to do reliable communication to old clients I have to do very very bad things :-)
14:48 sapier      send --> waitforack --> sendnext
14:48 sapier      at least I can do this per channel and per peer so old clients wouldn't slow down new ones ... and tcp is evolving right now too maybe a first prototype is ready this year
14:51 sapier      why drop em if it's possible to keep?
14:52 sapier      most ppl (except of developers) greatly appreciate compatibility ;-)
14:53 sapier      no it doesn't actually fixing the old protocol gave me ability to do a better design for tcp
14:54 sapier      /usr/share/wireshark/init.lua
14:54 sapier      if you want to do this you may be interested in my improved version of it
14:55 sapier      I added the new messages (at least the names)
14:55 sapier      no you have to add a dofile at end
14:59 celeron55   http://wiki.wireshark.org/Lua <- that recommends using tshark flag -X
15:00 celeron55   wireshark (the gui) takes the same kind of option
17:24 VargaD      sapier: what do you plan to send on tcp and waht on udp?
17:24 sapier      same thing as now unreliables are sent through udp while reliable date will be sent through tcp
18:05 VargaD      I'm wondering, where does the minetest name come form?
18:43 Calinou     it was supposed to be a "test"
18:44 sapier      wasn't minetest designed to be perfect? *smile*
19:30 ShadowNinja #1074 seems good, with the last line change removed.
19:30 ShadowBot   https://github.com/minetest/minetest/issues/1074
19:38 sapier      for old clients nothing will change but for all new ones (with enabled tcp support) reliable data will be sent through tcp only
19:46 ShadowNinja Older clients will still get faster UDP though, right?
19:47 sapier      if you are talking about my udp fixes yes but due to compatibility reasons old clients using udp may not benefit as much as I was hoping
19:47 ShadowNinja And this will be in a library that MT links to so that other applications can use it, right?
19:48 sapier      there's a bug in current udp reliability handling breaking it on high packet loss, only way to do reliable communication to old clients is send one by one
19:48 sapier      not by now
19:49 sapier      splitting it to a library would be another 1-2 weeks of work
19:53 sapier      to make this a library design needs to be way more clean then even my fixed version is
19:55 ShadowNinja O.K. That's a project for you to work on as soon as it works. :-)  And what do you think about that PR?
19:56 sapier      haven't had a look at it by now
19:56 sapier      that simple?
19:57 ShadowNinja Should be.  I looked into the Rollback error and it was SQLite complaining that the directory didn't exist.
19:58 sapier      did you try it?
19:58 ShadowNinja Nope, not yet.
19:59 sapier      if you wanna merge it you should try it first, I'm busy writing tcp support right now