| blast007 | we do some weird thing where we store a pointer value in a BZDB variable called poll. | 00:02 |
|---|---|---|
| TimRiker | that is a weird thing. | 00:09 |
| *** BZuser804 <BZuser804!~BZuser804@76.8.32.203> has joined #bzflag | 03:13 | |
| *** BZuser804 <BZuser804!~BZuser804@76.8.32.203> has quit IRC (Client Quit) | 03:13 | |
| BZNotify | master @ bzflag: timriker pushed 1 commit (https://github.com/BZFlag-Dev/bzflag/compare/007402d3fac8...ec3cb062326d): | 03:28 |
| BZNotify | master @ bzflag: timriker ec3cb0: one line (https://github.com/BZFlag-Dev/bzflag/commit/ec3cb062326d85c0876929a18dc1a1ef10d657ff) | 03:28 |
| TimRiker | How do I test plugins? I see plugins/R2/R2.la for example, but no R2.so | 04:58 |
| blast007 | do you have plugins/R2/.libs/R2.so ? | 05:07 |
| TimRiker | ah! yes. | 05:07 |
| TimRiker | plugins do not get loaded into autocomplete? | 05:17 |
| blast007 | autocomplete? | 05:24 |
| BZNotify | master @ bzflag: timriker pushed 1 commit (https://github.com/BZFlag-Dev/bzflag/compare/ec3cb062326d...fcac57df69e5): | 05:26 |
| BZNotify | master @ bzflag: timriker fcac57: check local plugin path if DEBUG (https://github.com/BZFlag-Dev/bzflag/commit/fcac57df69e5ab6c03d8be63da9411b4ae6469de) | 05:26 |
| TimRiker | in bzadmin you can hit tab to complete partial commands. however there's no command to list all commands, no default help, etc. | 05:37 |
| TimRiker | and tab is used in game, not captured when chatting, so I guess tab complete does not work there either. | 05:37 |
| TimRiker | /playerlist crashes when the sample player is attached. not sure why yet. perhap it does not fill in an address. | 05:38 |
| blast007 | does /? work in bzadmin? | 06:01 |
| blast007 | if so, then /b? might also work to list all commands starting with b | 06:01 |
| *** Krunch <Krunch!~Krunch@92.243.9.71> has joined #bzflag | 06:05 | |
| Krunch | hey, just noticed the new feature_ip6 branch; i actually researched the topic a bit and took notes here that you may find of interest https://docs.google.com/document/d/1r8gTVeIAJGFXY8GRJAAWXlJis9LR_J1I_oMXD3RJ_yo/edit?usp=sharing | 06:14 |
| Krunch | i started actually writing some code for it but the state of build and test system make it very tedious | 06:14 |
| Krunch | also logging | 06:16 |
| TimRiker | the branch was a partial test by me just to see what is needed. I should delete it. I am looking at what was changed there and trying to clean up our network code a bit as we move towards supporting ipv6. blast has some sample code, and so do I outside of the game. | 06:16 |
| *** I_Died_Once <I_Died_Once!~I_Died_On@c-73-184-170-223.hsd1.ga.comcast.net> has quit IRC (Ping timeout: 268 seconds) | 06:17 | |
| Krunch | is there A Plan written down somwhere? what's a good way for me to contribute? | 06:19 |
| Krunch | (fwiw i discussed this a bit with JeffM and blast007 back in January 2018; this got to the bottom of my priority list for a few years and i started looking at it again at the beginning of this year) | 06:23 |
| TimRiker | we should start by passing sockaddr structs around so that we have a AF. then move from things like inet_ntoa over to inet_pton same for other functions that don't support IPv6. | 06:24 |
| TimRiker | all this can happen while using sockaddr_in directly. In theory once we are using calls that support IPv6, we can move to sockaddr_in6 and not have to change a ton of code. blast007 is hoping for multiple listening addresses, so IPv4 users get AF_INET and ipv6 users get AF_INET6. | 06:26 |
| TimRiker | That would be nice, but might be more complication than it's worth. I'd be happy with sockaddr_in6 everywhere and custom toString calls that hide the ::ffff: portion of the address. | 06:27 |
| TimRiker | your doc points out that masks are also an issue. banlist could presumably have ipv4 and ipv6 masks for bans, etc. | 06:27 |
| TimRiker | my dev servers are ipv6 enabled. forum-dev my-dev www-dev images-dev, etc. | 06:30 |
| BZNotify | bzflag-bzfls: timriker review_requested pull request #1 "Token nameport support" (https://github.com/BZFlag-Dev/bzflag-bzfls/pull/1) | 06:32 |
| TimRiker | I've changed token handling to avoid pinning to an IP. Waiting on blast to review and potentially add to the prod server. It won't be IPv6 enabled yet, but the code to add/remove/list that removes the ip pinning should work there. | 06:32 |
| TimRiker | I think clients that have v6 will likely default to v6. I suppose we could ping both ips and see which is faster, then jump on that one. The server tracks the one ip for the client, so we can't move back and forth without a reconnect/rejoin. | 06:34 |
| Krunch | i think supporting ipv4+ipv6 clients at the same time is a lot of added complication that might not be needed until after we have a version that works with ipv6 at all? | 06:35 |
| TimRiker | I'm happy to give people rights to the dev boxes if desired. I'd like to auto-build bzfs and auto-restart it, but I've not set that up. I used to run a server like that many years ago. | 06:35 |
| TimRiker | should be easy enough. clients don't see each other's ip address. only the server and admin/ban commands show that now. | 06:35 |
| TimRiker | it used to be that all clients saw the ips. | 06:36 |
| TimRiker | I'm NOT suggesting that 2.4 ipv4 only clients will be able to see 2.6? ipv6 clients. I'm suggesting that 2.6? will work for ipv4 only, ipv6 only and mixed games. | 06:36 |
| Krunch | right now i have just updated net/network.cxx to familiarise myself with the build but i struggle to verify my change works at all | 06:37 |
| TimRiker | if the server only support v4, everything will be AF_INET. if the server supports v6, probably everything will be AF_INET6, even v4 clients. they will be ::ffff:a.b.c.d on the server, but playerlist etc will show them as only a.b.c.d | 06:38 |
| Krunch | yes, i see how to make mixed clients work; i am just saying that's more work on top of making ipv6 work and that the latter can be done without the former but without digging a hole that make it harder/impossible in the future | 06:39 |
| TimRiker | bans, playerlist, etc will need to handle either family sockaddr structs. | 06:39 |
| TimRiker | easier than supporting two branches with a v4 only branch. | 06:40 |
| TimRiker | what did you change on network.cxx that you are trying to test? | 06:40 |
| Krunch | i was not thinking of keeping two separate branches; more like having a --ipv6 flag at run time | 06:40 |
| TimRiker | -i 0.0.0.0 should do just that. | 06:41 |
| TimRiker | well, the reverse. the default would be -i :: which would enable both clients. -i 0.0.0.0 will disable v6 clients as we would not be listening there. | 06:42 |
| TimRiker | I expect there will be people with broken v6 routing that will want to use -i 0.0.0.0 | 06:43 |
| Krunch | my change: https://paste.debian.net/plainh/60d157d5 (i should probably create myself a github account at some point) | 06:44 |
| TimRiker | to help out, pick one of those tasks and make a pull request. convert inet_aton() calls to inet_pton(). etc. | 06:45 |
| Krunch | one of which tasks? | 06:45 |
| TimRiker | that's a good start. I'd wrap curl_ur inside the parseURL call and then remove our parsing. I think curl might support username and password in the url, which ours does not. I don't know if it handles https://[2001::1]:port format urls. we might need that for v6 servers without dns. | 06:49 |
| Krunch | the doc claims it does support the ipv6 url format | 06:50 |
| TimRiker | ones that we've talked about. you can open a github "enhancement" for the curl_url change, then fork, push your changes, and create a pull request. | 06:50 |
| TimRiker | excellent. | 06:50 |
| Krunch | i still struggle to verify it actually works as there is no test i can find and last time i tried HEAD was not running correctly in the first place and there doesn't seem to be a working logging system either | 06:53 |
| Krunch | or is the strategy to just update code and debug it before release? | 06:54 |
| TimRiker | I'm running two servers on master and testing bzadmin and bzflag against them. The catch is that since my token/key changes are not on the prod list server, I'm running on the dev list server. | 06:54 |
| Krunch | ok so i just send code and you debug it? | 06:54 |
| TimRiker | create a pull request. | 06:55 |
| TimRiker | note: parseURL() is not in master anymore. | 07:01 |
| TimRiker | we probably want to use getaddrinfo() in most places where we have inet_aton() now. | 07:06 |
| Krunch | i can't find the commit that removed parseURL | 07:24 |
| TimRiker | there have been quite a few commits on master. it's in there someplace. just update your patch against master and submit a pull request. | 07:27 |
| Krunch | yeah, i just don't understand how git or github works because the change should be between ee23 and a408 but it claims there isn't anything there https://github.com/BZFlag-Dev/bzflag/commits/master/src/net/network.cxx | 07:28 |
| Krunch | i think i am done trying to make sense of git/github for today; might try again next week | 07:28 |
| TimRiker | looks like commit ee2313eeabae4f3a899737581ace21b9007fdd08 | 07:31 |
| TimRiker | https://github.com/timriker/ipv6 has some examples I'm playing with. | 07:38 |
| BZNotify | master @ bzflag.org: allejo pushed 1 commit (https://github.com/BZFlag-Dev/bzflag.org/compare/84884228cc2d...fc3ff3550230): | 07:46 |
| BZNotify | master @ bzflag.org: allejo fc3ff3: Remove redundancy of BZDB settings in flag docs (https://github.com/BZFlag-Dev/bzflag.org/commit/fc3ff3550230da36b50265333ecf2e2006310199) | 07:46 |
| *** L4m3r_ <L4m3r_!~L4m3r@user/L4m3r> has joined #bzflag | 08:10 | |
| *** L4m3r <L4m3r!~L4m3r@user/L4m3r> has quit IRC (Ping timeout: 255 seconds) | 08:11 | |
| *** Cobra_Fast_ is now away: vacant | 08:29 | |
| *** Cobra_Fast_ is back | 08:29 | |
| TimRiker | https://github.com/timriker/ipv6/blob/master/getaddrinfo-server.c parses name<space>service from the command line and listens there on udp. | 08:34 |
| *** L4m3r_ is now known as L4m3r | 09:19 | |
| *** Sgeo <Sgeo!~Sgeo@user/sgeo> has quit IRC (Read error: Connection reset by peer) | 10:11 | |
| *** Cobra_Fast_ is now away: vacant | 13:41 | |
| *** Cobra_Fast_ is back | 13:41 | |
| *** I_Died_Once <I_Died_Once!~I_Died_On@c-73-184-170-223.hsd1.ga.comcast.net> has joined #bzflag | 14:15 | |
| *** FastLizard4 is back | 15:25 | |
| *** FastLizard4 is now away: AWAY from keyboard | 15:55 | |
| *** FastLizard4 is now away: GONE - Screen Detached and Disconnected from IRC (I'm probably asleep, at work, or doing something in real life) | 16:12 | |
| *** tupone <tupone!~tupone@gentoo/developer/tupone> has quit IRC (Remote host closed the connection) | 16:13 | |
| *** tupone <tupone!~tupone@gentoo/developer/tupone> has joined #bzflag | 16:17 | |
| *** Sgeo <Sgeo!~Sgeo@user/sgeo> has joined #bzflag | 18:58 | |
| *** jswldgr <jswldgr!~jswldgr@047-229-253-227.res.spectrum.com> has joined #bzflag | 20:17 | |
| *** jswldgr is now known as asparticus | 20:22 | |
| *** Flash <Flash!~Flash@user/flash> has quit IRC (Read error: Connection timed out) | 22:27 | |
| *** Flash <Flash!~Flash@user/flash> has joined #bzflag | 22:28 | |
| *** FastLizard4 is back | 22:46 | |
| BZNotify | master @ bzflag: timriker pushed 1 commit (https://github.com/BZFlag-Dev/bzflag/compare/fcac57df69e5...d77287f45bac): | 23:51 |
| BZNotify | master @ bzflag: timriker d77287: why is shotID always the same? (https://github.com/BZFlag-Dev/bzflag/commit/d77287f45bac42142e60dd2135f831ffd77373c1) | 23:51 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!