| 01:47.22 | *** join/#bzflag Kutakizukari (~Kutakizuk@169.sub-70-210-6.myvzw.com) |
| 02:00.21 | *** join/#bzflag joevano (~joevano@bzflag/developer/JoeVano) |
| 02:00.22 | *** mode/#bzflag [+v joevano] by ChanServ |
| 02:05.30 | *** join/#bzflag Kutakizukari (~Kutakizuk@2600:1004:b12f:935f:d886:ded2:394b:3fa0) |
| 02:11.37 | *** join/#bzflag Kutakizukari_ (~Kutakizuk@2600:1004:b12f:935f:d886:ded2:394b:3fa0) |
| 02:13.17 | *** join/#bzflag Kutakizukari__ (~Kutakizuk@2600:1004:b12f:935f:d886:ded2:394b:3fa0) |
| 02:22.35 | Not-311d | [02bzflag-import-3] 07JeffM2501 pushed 032 commits to 03v2_4_x [+0/-0/±6] 13http://git.io/kdgT-A |
| 02:22.36 | Not-311d | [02bzflag-import-3] 07JeffM2501 036a808d4 - proper int to bool conversion |
| 02:22.37 | Not-311d | [02bzflag-import-3] 07JeffM2501 03a7bf2b5 - Merge branch 'v2_4_x' of https://github.com/BZFlag-Dev/bzflag-import-3 into v2_4_x |
| 02:41.16 | *** join/#bzflag zuii (~akos@BC06BD4B.catv.pool.telekom.hu) |
| 03:06.59 | *** join/#bzflag ahs3- (~ahs3-@2602:306:bc22:4d20:dc2:e9a8:9441:b259) |
| 03:25.32 | *** join/#bzflag therealtbe (~AresHandl@pool-74-109-244-171.pitbpa.fios.verizon.net) |
| 03:53.03 | Flash | Patlabor221: is there a compelling reason to handle MsgUDPLinkEstablished down at the network layer, rather than at the game layer? |
| 03:53.34 | Patlabor221 | It used to be optional, now it's required |
| 03:53.58 | Patlabor221 | It's a com channel thing so it makes more sense to be in the net layer then the game layer |
| 03:54.07 | Patlabor221 | it's not a game message |
| 03:54.34 | Flash | the thing is, sending messages ought to require a fully connected game interface |
| 03:54.43 | Flash | otherwise, it exposes us to DoS |
| 03:54.58 | Patlabor221 | except for the messages that we allow yes |
| 03:55.06 | Patlabor221 | yes you can block that untill they fully ID |
| 03:55.14 | Patlabor221 | won't matter once we remove that UDP system |
| 03:55.42 | Flash | I could see the game delegating the handling of the message back down to the network layer |
| 03:55.57 | Flash | as it is, "handling" that message allocates an un-connected player record |
| 03:56.35 | Patlabor221 | I'd set a flag in the networkhandler like "can UDP" and set that form the game |
| 03:56.54 | Patlabor221 | then let the networking layer read that |
| 03:56.59 | zuii | tcp is enough for gaming? |
| 03:57.05 | Patlabor221 | no |
| 03:57.06 | Flash | I think that flag already exists; today it is set natively in response to the packet |
| 03:57.10 | Patlabor221 | TCP is overkill |
| 03:57.40 | zuii | you want someking udp based protocol? |
| 03:57.44 | Patlabor221 | that current flag is if the client wants UDP, this would be if the server is going to allow UDP |
| 03:57.47 | Patlabor221 | zuii, no |
| 03:58.10 | Patlabor221 | Flash, so I think we need 2 flats for now |
| 03:58.17 | Patlabor221 | flags |
| 03:58.37 | Patlabor221 | zuii, we don't need TCP and it's overhead for update messages |
| 03:59.08 | Patlabor221 | there is no need to retransmit an old postion when we have new postion to send |
| 03:59.38 | Patlabor221 | Flash, I'd add "AllowUDPConnect" to NetHandler and set it only after a full join |
| 04:00.03 | Patlabor221 | then nethandler can know to ignore that message |
| 04:02.15 | Flash | In the case of this particular attack, is there any harm in treating multiple MsgUDPLinkEstablished messages as malicious? |
| 04:02.55 | Patlabor221 | since UDP is connectionless it may be legit to send multiples |
| 04:03.05 | Flash | no, that's a TCP packet |
| 04:03.17 | Patlabor221 | ahh |
| 04:03.24 | Patlabor221 | I'd say "maybe" thing |
| 04:03.26 | Patlabor221 | then |
| 04:03.28 | therealtbe | you could allow like 10 if you wanted to. he is sending like 50k per second |
| 04:03.36 | Patlabor221 | I'd still ignore it untill full connect |
| 04:03.50 | therealtbe | it might be somewhat normal to see 2 or 3 in the beginning connection from client, but seeing 50k ? cmon |
| 04:04.08 | Patlabor221 | that means storing counters and such |
| 04:04.15 | Patlabor221 | dooable if needed but probalby overkill |
| 04:04.21 | Patlabor221 | full connect is probalb good enough for now |
| 04:04.24 | Flash | what is the best thing to do in the event of a flood? stop reading the socket or drain and drop? |
| 04:04.32 | Patlabor221 | kill the socket |
| 04:04.37 | Patlabor221 | let the OS flush the buffer |
| 04:04.39 | therealtbe | tcpkill the IP |
| 04:04.42 | allejo | death! |
| 04:04.49 | Flash | I don't think so, because it will spawn a new socket |
| 04:05.09 | therealtbe | so leave it open and just stop processing it |
| 04:05.09 | Flash | and, it has bytes (from the server) unread, so the socket won't close for a couple of minutes |
| 04:05.26 | Flash | any downside to letting the TCP buffer fill up? |
| 04:05.33 | Patlabor221 | if it's TCP, it'd have to go thru the accept process again, and it would not start with the magic |
| 04:05.50 | Patlabor221 | so I don't think it will remake a socket that wesee |
| 04:06.17 | therealtbe | he would have to add another while loop to his script to do that... |
| 04:06.23 | Patlabor221 | different OSs handle TCP buffers diferently |
| 04:07.04 | therealtbe | I don't think it is the OS, but that it triggers the server to launch 50k reverse dns lookups. |
| 04:07.11 | Flash | the script is pretty simple; if one socket dies, open another. Then flood |
| 04:07.29 | Patlabor221 | with that packet? |
| 04:07.37 | Patlabor221 | or with the magic then the packet? |
| 04:07.47 | Flash | magic , then packet |
| 04:08.07 | Patlabor221 | yeah then there isn't much we can do other then block the IP at the system level |
| 04:08.14 | therealtbe | first of all the server should only listen to TCP packets from one inbound port from each client... the way to flood is he opening 5x tcp connections at the same time from the same IP |
| 04:08.34 | therealtbe | second of all, only execute reverse dns lookup once and cache results |
| 04:08.40 | Patlabor221 | therealtbe, that would break some NAT people |
| 04:08.43 | Flash | 5x may be overkill. I suspect one can do it |
| 04:08.52 | therealtbe | oook limit to... 100 |
| 04:08.53 | therealtbe | or 1000 |
| 04:08.54 | Flash | and yes, I already told therealtbe about the NAT issue |
| 04:08.58 | Patlabor221 | therealtbe, feel free to reqwrite the sever archetecture |
| 04:09.01 | therealtbe | limit to 1000 per NAT |
| 04:09.10 | therealtbe | he doesn't make lag til he hits over 11k |
| 04:09.15 | therealtbe | and then it spikes to 50k real quick |
| 04:09.30 | Patlabor221 | since he sends the magic we only have 2 choices |
| 04:09.45 | therealtbe | if it cache dns lookups he can spam all day, it won't be enough to make a delay |
| 04:09.50 | Patlabor221 | let it go, or implement an internal reconnect prevention system |
| 04:10.07 | Flash | my first cut at that kill the list server and strayer |
| 04:10.19 | Patlabor221 | yeah it's hard |
| 04:10.32 | Patlabor221 | strayer doesn't send a UDP message |
| 04:10.40 | Patlabor221 | there are services we allow with out a full connect |
| 04:11.04 | Patlabor221 | how often are his reconnects? |
| 04:11.10 | therealtbe | he does not reconnect |
| 04:11.14 | therealtbe | he opens 5 ports simulataneously |
| 04:11.23 | Patlabor221 | so then if we kill all 5 he's done? |
| 04:11.28 | Patlabor221 | or does he redo it? |
| 04:11.34 | Patlabor221 | I can kill those 5, that's easy |
| 04:11.39 | therealtbe | he would have to hit up arrow and execute another copy of his script. no one has battled him live |
| 04:11.45 | Patlabor221 | I can't easaly prevent the reconnect |
| 04:11.50 | therealtbe | each of the scripts, one on each port, spams about 50k messages |
| 04:11.55 | Flash | each killed socket is replaced by another |
| 04:12.02 | therealtbe | these launch reverse DNS lookups, which takes a long time. |
| 04:12.06 | Patlabor221 | that's the reconnect |
| 04:12.48 | Patlabor221 | so If I kill them on the first message, it'll imediately reconnect with a new socket right? |
| 04:12.55 | Flash | yes |
| 04:13.11 | therealtbe | http://pastebin.com/2xe5ejNm |
| 04:13.12 | Flash | plus you now have a socket in PENDING_WAIT |
| 04:13.13 | Patlabor221 | ok so then he'll reconnect multiple times per second |
| 04:14.07 | Patlabor221 | that we can detect and differentiate from the stats site |
| 04:14.13 | therealtbe | so yes ok it's wrapped in multiple while loops; it will catch the socket error and reconnect |
| 04:14.49 | Patlabor221 | ponders |
| 04:15.11 | Flash | One thought I had is that there is precious little reason to accept partial packets from clients |
| 04:15.25 | Flash | either the whole packet comes in a single read, or we junk it |
| 04:15.29 | Patlabor221 | not always |
| 04:15.40 | Patlabor221 | some messages are buffered |
| 04:15.49 | Patlabor221 | the protocoll is designed to allow for partials |
| 04:16.01 | Flash | I know it is; the question is, is that overkill? |
| 04:16.04 | Patlabor221 | bit things like map send, or flag lists |
| 04:16.16 | Flash | that is server to client |
| 04:16.17 | Patlabor221 | we can't be sure what each OSs max packet will be |
| 04:16.20 | Patlabor221 | both |
| 04:16.27 | Patlabor221 | mostly client tho |
| 04:16.29 | Flash | client sends big packets to the server? |
| 04:16.46 | Patlabor221 | techincaly a chat message could be large |
| 04:16.54 | Patlabor221 | but yes client to server is more rare |
| 04:17.01 | Flash | the server is much more vulnerable to malicious clients than clients are to malicious servers |
| 04:17.07 | Patlabor221 | sure |
| 04:17.12 | therealtbe | simple rate limiting on messages per second |
| 04:17.13 | Patlabor221 | the server is horribly designed :) |
| 04:17.19 | Patlabor221 | bzflag is NOT an internet game |
| 04:17.21 | therealtbe | this is outside 3stdev of normal traffic |
| 04:17.53 | Patlabor221 | but there are some things we can do |
| 04:18.12 | therealtbe | you can allow a ratio of partial packets to good packets to give leeway to legit clients, and in his case 100% partial will be a bad ratio |
| 04:18.23 | therealtbe | allow 1% partials or something like this |
| 04:18.28 | Flash | in this case, it's not an issue |
| 04:18.32 | Patlabor221 | therealtbe, prone to failure |
| 04:18.37 | Patlabor221 | and it's all going away |
| 04:18.39 | Flash | he sends no partials, because it's a very small packet |
| 04:18.51 | Patlabor221 | yeah, I can block that packet, and cache the DNS |
| 04:19.00 | Patlabor221 | that should minimize the effect |
| 04:19.16 | therealtbe | dns is the only thing that causes lag; if it's cached he should be able to send a billion UDPLinkEst messages who cares |
| 04:19.18 | Patlabor221 | then we can look at a system to check for multiple reconnects per second and auto block those IPs |
| 04:19.29 | Patlabor221 | therealtbe, he'll find another message to send |
| 04:19.38 | therealtbe | so what |
| 04:19.42 | Patlabor221 | bzflag was not designed for this |
| 04:19.45 | therealtbe | if it doesnt have a way to cause lag the server will just say ok ok ok ok ko |
| 04:19.51 | Flash | I'll look back at what I mistakenly committed; I had most of that, although the intervals may have been wrong |
| 04:19.55 | therealtbe | right now it says ok... dns lookup.... ok .... dns lookup.... |
| 04:20.21 | therealtbe | also we could just script looking at nestat for large Recv-Q and add to firewall and put it in crontab |
| 04:20.25 | Patlabor221 | therealtbe, the dns lookup is supposed to be async, so it is no blocking, but caching it will help |
| 04:20.37 | Patlabor221 | but he can also just flood the system with processing messages |
| 04:23.37 | therealtbe | sudo iptables -A INPUT -s `netstat -atn | awk ' $2 > 10000 ' | awk '{print 5}' | awk -F: '{print $1}` -j DROP |
| 04:23.45 | therealtbe | put that in the crontab every 15s |
| 04:23.54 | therealtbe | won't have any more problems from that fuck |
| 04:24.15 | therealtbe | is smarter than hacker in one liner |
| 04:24.17 | Flash | you realize that is completely outside the game? |
| 04:24.26 | Flash | talk to the server owner |
| 04:25.21 | Patlabor221 | therealtbe, language please |
| 04:25.38 | Patlabor221 | well really this should be blockable at the firewall level :) |
| 04:25.45 | Patlabor221 | thats the right place to do it |
| 04:27.23 | therealtbe | yeah any decent firewall should catch this. |
| 04:27.52 | Flash | that is no doubt true, and still outside the scope of the game |
| 04:28.02 | therealtbe | fine how about this for scope: |
| 04:28.07 | Flash | whether the server owner has a firewall remains to be seen |
| 04:28.38 | therealtbe | the server should not respond to >1000 MsgUDPLinkEstablished messages from one IP address and try to execute reverse dns lookups thus lagging everyone who is actually conneted and playing |
| 04:28.41 | therealtbe | is that within scope |
| 04:28.51 | Flash | yes |
| 04:29.11 | Flash | my first attempt at that failed as already described. |
| 04:29.14 | Flash | working on it |
| 04:29.42 | therealtbe | can we just turn off DNS ? so what we can't hostban, he is using proxys anyways |
| 04:30.00 | therealtbe | the dns lookup could happen one time on hostban, instead of in response to client conneting |
| 04:30.03 | Flash | since he's not reconnecting, DNS is not the real problem |
| 04:31.05 | therealtbe | what i'm saying is, can you amke hte response to multiple UDPLinkEstablished messages instantaneous from the server side? |
| 04:31.10 | therealtbe | this way there should be no way to create lag |
| 04:31.19 | Flash | it can be done |
| 04:31.22 | therealtbe | the messages are small and the server is fast, it should just process them super fast |
| 04:31.24 | Patlabor221 | therealtbe, the server should not accept a UDP accept message untill i has logged in |
| 04:31.26 | Patlabor221 | that's easy |
| 04:31.30 | Patlabor221 | I'm fixing that now |
| 04:31.33 | therealtbe | it is a TCP message |
| 04:31.37 | Patlabor221 | yes I know |
| 04:31.47 | Patlabor221 | the message conrolls the UDP mode ;) |
| 04:32.19 | Patlabor221 | I will prevent that message from being accepted and drop the connection |
| 04:32.44 | Patlabor221 | and I can cache the DNS |
| 04:32.57 | therealtbe | DNS isn't really needed until/if hostban time |
| 04:33.15 | Patlabor221 | hostban starts at the connect stage |
| 04:33.17 | therealtbe | no reason to do that in the beginning handshake |
| 04:33.25 | Flash | you are like a dog with a bone, aren't you? DNS is not the issue |
| 04:33.32 | therealtbe | sorry what is the issue then |
| 04:33.33 | Patlabor221 | actualy there is, you don't want ANY messages from a banned person |
| 04:33.56 | Patlabor221 | the issue is our server code |
| 04:34.18 | Flash | the issue is that he is sending messages that are handled at the bzfs protocol level faster than the server can handle them |
| 04:34.23 | Flash | including logging them |
| 04:35.11 | therealtbe | oh, so it is the IO load on the logging system that slows everybody down? |
| 04:35.17 | Flash | probably |
| 04:35.35 | therealtbe | well that's easy to prove, just turn off logging |
| 04:35.35 | Flash | the biggest failure in my initial implementation was allowing 1 connection every 1.5 seconds |
| 04:35.48 | Flash | apparently, the list server and strayer connect more frequently than that |
| 04:35.53 | Patlabor221 | well the game and protocoll level run in the same thread, so I don't see how one is "faster" then the other |
| 04:36.17 | Patlabor221 | flash yeah you can't use 1 |
| 04:36.28 | Patlabor221 | I'd block when it was 10 in 5 seconds or something |
| 04:36.36 | Patlabor221 | that should be tollerant of valid reconnects |
| 04:36.54 | Flash | I figured 1.5 was valid for a human clicking. I figured wrong |
| 04:37.10 | Flash | (actually, I didn't know about the list server or strayer's connections) |
| 04:37.13 | Patlabor221 | the problem is you did "one" :) |
| 04:37.19 | Patlabor221 | you need to allow for a few dupes |
| 04:37.23 | Patlabor221 | just not a lot |
| 04:37.39 | Flash | ok |
| 04:38.00 | Flash | I'm not pursuing fixing this ATM if you are on it, Patlabor221 |
| 04:38.32 | Patlabor221 | I'll do the blocking of the message |
| 04:38.45 | Patlabor221 | I don't want to implement a full blown blocking system |
| 04:38.58 | Patlabor221 | well a blocking system based on reconnects |
| 04:39.14 | Flash | most of it is here https://github.com/jwmelto/bzflag-import-3/commit/beb7a3f7636e61b3e3bd3c144cee3161c4408c26 |
| 04:39.25 | Flash | as you point out, it was overly restrictive |
| 04:39.38 | Patlabor221 | I mostly don't want to do t because I know we are probalby changing networking backends |
| 04:40.15 | Flash | for 2.6, I agree. However, since this is a present and serious issue on the most popular server, it seems to merit attention to me |
| 04:40.35 | Patlabor221 | yeah hence why I'm willing to block the messasge |
| 04:40.52 | Patlabor221 | but I think that a conenction based ban sytem needs research |
| 04:41.03 | Patlabor221 | actualy the connection based system can be prototyped in a plugin |
| 04:41.10 | Patlabor221 | that'd be the place to start with that |
| 04:41.18 | Patlabor221 | then get some servers to test it |
| 04:46.23 | Patlabor221 | god I hate ares |
| 04:48.32 | therealtbe | yeah, everytime I compile a new version. |
| 04:48.49 | therealtbe | then I have to go look in my notes for the right options |
| 04:49.40 | Flash | do we know how often the list server and strayer connect |
| 04:50.59 | therealtbe | what ports do those use |
| 04:51.18 | Flash | they connect to the server like anyone else |
| 04:54.23 | Patlabor221 | yeah they have to use the game port, they are quering the game |
| 04:54.42 | Patlabor221 | the list server backconnections everytime the server pushes an update |
| 04:54.46 | therealtbe | i've seen the list serv connect like twice in the last minute |
| 04:54.51 | Patlabor221 | it's t make sure they have the reported port open |
| 04:54.57 | therealtbe | I haven't seen a strayer connection but it might be faster than my refresh |
| 04:54.59 | Patlabor221 | depends how often he server updates |
| 04:55.16 | Patlabor221 | stats tends to connect every 5 min |
| 04:55.29 | Patlabor221 | but I think it makes multiple connections when it polls |
| 04:55.35 | Flash | apparently, the list server and strayer connected faster than 1.5 secs, since I locked them out |
| 04:55.50 | Patlabor221 | I belice strayer does 2 |
| 04:56.14 | Patlabor221 | one for game stats one for plaer stats |
| 04:57.05 | Patlabor221 | the game client knows the address of the list server it contacted, so that can be excluded pretty easy |
| 04:58.51 | therealtbe | listserv appears top open up to 4 connections at a time |
| 04:59.16 | Flash | hence why we have to allow more than one connection/IP |
| 04:59.47 | therealtbe | or just white list it as a special case |
| 05:00.02 | therealtbe | or revisit the listserv design... |
| 05:05.10 | Patlabor221 | well the game and list server need to be rewriten |
| 05:05.15 | Patlabor221 | that's a known issue |
| 05:05.27 | Patlabor221 | but it's a huge task |
| 05:07.36 | therealtbe | if flash's thing works, just set it to 100 or some large number |
| 05:07.52 | therealtbe | but that will only work after someone recognizes the attack and starts killing his sockets |
| 05:09.16 | Patlabor221 | his "thing" should be tested in a plug-in first |
| 05:09.35 | therealtbe | meaning he will still be able to open the 5 original ports and flood the server, then only when someone notices and firewalls, it will rate limit his reconnections. |
| 05:09.36 | Patlabor221 | or at the very least a fork |
| 05:10.01 | Flash | I made a branch in my fork in anticipation of working on it |
| 05:10.37 | Patlabor221 | k |
| 05:10.49 | Patlabor221 | I still think you'll get more testing by doing it in a plug-in |
| 05:10.53 | Flash | I have learned my lesson (mostly) |
| 05:11.09 | Patlabor221 | won't be as efficent, but it'll be easier for owners to test and turn of if it fails |
| 05:12.01 | Patlabor221 | how do I get an int for the IP from a sockaddr_in ? |
| 05:12.07 | Patlabor221 | I need a key for my cache map |
| 05:12.13 | Flash | I used a string |
| 05:12.33 | Patlabor221 | based on what? |
| 05:12.54 | Flash | <PROTECTED> |
| 05:13.04 | Flash | method on Address |
| 05:13.14 | Flash | <PROTECTED> |
| 05:13.51 | Patlabor221 | ahh we have a utility |
| 05:14.12 | Flash | clever people, those Chinese |
| 05:14.22 | therealtbe | inet_ntoa |
| 05:20.01 | Not-311d | [02bzflag-import-3] 07JeffM2501 pushed 031 commit to 03v2_4_x [+0/-0/±1] 13http://git.io/XV4Ylg |
| 05:20.02 | Not-311d | [02bzflag-import-3] 07JeffM2501 0361eac31 - double's ain't floats (fixes warnings) |
| 05:23.44 | Patlabor221 | I've writen the code for the cache and the message block, I just need to test it |
| 05:41.42 | *** join/#bzflag Patlabor221 (~JeffM@2605:e000:7d02:c700:7ca9:d7ae:c4f4:8b32) |
| 05:51.46 | Patlabor221 | ok, basic testing of the caching and the blocking seems to be working. |
| 05:53.37 | Not-311d | [02bzflag-import-3] 07JeffM2501 pushed 031 commit to 03v2_4_x [+0/-0/±5] 13http://git.io/T82FHw |
| 05:53.38 | Not-311d | [02bzflag-import-3] 07JeffM2501 03b225e8b - Cache reverse DNS lookups (note we may need to flush the cache every so often) Let NetHandler know if we are allowing UDP negotiation or not. Only allow NetHandler to do UDP negotiation after we have a full join |
| 06:02.27 | Patlabor221 | that should help somewhat |
| 06:47.02 | *** join/#bzflag Deadalot (~philb@67.8.155.99) |
| 06:48.11 | Deadalot | Don't know if it made it over here or not. The hacker that has been plagueing all of the bzflag servers posted his script. http://pastebin.com/2xe5ejNm |
| 06:49.16 | allejo | yea. pat just committed some changes and i'm testing a firewall rule |
| 06:49.50 | Deadalot | Cool |
| 06:50.18 | Deadalot | Good on ya. Butting out now. |
| 06:53.10 | Patlabor221 | the fix is highly expirmental, it needs testing |
| 07:11.27 | *** join/#bzflag Snake12534 (uid45641@gateway/web/irccloud.com/x-iiydzwwbscpipsla) |
| 07:13.11 | Snake12534 | allejo: I decided to become a full time GD. |
| 07:14.29 | allejo | sets off a single fire cracker |
| 07:14.44 | Patlabor221 | graphic desgin is a fine field |
| 07:15.45 | allejo | he means game dev |
| 07:15.52 | Patlabor221 | I gathered that |
| 07:16.02 | allejo | and just making sure |
| 07:16.04 | allejo | ah* |
| 07:22.07 | Snake12534 | For that reason, I am not implementing bzflagRPG. |
| 07:23.55 | allejo | tears up |
| 07:25.01 | Snake12534 | Sorry allejo. |
| 07:25.14 | Snake12534 | Even I am dropping out of school. |
| 07:25.24 | Patlabor221 | why would you do that? |
| 07:25.25 | Snake12534 | (I'm not) ^ |
| 07:25.38 | Patlabor221 | stay in school kids, it's worth it |
| 07:25.46 | Snake12534 | Full time student :o |
| 07:25.47 | Snake12534 | Scary |
| 07:26.21 | Patlabor221 | not really |
| 07:27.05 | Patlabor221 | that reminds me, I should proalby backup my final proejct somewhere |
| 07:27.35 | allejo | would be a good idea |
| 07:27.41 | allejo | throws a butter lettuce party |
| 07:27.46 | allejo | it's locally grown too! |
| 07:27.54 | Patlabor221 | heh |
| 07:28.31 | Patlabor221 | thing ended up like 5 gigs |
| 07:28.58 | Patlabor221 | it's weird to not have to go to school tomorow like everyone else |
| 07:30.31 | allejo | lucky |
| 07:30.41 | Patlabor221 | no, just done |
| 07:30.54 | Patlabor221 | there is always an end date |
| 07:31.00 | Patlabor221 | one day you'll get t it |
| 07:31.13 | Patlabor221 | just don't wait till your over 40 |
| 07:32.10 | allejo | meh |
| 07:32.19 | Snake12534 | A winters day |
| 07:32.24 | Snake12534 | In a deep and dark December |
| 07:32.32 | Snake12534 | I am old |
| 07:32.36 | Snake12534 | Geeking through my window |
| 07:32.42 | Snake12534 | I am a rock. |
| 07:32.50 | Patlabor221 | I'd do it again if I had to |
| 07:32.56 | Patlabor221 | it's been worth it at work |
| 07:32.57 | Snake12534 | really? |
| 07:33.01 | Snake12534 | What degree? |
| 07:33.05 | Patlabor221 | more resposnibility |
| 07:33.15 | Patlabor221 | They are sending me to Germany in January |
| 07:33.19 | Patlabor221 | Computer Science |
| 07:33.25 | Snake12534 | Nice! |
| 07:33.32 | Snake12534 | Hows cs going for you? |
| 07:33.39 | Patlabor221 | I am done with it |
| 07:33.44 | Snake12534 | ? |
| 07:33.50 | Snake12534 | Quit? |
| 07:33.50 | allejo | oh great, now you're going to sound like you're yelling all the time, jeff |
| 07:34.02 | Patlabor221 | no, I finished |
| 07:34.08 | Patlabor221 | all clases done |
| 07:34.12 | Snake12534 | Oh |
| 07:34.12 | Patlabor221 | capstone complete |
| 07:34.15 | Snake12534 | I meant your job |
| 07:34.30 | Snake12534 | your job based on your CS degree |
| 07:34.56 | Patlabor221 | I work on software that controlls industrial robotics for manufacturing |
| 07:35.03 | Snake12534 | sounds sick |
| 07:35.10 | Snake12534 | you like the job? |
| 07:35.23 | Patlabor221 | I've been there for 20 years, so I'd hope I do |
| 07:35.49 | Snake12534 | When you wake up in the morning |
| 07:36.00 | Snake12534 | And you are getting ready for your job.. Are you happy? |
| 07:36.15 | Patlabor221 | sure, but I don't define myself through my work |
| 07:36.19 | Patlabor221 | it's a job, it pays the bills |
| 07:36.24 | Snake12534 | hm |
| 07:36.28 | Patlabor221 | it's full of nice people |
| 07:36.34 | Snake12534 | I have a slightly different perspective |
| 07:36.42 | Snake12534 | But that's cool :) |
| 07:36.51 | Patlabor221 | you are young, you will learn |
| 07:36.56 | Snake12534 | haha |
| 07:37.03 | Snake12534 | How do you know my age? :) |
| 07:37.13 | Patlabor221 | your history here is not unknown |
| 07:37.18 | Snake12534 | oh |
| 07:37.23 | Patlabor221 | nor is mine |
| 07:37.25 | Snake12534 | I've never met you before |
| 07:37.29 | Flash | chuckles |
| 07:37.37 | Snake12534 | nor flash |
| 07:37.45 | Patlabor221 | I have only met a few bzflag players in real life |
| 07:37.51 | Patlabor221 | less then 10 I belive |
| 07:37.59 | Snake12534 | In the game bzflag I meant |
| 07:38.06 | Snake12534 | I've only met one.. My brother |
| 07:38.16 | Patlabor221 | that does not quite count :) |
| 07:38.31 | Snake12534 | I've always had a dream of a BZ-con 2014 |
| 07:38.36 | Patlabor221 | I did get one of them a job at my work |
| 07:38.40 | Snake12534 | Where all BZ players meet |
| 07:38.43 | Snake12534 | Nice |
| 07:38.48 | Snake12534 | That's always nice |
| 07:39.13 | Patlabor221 | hmm no, that living room was quite full, it's probably more then 10 people now |
| 07:39.43 | Snake12534 | Living room? Did you host an event? |
| 07:39.47 | Patlabor221 | no |
| 07:40.02 | Snake12534 | Allejo: I am finally a boss! |
| 07:40.12 | Patlabor221 | just a meeting after a sigraph |
| 07:40.27 | Snake12534 | I am getting a 70/30 split for doing 10% as much work - the power |
| 07:40.42 | Snake12534 | What is your bzflag callsign? |
| 07:41.14 | Patlabor221 | same as my forum account, most people play registered |
| 07:41.20 | Patlabor221 | that is what the survey said, |
| 07:41.27 | Patlabor221 | over 80% IRRC |
| 07:42.02 | Snake12534 | What is iRrc |
| 07:42.11 | Patlabor221 | If I Recall Correctl |
| 07:42.16 | allejo | iirc* |
| 07:42.18 | Flash | correctly |
| 07:42.27 | Flash | is a grammar nazi |
| 07:42.27 | Patlabor221 | someting like that |
| 07:42.39 | Patlabor221 | tis late, fingers tired |
| 07:42.48 | allejo | Flash, but that was spelling |
| 07:42.53 | Flash | both |
| 07:42.57 | allejo | :p |
| 07:43.16 | Patlabor221 | 83% by current counts |
| 07:43.40 | Patlabor221 | only 3 out of 63 respondents oppose registration |
| 07:43.42 | Snake12534 | I wish the good outweighed the bad in life. |
| 07:43.55 | Patlabor221 | it does, you just have to look for it |
| 07:44.11 | Patlabor221 | or work for it |
| 07:44.29 | Flash | blanket statements... evil generally prevails, but that does not mean that there is not good to be found at every turn |
| 07:44.40 | Snake12534 | Patlabor221: The question was worded awkwardly IMO. It asked, "I'm already registered, it shouldn't be a problem!" |
| 07:44.42 | thrakt | "grammar nazi"... is that why they burned all those books? |
| 07:44.44 | Patlabor221 | Life is what we choose to make it |
| 07:45.05 | Flash | thrakt: only those with bad grammar! |
| 07:45.09 | Patlabor221 | Snake12534, it specificly said "I play registered all the time" |
| 07:45.14 | thrakt | :D |
| 07:45.15 | Patlabor221 | that should be fairly clear |
| 07:45.21 | Snake12534 | My memory is vague ;/ |
| 07:45.38 | Patlabor221 | then best not to comment onn vaugalities ;) |
| 07:45.41 | Snake12534 | I oppose registration, but I play register all the time. |
| 07:45.55 | Snake12534 | I play unregistered for years to test out the game. |
| 07:45.58 | Patlabor221 | then if we make that change you'll have to make a choice |
| 07:46.22 | Snake12534 | Played * |
| 07:46.51 | Patlabor221 | facebook registration is pretty eays anyway |
| 07:46.59 | Flash | Snake12534: the case against registration is weak |
| 07:47.06 | thrakt | not FB please |
| 07:47.14 | thrakt | I will pay you. |
| 07:47.28 | Patlabor221 | thrakt, ok, then myspace |
| 07:47.33 | thrakt | LOL |
| 07:47.34 | Patlabor221 | or excite |
| 07:47.40 | Flash | ApplePay |
| 07:47.41 | Patlabor221 | maybe lycos has an API |
| 07:47.41 | Snake12534 | allejo: I wanted to make Rico world in to a league via timer and adverise groups implementation |
| 07:47.56 | Patlabor221 | naw my phone doesn't do apple pay |
| 07:48.01 | Patlabor221 | I am not hip enough |
| 07:48.06 | Snake12534 | Rico world is ridiculously fun.. Played it for a long time, great CTF |
| 07:48.13 | Snake12534 | Others agreed |
| 07:48.15 | Flash | and you don't have a big enough pocket |
| 07:48.17 | thrakt | I like the idea... but I've been able to get away with only Google so far |
| 07:48.29 | Flash | Google is the devil incarnate |
| 07:48.35 | thrakt | Jah. |
| 07:48.36 | Patlabor221 | yeah, I will not be gettig another iPhone when this one is up |
| 07:48.46 | Patlabor221 | the base 6 is too big |
| 07:48.56 | Flash | forget about the NSA ... Google TELLS you they are reading your e-mail and sending you ads based on it |
| 07:49.14 | Patlabor221 | thrakt, we won't be using a third party, the question was a curiosity to see how many would directly oppose it |
| 07:49.15 | thrakt | +1 for disclosure? |
| 07:49.19 | Snake12534 | Flash: how? |
| 07:49.34 | Flash | if you use gmail, it's in their terms and conditions |
| 07:49.36 | Snake12534 | Flash: how do they read millions of emails? |
| 07:49.41 | Patlabor221 | I find the google adds quite well targeted |
| 07:49.45 | Flash | lots of computers |
| 07:49.51 | Patlabor221 | in lot of raks |
| 07:49.53 | thrakt | Patlabor221: cool. BTC IRC community uses PGP keys/circle of trust... |
| 07:49.59 | allejo | Snake12534, uh go for it if you want. i don't think i ever publicly released that map so if you get it organize then i will |
| 07:50.14 | Patlabor221 | thrakt, we are a silly tank game, no need for that |
| 07:50.30 | Patlabor221 | you should not be sending sensitive info over a silly tank game :) |
| 07:50.31 | Snake12534 | allejo: I was playing it last night? |
| 07:50.47 | Flash | thrakt: I've proposed PKI authentication before. He always shoots me down |
| 07:50.58 | thrakt | Just sayin... if there was a plan to enforce reg, etc... that would be an "in house" solution |
| 07:51.15 | Snake12534 | allejo: well the thing is, I dont have a library folder in users -> Brandon |
| 07:51.17 | Patlabor221 | it's just a bit more complicated then we need, it's not a flawed system, it's just not the right one fo what we need. |
| 07:51.18 | allejo | Snake12534, i'm hosting it yes. don't know if i released it for others to use (i.e. licensed it) |
| 07:51.21 | thrakt | Flash: I guess I missed those debates |
| 07:51.28 | Snake12534 | I am using constitution's |
| 07:51.30 | Patlabor221 | thrakt, the one we have now..... |
| 07:51.31 | Flash | debate is a strong term |
| 07:51.50 | Patlabor221 | yeah it was not a debate |
| 07:52.02 | Snake12534 | I am using Constitution Nogamma version. |
| 07:52.09 | Patlabor221 | he offered a sugestion, had good information on it, but it has not shown to be a goodfit |
| 07:52.28 | thrakt | All good. |
| 07:52.43 | Flash | one of the big issues is that PKI and UDP don't go well together |
| 07:52.45 | Patlabor221 | he isn't the frist to sugest it, we even paid someone to research authenticaiton systems for our needs and that was included |
| 07:52.53 | Snake12534 | Meaning, I can't access the worlds. |
| 07:53.02 | Patlabor221 | yeah when you arn't coneciton oriented you have to get fancy |
| 07:53.25 | Patlabor221 | well google paid them no use actualy |
| 07:53.34 | Patlabor221 | twice even I think |
| 07:54.53 | thrakt | Pat, Flash: Lost in cross-channel there. Understand that PKI, et al auth systems not likely. |
| 07:55.55 | thrakt | Anyone played with LibreSSL yet? |
| 07:56.07 | Flash | what's to play with? |
| 07:56.20 | thrakt | Dunno.. OpenSSL replacement/rewrite |
| 07:56.32 | thrakt | For OBSD systems... |
| 07:56.57 | Flash | No motivation to spend time on a replacement to something that works fine |
| 07:57.26 | Patlabor221 | but but but cryptography SOLVES EVERYTHING!!!!!!! ;) |
| 07:57.27 | Flash | which is most of Patlabor221's argument against changing auth ... |
| 07:58.00 | Patlabor221 | well there are changes I want to auth, I'm just happy with the token system, it is secure enough and easy to maintain |
| 07:58.19 | Patlabor221 | our entire backend needs to be decoupled from phpbb |
| 07:58.40 | Patlabor221 | token + https should be fine for public servers |
| 07:58.52 | Constitution | we've been saying that almost since it was first coupled with phpbb :) |
| 07:58.54 | Patlabor221 | I would not require auth for private or lan servers |
| 07:59.10 | Patlabor221 | Constitution, yup, none of us did that coupling |
| 07:59.15 | Patlabor221 | Tim did it |
| 07:59.27 | Patlabor221 | but nobody has made a user manager |
| 07:59.36 | Patlabor221 | so we don't have anything to replace it with |
| 07:59.52 | Constitution | wasn't that a GSoC project? |
| 07:59.53 | Patlabor221 | and now anything we do would have to have a migration plan for the userDB |
| 08:00.10 | Patlabor221 | Constitution, twice, the LDAP implementation, never got finished |
| 08:00.28 | Patlabor221 | and never had a user manager, assumed ldap admin via wiki or forums |
| 08:01.37 | Patlabor221 | anything we do will need some sort of user friendly front end, the backend doesn't really matter, that's the main reason I am not fond of changing the auth transport system now |
| 08:02.14 | Patlabor221 | once we have a useable front end, then would be the time to find the best backend and transport layer for our needs |
| 08:02.37 | thrakt | Sounds like a typical Friday night out to me. |
| 08:02.55 | Snake12534 | Sunday * |
| 08:02.58 | Patlabor221 | then please write one up, we have several pages of specification on the wiki IIRC |
| 08:03.41 | thrakt | I was being punny: "once we have a useable front end, then would be the time to find the best backend and transport" |
| 08:03.42 | thrakt | <PROTECTED> |
| 08:04.13 | thrakt | Fail. |
| 08:04.27 | thrakt | If you ewre to change the bb what would you change it to? |
| 08:04.29 | Patlabor221 | IRC is not an awesome protocoll :) |
| 08:04.48 | Patlabor221 | our initaly plan is to just have a standalone auth db in mysql |
| 08:05.01 | Patlabor221 | then have the forums auth against that as well as the game |
| 08:05.10 | Patlabor221 | we can write an auth plugin to phpbb |
| 08:05.20 | Patlabor221 | then we'd remove user management from phpbb |
| 08:05.35 | thrakt | Meaning groups & such? |
| 08:05.41 | Patlabor221 | yeah |
| 08:05.50 | Patlabor221 | and allow nested groups and a few new features |
| 08:06.11 | Patlabor221 | then it'd all be decoupled and we could fiddle with each bit independently |
| 08:06.57 | Patlabor221 | but once we had a seperate user DB, then I'd push for an official leauge solution that integrated with that new userDB |
| 08:07.12 | Patlabor221 | get all our services going, then we'll have a good idea of what our needs are |
| 08:08.23 | thrakt | Cool. Where does that rank on the "Likely to get done" list? |
| 08:08.33 | Patlabor221 | same as all the others, low |
| 08:08.41 | Patlabor221 | someone has to want to do it |
| 08:08.47 | Patlabor221 | and have the ability |
| 08:08.57 | Patlabor221 | so far none have stepped forwared to build the user system |
| 08:09.04 | Patlabor221 | a few have started but none ever finished |
| 08:10.09 | thrakt | Would hackathons require too much of existing devs? |
| 08:10.34 | Patlabor221 | I have never seen them work out |
| 08:20.42 | Patlabor221 | the bzflag codebase is not very condusive to changes |
| 08:21.03 | Patlabor221 | and people tend to work on things they want to work on, not things that need to get done |
| 08:21.27 | Patlabor221 | at least that's what I've seen in my experience |
| 08:33.50 | Patlabor221 | makes sleep |
| 08:34.25 | Flash | when you make it, can I have it? |
| 08:38.17 | allejo | mmm sleep |
| 09:11.59 | *** join/#bzflag DTRemenak (~DTRemenak@about/essy/CrazyCoder/DTRemenak) |
| 09:11.59 | *** mode/#bzflag [+v DTRemenak] by ChanServ |
| 11:33.35 | *** join/#bzflag Patlabor221 (~JeffM@2605:e000:7d02:c700:c108:c3fa:1d38:b041) |
| 11:39.46 | blast007 | therealtbe: the list server opens a connection to a game server every time an 'ADD' message is sent, which is when a server starts, when a player joins or leaves, and every so often if nobody has join or left recently (to keep the server on the list) |
| 14:30.41 | *** join/#bzflag DTRemenak|RDP (~DTRemenak@about/essy/CrazyCoder/DTRemenak) |
| 14:30.41 | *** mode/#bzflag [+v DTRemenak|RDP] by ChanServ |
| 14:44.11 | *** join/#bzflag Patlabor221 (~JeffM@2605:e000:7d02:c700:b4f4:185a:b8cc:e714) |
| 15:07.41 | *** join/#bzflag zlsa (~zlsa@static-205-157-147-198.ded.unwiredbb.net) |
| 15:07.41 | *** join/#bzflag zlsa (~zlsa@unaffiliated/zlsa) |
| 15:33.58 | *** join/#bzflag Snake12534 (uid45641@gateway/web/irccloud.com/x-vyxefdrvrdubxcbd) |
| 15:39.20 | *** join/#bzflag spldart (~james@bzflag/contributor/spldart) |
| 15:39.20 | *** mode/#bzflag [+v spldart] by ChanServ |
| 15:42.53 | spldart | a bit off topic but lots of users here. Recommend a good emulator to run windows xp in debian stable please? |
| 15:52.35 | brad | virtualbox? |
| 15:52.53 | spldart | k |
| 15:52.56 | spldart | thanks |
| 16:18.11 | *** join/#bzflag zlsa (~zlsa@static-205-157-147-198.ded.unwiredbb.net) |
| 16:18.11 | *** join/#bzflag zlsa (~zlsa@unaffiliated/zlsa) |
| 16:34.11 | *** join/#bzflag TimRiker (~TimRiker@bzflag/projectlead/TimRiker) |
| 16:34.11 | *** mode/#bzflag [+o TimRiker] by ChanServ |
| 16:39.47 | *** join/#bzflag zlsa (~zlsa@static-205-157-147-198.ded.unwiredbb.net) |
| 16:39.58 | *** join/#bzflag zlsa (~zlsa@unaffiliated/zlsa) |
| 17:06.01 | *** join/#bzflag Patlabor221 (~JeffM@107-209-61-107.lightspeed.irvnca.sbcglobal.net) |
| 17:13.48 | *** join/#bzflag zlsa (~zlsa@static-205-157-147-198.ded.unwiredbb.net) |
| 17:13.48 | *** join/#bzflag zlsa (~zlsa@unaffiliated/zlsa) |
| 17:14.33 | blast007 | Patlabor221: I don't think DNS was really an issue anyways |
| 17:15.02 | Patlabor221 | probalby not, but caching seemed like a good idea anyway |
| 17:15.09 | blast007 | the only thing that the sending of that packet seemed to trigger based on what I'm seeing in the code is setting a boolean value to true and printing a log message |
| 17:15.11 | Patlabor221 | want me to tie it to a flg or something? |
| 17:15.34 | blast007 | does ares have it's own cache? |
| 17:15.39 | Patlabor221 | dunno |
| 17:15.39 | blast007 | its* |
| 17:15.44 | Patlabor221 | I loathe ares ;) |
| 17:16.02 | Patlabor221 | so it's just flooding out the entire message buffer? |
| 17:17.04 | blast007 | the script just sends repeated messages in an infinite loop, each which triggers a log message, so it's wasting network bandwidth, CPU cycles, and disk IO |
| 17:17.23 | Patlabor221 | then yeah that's firewall blockage material |
| 17:17.59 | blast007 | has mentioned it's probably something that a Snort ruleset could take care of |
| 17:18.27 | Patlabor221 | yeah thats why I said the proper fix is at the firewal |
| 17:18.35 | Patlabor221 | I'll pull the cache, I am not confident of it |
| 17:18.50 | Patlabor221 | I do't mind leaving the UDP block as that may help prevent other exploits down the line |
| 17:19.22 | blast007 | if someone is worried about DNS issues, they could just run a local DNS cache |
| 17:19.33 | Patlabor221 | yeah |
| 17:19.40 | Patlabor221 | and I bet your right, it'll be cached at the lib level |
| 17:20.02 | blast007 | well, it *might* be... we've had issues before where the list server IP was cached forever... |
| 17:20.09 | Patlabor221 | heh |
| 17:20.22 | blast007 | made for interesting times when we had to move to a new server |
| 17:20.51 | Patlabor221 | yeah and curl does use ares |
| 17:21.50 | Patlabor221 | I will leave one of my changes in for AresHandler, changing from a malloc/freed char* to a std::string for hostname storage |
| 17:22.13 | blast007 | k |
| 17:23.16 | Patlabor221 | would you be against the inital magic read checking for more data to see if there are invalid messages and just not accepting the socket? |
| 17:23.19 | blast007 | ah, turns out that it's *cURL* that has a DNS cache ;D |
| 17:23.24 | Patlabor221 | that'd prevent the loging from hitting |
| 17:24.11 | blast007 | are we sure that message never happens before MsgEnter? |
| 17:24.49 | Patlabor221 | not in legit code and not in any of my testing |
| 17:25.10 | Patlabor221 | and from what I can tell, we will resend it if it fails |
| 17:29.44 | Patlabor221 | nope I lie, it could come before and only once |
| 17:29.45 | Patlabor221 | fixing |
| 17:30.40 | blast007 | so it will only come once? we could kick if it comes twice, then. |
| 17:30.58 | Patlabor221 | it still comes after another message |
| 17:31.05 | Patlabor221 | I just have to chagne where I set my flag |
| 17:31.13 | Patlabor221 | it can not be the first message after magic |
| 17:31.31 | blast007 | okay |
| 17:32.33 | Patlabor221 | ahh we do send it twice... odd |
| 17:32.39 | blast007 | heh |
| 17:32.47 | Patlabor221 | in 2 differently named functions |
| 17:33.02 | blast007 | ahh.. legacy code.... ;D |
| 17:33.04 | Patlabor221 | we have a function named sendUDPlinkRequest |
| 17:33.16 | Patlabor221 | yet in the ServerLink constructor we send the message directly |
| 17:33.44 | blast007 | do we just say "screw it" and just fix it properly for 2.6? |
| 17:33.51 | Patlabor221 | still looking |
| 17:33.55 | Patlabor221 | but maybe |
| 17:36.10 | Patlabor221 | ok, we don't send it untill AFTER we have received our connection responce |
| 17:36.28 | Patlabor221 | so it can never be part of the same packet that has the magic |
| 17:37.17 | Patlabor221 | so I should be able to check for that |
| 17:37.23 | Patlabor221 | no message can be part of the magic |
| 17:37.31 | Patlabor221 | the buffer must be empty at that point |
| 17:38.39 | Patlabor221 | so I'll see if I can block it at the accept level, that'd be before logging got it |
| 17:38.52 | Patlabor221 | it'll be at the level where we determine if it's BZFS or HTTP |
| 17:39.59 | therealtbe | how about generic prevention of logging repeat messages? |
| 17:40.25 | Patlabor221 | that's outside the scope of a small change for 2.4.x |
| 17:40.38 | therealtbe | Flash's hunch is that the logging of the thousands of messages is slowing down Disk IO and creating lag |
| 17:40.49 | therealtbe | generic rate limiter on logging frequency ? |
| 17:40.55 | therealtbe | something simple |
| 17:41.05 | Patlabor221 | yeah we know how to do it, it's still outside the scope |
| 17:41.24 | Patlabor221 | even the simplest rate limiter is a sizeable proejct that could break valid connections |
| 17:41.34 | Patlabor221 | the firewal is where that should be blocked |
| 17:41.54 | therealtbe | I agree any decent firewall should block this |
| 17:42.12 | therealtbe | however, no offense but nobody has even established the actual method of causing lag |
| 17:42.32 | therealtbe | Flash's hunch seems like a good place to start, so making sure the logging doesn't bog the server could be a quick fix |
| 17:42.35 | Patlabor221 | just fill the message buffer, it'll make lag |
| 17:42.41 | Patlabor221 | bzfs runs in a SINGLE thread |
| 17:42.59 | Patlabor221 | making sure logging dosn't bog the server is not a trivial problem to solve |
| 17:43.03 | Patlabor221 | given how bzflag is writen |
| 17:43.30 | therealtbe | it seems like you could add a counter to how many messages per second, and then just stop logging them if it goes over a set rate. |
| 17:43.58 | therealtbe | it would require a counter, a set variable, and an if statement... |
| 17:44.24 | therealtbe | I am happy to see the activity around this but I hope it actually solves the problem, that's all |
| 17:44.25 | Patlabor221 | logging is not on by default |
| 17:44.35 | Patlabor221 | and it's more then just a single variable |
| 17:44.59 | Not-311d | [02bzflag-import-3] 07JeffM2501 pushed 031 commit to 03v2_4_x [+0/-0/±3] 13http://git.io/4mj-vQ |
| 17:45.00 | Not-311d | [02bzflag-import-3] 07JeffM2501 0340013d4 - remove DNS cache, it should happen at the ares or system level Remove UDP blocking until MsgEnter, it is valid to send it. Will look into other methods. |
| 17:47.07 | Patlabor221 | blast007, want to do a test for me? |
| 17:47.39 | Patlabor221 | or anyone with a public server? |
| 17:50.23 | *** join/#bzflag TimRiker (~TimRiker@bzflag/projectlead/TimRiker) |
| 17:50.23 | *** mode/#bzflag [+o TimRiker] by ChanServ |
| 17:53.38 | *** join/#bzflag zlsa (~zlsa@static-205-157-147-198.ded.unwiredbb.net) |
| 17:53.38 | *** join/#bzflag zlsa (~zlsa@unaffiliated/zlsa) |
| 17:56.15 | *** join/#bzflag Patlabor (~JeffM@76-218-80-21.lightspeed.irvnca.sbcglobal.net) |
| 17:59.50 | Patlabor221 | so in Git, how do I get my fork to be a copy of the master upstream repo again, getting rid of my changes to a branch? |
| 18:00.21 | blast007 | delete it and refork is how I would do it ;) |
| 18:00.38 | Patlabor221 | yeah that's my last resort, I am curious if there is a valid way |
| 18:00.45 | Patlabor221 | I'd like to do it on a per branch basis |
| 18:00.53 | Patlabor221 | I have things in 2_6 I don't want to blow out |
| 18:01.07 | therealtbe | git reset --hard HEAD |
| 18:01.19 | BulletCatcher | Rename your forked branch, then pull the remote branch again. |
| 18:01.48 | therealtbe | git reset && git checkout . && git clean -fdx |
| 18:02.04 | blast007 | therealtbe: I'm assuming he has locally commited changesets already.. |
| 18:02.23 | Patlabor221 | I have changes I pushed to my fork |
| 18:02.35 | Patlabor221 | looking into BulletCatcher's sugestion |
| 18:02.50 | therealtbe | best to just start clean |
| 18:03.26 | BulletCatcher | After you have renamed your forked branch you can cherry pick any commits you want to keep. |
| 18:03.55 | BulletCatcher | When you're really done with it, you can delete the renamed branch. |
| 18:05.06 | blast007 | therealtbe: and that wasn't Flash's hunch, it was something I mentioned that he just passed on. and it wasn't the only issue, just one possible piece that added up to causing issues. |
| 18:06.10 | BulletCatcher | A "git branch --unset-upstream" on the renamed branch will completely disconnect it from the remote repo. |
| 18:06.55 | blast007 | what if the remote remo (his fork) is what he wants to remove it from? |
| 18:07.05 | blast007 | repo* |
| 18:11.00 | Patlabor221 | having a branch with the murder machine isn't a bad idea |
| 18:12.07 | BulletCatcher | It looks like "git push origin :badbranch" will delete badbranch from a remote repo. |
| 18:12.29 | Patlabor221 | with the : ? |
| 18:12.35 | BulletCatcher | yup |
| 18:12.37 | blast007 | the actual changesets will remain, though, right? |
| 18:12.46 | Patlabor221 | so cryptic :) |
| 18:14.29 | BulletCatcher | If nothing else points to the tip of the deleted branch then its changesets will be invisible and eventually garbage collected. |
| 18:15.32 | blast007 | huh. so history can be removed without having to "force" it? |
| 18:16.37 | BulletCatcher | This is one way to do that. |
| 18:17.28 | blast007 | what if, for instance, the v2_4_x branch were deleted. Would all the changesets for 2.3/2.4 go away, or would they stay since v2_6_x is based on the v2_4_x branch? |
| 18:18.16 | BulletCatcher | I don't know. Make a temoporary clone and see for yourself. :-) |
| 18:18.25 | blast007 | hehe |
| 18:19.06 | BulletCatcher | Seriously, that is an excellent way to experiment with Git. |
| 18:20.41 | blast007 | yeah, I'll have to read up on the garbage collecting to know how to make it trigger so I can have some sort of controlled experiment |
| 18:20.50 | BulletCatcher | "git gc" |
| 18:21.21 | BulletCatcher | "git gc --prune=now" |
| 18:22.23 | BulletCatcher | Here's my final answer: git reflog expire --expire=now --all ; git gc --prune=now |
| 18:22.38 | BulletCatcher | I recommend doing that only on a throw-away clone. |
| 18:23.29 | blast007 | right |
| 18:23.43 | blast007 | but that's all for a local repo, not the remote copy |
| 18:24.16 | blast007 | "Github runs gc on a periodic basis. If you want have the files removed from the repo immediately, you'll have to delete and recreate the repo." (whatever "periodic" is) |
| 18:24.37 | BulletCatcher | Normally you don't have to worry about garbage collection. It'll just happen. |
| 18:25.34 | BulletCatcher | The risk of renaming/deleting a branch on a shared repo is that someone else will not get the memo and will push the old branch back to it from their own clone. |
| 18:26.32 | BulletCatcher | Each clone with write access to the shared repo must manually do the rename/delete for it to take permanent effect. |
| 18:27.18 | Patlabor221 | ok so I think I have my branch renamed |
| 18:27.30 | Patlabor221 | how do I sync my fork to the main bzflag again? |
| 18:27.51 | *** join/#bzflag khonkhortisan (~Khonkhort@2601:8:8600:9b0:4a5b:39ff:fe94:be23) |
| 18:28.56 | Patlabor221 | do I want to rebase? |
| 18:29.13 | BulletCatcher | Which way do you want the data to move? |
| 18:30.20 | Patlabor221 | I want bzflag-dev's v2_4_x branch to be in my fork |
| 18:30.27 | Patlabor221 | since I should not have a branch named that anymore |
| 18:31.09 | Patlabor221 | I snapped off the geno plugin into it's own branch, now I want my fork to be clean back to the main repo so I can do pull requests with out having the geno stuff |
| 18:33.25 | BulletCatcher | So far this is on your private clone of https://github.com/JeffM2501/bzflag-import-3, right? |
| 18:33.30 | Patlabor221 | yeah |
| 18:33.57 | Patlabor221 | I think I did it |
| 18:34.30 | Patlabor221 | I added master as a remote, rebased from master/2.4.x to my 2.4.x and then did a checkout/switch with force |
| 18:34.52 | Patlabor221 | or... I'm looking at the wrong thing on the website :) |
| 18:35.01 | BulletCatcher | That's roughly what I was going to suggest. |
| 18:35.12 | blast007 | don't forget the goat |
| 18:35.18 | Patlabor221 | does the website take a while to sync? |
| 18:35.38 | BulletCatcher | Not usually. Have you pushed to GitHub. |
| 18:35.39 | BulletCatcher | ? |
| 18:35.55 | Patlabor221 | several times |
| 18:37.00 | BulletCatcher | I don't see it there, either. |
| 18:37.25 | blast007 | the other option of course would be to delete and re-create the fork, and then push (or whatever the term would be) the changes you want to keep from your old local repo into your new local repo |
| 18:38.20 | Patlabor221 | ok, I got my branch to finaly show up |
| 18:38.57 | BulletCatcher | I see it too, now. |
| 18:38.58 | BulletCatcher | yay |
| 18:39.31 | BulletCatcher | Was there a time delay or did you do something different? |
| 18:39.42 | Patlabor221 | I did some things different |
| 18:41.57 | BulletCatcher | If you are satisfied, a "git remote remove" will undo the "git remote add" you did earlier and keep things less complicated. |
| 18:43.05 | Patlabor221 | all I'm stuck on now is getting bzflag-dev/v2_4_x into my fork |
| 18:43.26 | Patlabor221 | I got this to exist |
| 18:43.27 | Patlabor221 | https://github.com/JeffM2501/bzflag-import-3/tree/v_2_4_xGenoPlugin |
| 18:46.23 | Patlabor221 | got it :) |
| 18:46.37 | Patlabor221 | my GUI tool was still set to upstream |
| 18:47.00 | Patlabor221 | now, https://github.com/JeffM2501/bzflag-import-3/commits/v2_4_x doesn't have any of the geno stuff |
| 18:48.08 | Patlabor221 | thank you BulletCatcher that's what I was hoping I could do |
| 18:48.39 | Patlabor221 | then I don't have to loose the humorously named plug-in |
| 18:49.34 | *** join/#bzflag zlsa (~zlsa@static-205-157-147-198.ded.unwiredbb.net) |
| 18:49.34 | *** join/#bzflag zlsa (~zlsa@unaffiliated/zlsa) |
| 18:49.35 | BulletCatcher | You may need to do some extra work to have the v2_6_x branch have the correct merges from v2_4_x instead of v_2_4_xGenoPlugin. |
| 18:50.06 | Patlabor221 | if I have to I'll rename my 2_6, rebase it and cherry pick |
| 18:51.24 | BulletCatcher | I'm not sure that will preserve the merges. |
| 18:51.56 | Patlabor221 | the code I need to move over isn't quite mergable :) |
| 18:52.02 | Patlabor221 | it replaces the networking system |
| 18:52.20 | Patlabor221 | it'll get broken when you do the final repo import anyway |
| 18:52.25 | Patlabor221 | so i'm kinda waiting on that |
| 18:53.07 | BulletCatcher | I have been working on the svn2git process a lot in the last week, and it is getting close. |
| 18:53.15 | Patlabor221 | blast007, try this on your server when you have a chance, https://github.com/JeffM2501/bzflag-import-3/commit/471bc4ca197b3c11776c8c404c4223c93154e26a |
| 18:56.29 | BulletCatcher | BTW, the "don't leak player records" commit in pull request #11 is a good one that hasn't made it into the main v2_4_x branch. |
| 18:57.34 | BulletCatcher | You could merge #11 safely because the geno changes were undone, or just do the same thing in a commit directly on v2_4_x. |
| 19:01.50 | Not-311d | [02bzflag-import-3] 07JeffM2501 pushed 031 commit to 03v2_4_x [+0/-0/±1] 13http://git.io/VRGUVw |
| 19:01.51 | Not-311d | [02bzflag-import-3] 07JeffM2501 030a61018 - Don't leak a player record in this unused sample. Increment wins with new increment function. |
| 19:02.07 | Patlabor221 | I just redid it, there was a better way |
| 19:02.16 | Patlabor221 | since I added the increment funciton |
| 19:02.40 | BulletCatcher | That gets it done. |
| 19:03.47 | Patlabor221 | there is no real world problem based on the old code, I know of no custom plug-ins that used that part |
| 19:04.13 | Patlabor221 | I just wanted to have the sample be good because people do't understand they need to free player records :) |
| 19:29.42 | khonkhortisan | "I unloaded my plugin, now it's not storing any variables" - wrong |
| 19:33.49 | blast007 | khonkhortisan: what? |
| 19:34.15 | khonkhortisan | um, nvm if it doesn't make sense |
| 20:16.15 | *** join/#bzflag TimRiker (~TimRiker@bzflag/projectlead/TimRiker) |
| 20:16.15 | *** mode/#bzflag [+o TimRiker] by ChanServ |
| 20:18.22 | *** join/#bzflag bzbugsy (~mortal@unaffiliated/bzbugsy) |
| 20:27.09 | joevano | bzflag.org is down... |
| 20:33.34 | Patlabor221 | joevano, we proably just got bought by EA |
| 20:34.57 | joevano | Probably.. hopefully it was for 3 billion dollars |
| 20:35.14 | joevano | lots of potential... ;-) |
| 20:43.44 | joevano | and we are back |
| 20:43.55 | joevano | deal fell through |
| 20:44.19 | blast007 | nah, that was just the initial flood of new players making it go down for a bit |
| 20:51.59 | Patlabor221 | yeah Origin is so full of people |
| 20:52.07 | Patlabor221 | at least it's not ubisoft |
| 20:53.08 | blast007 | I wanted to like ubisoft... |
| 20:54.00 | Patlabor221 | uPlay activly prevents me from playing games |
| 20:54.12 | blast007 | just made it hard when I lost like 10 hours of progress because their stupid uplay cloud saving thing was broken, and then all the DRM crap |
| 20:56.39 | Patlabor221 | just added DRM to his product |
| 20:57.27 | blast007 | I think when I was having issues I was still playing over a crappy 2.4GHz bridged wifi setup in a household with 2.4GHz cordless phones :P |
| 20:57.53 | Patlabor221 | heh |
| 20:58.05 | Patlabor221 | "mom get off the phone, I need to get on the internet!" |
| 20:58.06 | blast007 | though to be honest, I'm even less of a fan of modern day EA |
| 20:58.12 | blast007 | pfft, that was basically it :) |
| 20:58.31 | Patlabor221 | with voip, it's now "Get off the internet, I need to use the phone" |
| 20:58.32 | blast007 | "I'll get you some DECT 6.0 phones!" "But I like these!" |
| 20:59.09 | blast007 | still rocking a WRT54G there to this day |
| 21:25.21 | *** join/#bzflag TimRiker (~TimRiker@bzflag/projectlead/TimRiker) |
| 21:25.21 | *** mode/#bzflag [+o TimRiker] by ChanServ |
| 21:56.40 | gnu-sense | Gosh that last burst of chat was hilarious. |
| 21:57.23 | *** join/#bzflag khonkhortisan (~Khonkhort@c-24-18-42-215.hsd1.wa.comcast.net) |
| 22:01.06 | gnu-sense | list-server should go down more often. Oh I see it is. hmm. maybe I shouldn't have checked "require registration" in the survey... |
| 22:01.31 | Patlabor221 | if we required registration we'd have more then one list server ;) |
| 22:05.58 | Patlabor221 | or facebook |
| 22:17.52 | blast007 | gnu-sense: only went down because I'm too dang nice and wanted to get someone another chance |
| 22:18.18 | Patlabor221 | bit you in the ass didn't it? |
| 22:19.08 | blast007 | yep :) |
| 22:22.03 | blast007 | back on the masterban and firewall lists he went |
| 22:23.19 | *** join/#bzflag catt (49ac7320@gateway/web/freenode/ip.73.172.115.32) |
| 22:23.39 | catt | =( |
| 22:23.46 | blast007 | ? |
| 22:24.00 | catt | someone is hacking mofo again |
| 22:24.06 | catt | his name is glo |
| 22:24.27 | catt | bragginbg about using a DOS script he found on google |
| 22:24.29 | blast007 | hacking isn't even the right term... |
| 22:24.36 | blast007 | attacking would be |
| 22:24.39 | catt | nDOS attack |
| 22:24.42 | Patlabor221 | yeah that's the thing to do this week it seems |
| 22:24.47 | catt | -n |
| 22:24.53 | blast007 | ~n |
| 22:24.53 | infobot | hmm... n is the disease bzflag players get when they play too much. There is no cure. |
| 22:25.02 | catt | lol |
| 22:25.05 | catt | ye |
| 22:25.41 | catt | 20 people were playing and having fun. now it's empty.. |
| 22:25.44 | catt | =( |
| 22:26.15 | Patlabor221 | they could try my fix, it could be worse |
| 22:26.31 | catt | never an admin around when you need them. that's how it is at mofo's |
| 22:26.41 | Patlabor221 | that's how it is at most servers |
| 22:27.09 | catt | busiest server on bzflag. but only a handful of admins |
| 22:28.20 | blast007 | neat, client just closed down while on mofo ;) |
| 22:28.28 | catt | ~moof glo |
| 22:28.28 | infobot | ACTION thwacks glo soundly with a cuddly large-antlered moof |
| 22:29.00 | catt | o well, thanks for hearing me vent ;P |
| 22:29.14 | *** part/#bzflag catt (49ac7320@gateway/web/freenode/ip.73.172.115.32) |
| 22:34.22 | *** join/#bzflag Patlabor221 (~JeffM@12.167.61.5) |
| 23:26.52 | *** join/#bzflag Patlabor (~JeffM@107-209-61-107.lightspeed.irvnca.sbcglobal.net) |