| *** tpb has joined #tp | 00:00 | |
| *** ChanServ sets mode: +o tpb | 00:00 | |
| shenki | glew: $ ssh-keygen -t rsa | 00:02 |
|---|---|---|
| glew | shenki: thanks, that prompts me to save it in a file, should i name this something specific? | 00:03 |
| shenki | that will produce a file ~/.ssh/id_rsa.pub, thats the one you want to send to mithro | 00:03 |
| glew | ok | 00:03 |
| shenki | use the defaults | 00:03 |
| glew | well the terminal is asking my to enter a file to save the key in | 00:04 |
| shenki | ensure you use a strong password, as this file will allow you to login to our server | 00:04 |
| glew | ok | 00:05 |
| shenki | so if you use the defaults for the file name, ~/.ssh/id_rsa is the password protected private key: make sure no one else gets a copy of that. | 00:06 |
| shenki | and ~/.ssh/id_rsa.pub is the public key; you can share that with anyone. we will place that in a file ~/.ssh/authorized_keys,which gives you access to ourserver without you setting a password on the server | 00:08 |
| *** jmtan has quit IRC | 00:08 | |
| glew | ok thanks for the help | 00:10 |
| shenki | no worries | 00:10 |
| tansell | ezod, pong | 00:21 |
| glew | tansell, hey | 00:40 |
| *** nash has joined #tp | 00:48 | |
| *** Erroneous has quit IRC | 01:17 | |
| *** glew has quit IRC | 02:44 | |
| *** peres has joined #tp | 03:35 | |
| *** mithro has joined #tp | 03:44 | |
| llnz | brb | 05:24 |
| *** llnz has quit IRC | 05:24 | |
| *** Epyon has joined #tp | 06:20 | |
| *** nash has quit IRC | 07:27 | |
| *** nash has joined #tp | 07:28 | |
| *** cahirwpz has joined #tp | 08:13 | |
| *** llnz has joined #tp | 09:50 | |
| *** nash has quit IRC | 10:04 | |
| tansell | cahirwpz, ping? | 10:05 |
| cahirwpz | tansell, pong | 10:06 |
| tansell | cahirwpz, got time for a quick review of your progress so far? | 10:06 |
| cahirwpz | well... yes | 10:07 |
| tansell | so where are we at? You've decided that using libtpproto2-py is a good idea and started writing a client? | 10:08 |
| cahirwpz | yes - I've decided to use libtpproto2-py and fix all its bugs I encounter during tpserver-py development | 10:09 |
| cahirwpz | I wrote a simple client that's able to talk to tpserver-py to verify my work | 10:10 |
| cahirwpz | I'll extend this client with automated tests | 10:10 |
| cahirwpz | today I made a branch in tpserver-py (called gsoc2010), switched libtpproto-py to libtpproto2-py and started to port server to twisted | 10:11 |
| tansell | cahirwpz, can you push this server to github? | 10:12 |
| cahirwpz | as soon as it's functional and ready to be presented - yes | 10:12 |
| tansell | cahirwpz, I would prefer to push early and push often | 10:13 |
| tansell | then you are less likely to waste your time going in the wrong direction | 10:13 |
| tansell | cahirwpz, I would also like to add a sub-goal of trying to make the tpserver-py code more tested | 10:13 |
| cahirwpz | so which repository should I use - my own, or push the branch to original tpserver-py repo ? | 10:14 |
| tansell | cahirwpz, your own "working" repository for now - as your happy with your changes we can then push to tpserver-py with a "clean" patch | 10:15 |
| cahirwpz | and the other thing is that I'll probably frequently modify libtpproto2-py and server won't work without those changes | 10:15 |
| tansell | cahirwpz, you should be pushing both obviously | 10:16 |
| tansell | github lets you "fork" a repository really easily | 10:16 |
| cahirwpz | ok - I'll try to fork and push my changes today | 10:17 |
| cahirwpz | lets get back to review | 10:18 |
| cahirwpz | we settled that remote console is not what we want and we'll be waiting for administrator extensions for tp04 protocol | 10:18 |
| cahirwpz | also I know how to implement zeroconf support | 10:19 |
| tansell | cahirwpz, has twisted gained zeroconf support - or do you have to impliment that from scratch? | 10:21 |
| cahirwpz | the last thing which is quite vague to me is built-in http server and game data download | 10:22 |
| tansell | cahirwpz, that should be trivially easy for you to support - it's basically adding a http server to serve the battlexml and media files | 10:23 |
| cahirwpz | tansell, it does not have explicit support for mDNS, however it contains full DNS server implementation so... I guess it's got everything we need | 10:24 |
| cahirwpz | tansell, yes... but I still don't understand in what way mentioned data is related to games / players, why do we need authorization and so on... | 10:25 |
| tansell | cahirwpz, so you understand that there are graphical objects which are displayed to the user right? | 10:26 |
| cahirwpz | mhm | 10:28 |
| tansell | IE ship pictures | 10:28 |
| tansell | and such | 10:28 |
| tansell | the client uses HTTP to download these pictures | 10:29 |
| cahirwpz | mhm... and why we need authorization for that ? | 10:29 |
| tansell | another type of media is "battle descriptions" | 10:30 |
| tansell | battle descriptions are a blow-by-blow record of what happened in a battle | 10:30 |
| cahirwpz | something like battle log ? | 10:31 |
| tansell | yeah - exactly like a battle log | 10:31 |
| *** Agon-laptop has joined #tp | 10:31 | |
| cahirwpz | ok... something else which is considered to be downloadable game data ? | 10:32 |
| tansell | they are the two main components at the moment | 10:32 |
| cahirwpz | ok... and why do we need authorization here ? why users should be restricted to download all available battle logs or gfx files ? | 10:35 |
| tansell | users shouldn't be able to download other people's battle logs | 10:35 |
| cahirwpz | there's one-to-one mapping between a game and a battle log, correct ? | 10:36 |
| tansell | no | 10:36 |
| tansell | the battle log is for each "encounter" | 10:36 |
| tansell | IE it says "Ship A and B fought it out and Ship A won" | 10:36 |
| cahirwpz | and who should be able to access a single battle log ? | 10:37 |
| cahirwpz | owners of both ships ? | 10:37 |
| tansell | each person gets a personalized battle log | 10:38 |
| cahirwpz | shouldn't download of battle log be covered by tp protocol ? | 10:39 |
| tansell | cahirwpz, no, because the battle logs can be huge (megabytes in size) | 10:40 |
| cahirwpz | I can hardly find any good reason (for now) to keep it separated from TP protocol... | 10:40 |
| tansell | and battle logs have associated media contents | 10:40 |
| tansell | you don't want to that over the tp protocol as it has no way of doing things like resuming | 10:41 |
| cahirwpz | what do you mean by resuming and lack of it in tp protocol ? | 10:43 |
| tansell | cahirwpz, you can't resume a file mid-transfer | 10:43 |
| cahirwpz | tansell, ok - that's good explanation | 10:44 |
| cahirwpz | but on the other hand using tp protocol you can download records one-by-one | 10:45 |
| cahirwpz | isn't that exactly what one would want to download battle logs ? | 10:46 |
| tansell | cahirwpz, no - you want the complete battle log so you can seek around it | 10:46 |
| cahirwpz | each battle log entry should be transferred as single short record, and client would then download all records | 10:47 |
| cahirwpz | can't battle logs be divided into records ? | 10:48 |
| tansell | cahirwpz, that would take forever | 10:48 |
| tansell | cahirwpz, there could be thousands of actions in a single log | 10:48 |
| cahirwpz | hmm... XML files tends to be larger and more time consuming to create | 10:49 |
| tansell | cahirwpz, think of it like a log of the moves in a chess game | 10:49 |
| cahirwpz | still you need to create them on server side - presenting each move as binary record is less time consuming (in terms of CPU time and transfer time) | 10:50 |
| cahirwpz | am I wrong ? | 10:50 |
| tansell | cahirwpz, sure - but then we design yet another binary protocol - compressed xml turns out to be pretty good | 10:51 |
| cahirwpz | this soultion seems messy to me - some things will be implemented twice because of representation differences | 10:53 |
| tansell | cahirwpz, like? | 10:53 |
| tansell | the battle format is not an "over the air" format | 10:54 |
| tansell | it's closer to the graphic files then it is to the rest of the tp data | 10:54 |
| * llnz wanders off | 10:54 | |
| llnz | later all | 10:54 |
| *** llnz has quit IRC | 10:54 | |
| cahirwpz | ok - I assume for now that you're right - we will get back to this topic before I start implementing built-in http server | 10:55 |
| cahirwpz | for now I feel I know just enough to continue | 10:56 |
| cahirwpz | I've got another question about protocol design | 10:56 |
| cahirwpz | FEATURES response can return information about the server (for example - that it does support SSL) | 10:57 |
| tansell | yeah | 10:58 |
| cahirwpz | I thought it might be a good idea to add command StartTLS, and drop using two sockets one for TP and one for TP-over-SSL protocol | 10:59 |
| cahirwpz | such command would be invoked be client if it knows that server support TLS | 10:59 |
| cahirwpz | after OK response from server - both client and server would start TLS negotiation | 10:59 |
| tansell | cahirwpz, see the filter negotiation stuff | 11:00 |
| cahirwpz | this is similar to behaviour of FTP+TLS and I believe many protocols does it this way | 11:00 |
| cahirwpz | where should I look for it - TP protocol description ? | 11:01 |
| tansell | http://www.thousandparsec.net/tp/dev/documents/protocolxml.php#Desc_SetFilters | 11:01 |
| tpb | <http://ln-s.net/6Zy6> (at www.thousandparsec.net) | 11:01 |
| cahirwpz | so... should I support additional port - 6924 ? | 11:03 |
| tansell | cahirwpz, if you want to support HTTPS Tunneling | 11:03 |
| *** cahirwpz has quit IRC | 11:08 | |
| *** cahirwpz has joined #tp | 11:08 | |
| tansell | <cahirwpz> so... should I support additional port - 6924 ? | 11:10 |
| tansell | <tansell> cahirwpz, if you want to support HTTPS Tunneling | 11:10 |
| tansell | cahirwpz, I'm about to head home - any quick questions? | 11:10 |
| cahirwpz | it seems I was disconnected | 11:11 |
| cahirwpz | yes | 11:11 |
| cahirwpz | so tp:// tps:// and tp+http:// services are using port 6923, and tp+https:// is using 6924, correct ? | 11:11 |
| tansell | tps:// uses port 6924 | 11:11 |
| tansell | as it sets up the SSL connection *before* the Thousand Parsec connection | 11:12 |
| tansell | tp+https:// uses 443 normally | 11:12 |
| cahirwpz | do we need tps:// after all if TLS is supported by tp:// ? | 11:13 |
| tansell | because TLS is only new in tp:// and no client actually supports it yet | 11:13 |
| cahirwpz | BTW I was unable to connect tpclient-pywx and tpserver-py over tps:// | 11:14 |
| tansell | it's not well tested | 11:15 |
| cahirwpz | has anybody tested SSL lately ? | 11:15 |
| tansell | nope! | 11:16 |
| cahirwpz | then I suppose none of the clients works with tps:// ;-) | 11:16 |
| tansell | well, very little of TP has been well tested - something I hope that GSoC helps with | 11:16 |
| tansell | cahirwpz, we are kind stuck in a phase half way between protocols and have been for a while | 11:17 |
| cahirwpz | I hope I'll help you with that and provide you a *really* good server implementation | 11:18 |
| tansell | cahirwpz, is porting tpserver-py to TP04 on your todo list? | 11:18 |
| tansell | (it shouldn't be all that hard, tpserver-py already has quite a flexible object model) | 11:19 |
| cahirwpz | well if libtpproto2-py won't give us headaches then porting tpserver-py to TP04 will be fairly easy | 11:20 |
| cahirwpz | at least it's something I believe in now | 11:20 |
| tansell | okay, well I need to head home | 11:21 |
| tansell | I may (or may not) be on a bit later | 11:21 |
| tansell | cahirwpz, bloging your thoughts and progress is a good idea too | 11:22 |
| cahirwpz | ... it'll be hard to get used to that :/ | 11:22 |
| cahirwpz | but I suppose I'll have to sooner or later | 11:23 |
| cahirwpz | tansell, see you later... I'm getting back to code | 11:25 |
| *** mithro has quit IRC | 11:28 | |
| *** Epyon has quit IRC | 11:56 | |
| *** Epyon has joined #tp | 11:56 | |
| *** mithro has joined #tp | 12:04 | |
| *** Agon-laptop has quit IRC | 13:01 | |
| *** Agon-laptop has joined #tp | 14:42 | |
| *** jmtan has joined #tp | 15:12 | |
| *** Agon-laptop has quit IRC | 15:16 | |
| *** peres has quit IRC | 15:19 | |
| *** bisc has joined #tp | 15:29 | |
| *** Agon-laptop has joined #tp | 16:15 | |
| *** Erroneous has joined #tp | 17:03 | |
| *** cahirwpz has quit IRC | 17:32 | |
| *** jnengland77 has joined #tp | 17:35 | |
| *** jnengland77 has left #tp | 17:35 | |
| *** StupidIncarnate has joined #tp | 17:54 | |
| *** jmtan has quit IRC | 18:11 | |
| *** Agon-laptop has quit IRC | 18:20 | |
| *** tuna-fish has joined #tp | 18:38 | |
| *** tuna-fish has quit IRC | 18:55 | |
| *** StupidIncarnate has quit IRC | 19:56 | |
| *** Agon-laptop has joined #tp | 20:30 | |
| *** llnz has joined #tp | 20:32 | |
| llnz | morning all | 20:45 |
| *** glew has joined #tp | 20:53 | |
| *** bisc has quit IRC | 20:54 | |
| *** Epyon has quit IRC | 21:09 | |
| *** Agon-laptop has quit IRC | 23:15 | |
| *** mithro has quit IRC | 23:47 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!