Hello, Here is the log for the latest meeting. Only 2 people were present, so not a whole lot was discussed. The summary is:
* Moving meetings to Tuesday nights * Re-designing the CBuffer class. Details are on the Wiki, and check out the discussion on there as well. Hopefully more will be present this week. Jon -- ________________________________________________________ Jon Keating ICQ: 16325723 [EMAIL PROTECTED] MSN: [EMAIL PROTECTED] http://www.licq.org GPG: 2290A71F http://www.thejon.org HOME: Minamiashigara, Japan
--- Log opened Wed May 10 20:00:41 2006 20:00 <@emostar> OK, Let's get this started 20:01 <@emostar> Hmm. anyone here? :) 20:04 -!- BDick [EMAIL PROTECTED] has joined #licq 20:12 < Rapp> hi 20:12 < Rapp> i have been at lunch, but now i am here 20:12 <@emostar> Ahh hello.. just was reading a book 20:12 <@emostar> Did you get my e-mail about licq-main? 20:13 < Rapp> so, i haven't got much of an agenda here. 20:13 < Rapp> what mail? 20:13 -!- wwp [EMAIL PROTECTED] has quit ["brôôh"] 20:13 <@emostar> hmm.. guess my email goes to /dev/null cuz i send from a dynamic ip 20:13 < Rapp> just wanted to note, that i'll look at the two patches later and see if they work and maybe submit them then 20:13 <@emostar> have to set it up to use Uncle-Mid's smtp 20:13 < Rapp> ok 20:13 < Rapp> so that about licq-main? 20:14 <@emostar> the one from Erik looks fine 20:14 < Rapp> i meant 'so what' 20:14 <@emostar> can you maintin licq-main as well? it has some messages building up 20:14 < Rapp> yes, of course 20:14 <@emostar> the password is the same 20:14 < Rapp> i clean all the lists up once in a while 20:14 < Rapp> i cleaned devel yesterday or so 20:14 < Rapp> usually i look into them once a week 20:15 <@emostar> wasnt sure you knew about licq-main.. it had 19 messages last time i think 20:15 < Rapp> but the spam accumulates fairly fast 20:15 < Rapp> it had 40 once... 20:15 <@emostar> 40 is nothing :) 20:15 < Rapp> yup :) 20:15 <@emostar> i had so many that some messages where gone because they expired.. 20:15 < Rapp> it would be nice if SF had some spam filtering... greylisting or something 20:16 <@emostar> i upgraded the svn server to version 1.3.1 the other day 20:16 < Rapp> yes, thats great 20:16 < Rapp> i did a successful checkout yesterday 20:16 <@emostar> maybe the timeout error is gone... 20:16 < Rapp> um 20:16 < eNTi> sorry 20:17 < eNTi> i'M here now. 20:17 < Rapp> no, not quite :) i had one timeout yesterday, but after two tries or so it worked 20:17 -!- Varjat [EMAIL PROTECTED] has joined #licq 20:17 < Rapp> at least the authentication problem was gone 20:17 -!- Varjat_at_work [EMAIL PROTECTED] has joined #licq 20:17 <@emostar> well, i will make a wrapper to debug what's going on to see if i can find out where it locks up 20:18 -!- Varjat_at_work [EMAIL PROTECTED] has quit [Client Quit] 20:18 < eNTi> emostar> did you look at my error? 20:18 <@emostar> eNTi: yeah.. quick fix might be to use SVN 20:19 <@emostar> so the agenda for today is rather small 20:19 < eNTi> emostar> what caused this problem? is it related to the fact, that the users do not get updated on the icq server? 20:19 <@emostar> here is what I have, I will just list it up now 20:19 <@emostar> * Moving the meetings to Tuesday nights 20:19 < eNTi> haha 20:19 <@emostar> * Getting more API discussion on licq-devel 20:20 <@emostar> * Getting things into smaller more managable tasks that can be done as a team 20:20 <@emostar> I have created a task list on the wiki with different stages 20:21 < Rapp> ah, let's take a look.. 20:21 <@emostar> http://wiki.licq.org/Developer_Area 20:22 < Rapp> yes, seen it. 20:22 <@emostar> i have some local modifications that arent in svn yet that related to changing rms.. and adding some of the pipeline to the daemon 20:23 < Rapp> let me guess -- stage 3 is the hardest? :) 20:23 <@emostar> stage 3 involves the most labor.. but shouldnt be too hard 20:23 < Rapp> ah, ok 20:23 <@emostar> designing is the hardest ;) 20:24 < Rapp> :) 20:24 < Rapp> so at the moment we are at top 3 of the agenda. who is there anyway at the moment to build a team? 20:25 <@emostar> i got in the mail "Modern C++ Design" today.. so maybe i will find some good ideas from there 20:25 <@emostar> but i think the pipeline architecture is nice 20:25 < Rapp> yes 20:26 < Rapp> with pipeline you mean the plugin-api that we have discussed in the past, right/ 20:26 < Rapp> ? 20:26 <@emostar> hmm, we have rsLeo.. Joachim.. you.. me.. 20:26 <@emostar> Agrajag- will be doing the releases now 20:26 <@emostar> and im hoping more people will be attracted into helping out if they see the wiki, blog and these logs 20:27 < Rapp> thats to be hoped 20:27 < Rapp> good 20:27 <@emostar> yes, the pipline architecture is what i call it.. since all events go to each plugin like traveling down a pipe 20:27 <@emostar> there is one task I'd like to have done that has nothing to do with the new API... CBuffer could be vastly improved to be more dynamic. currently it requires a size to be passed to it, and the buffer has to be exactly that size 20:28 < Rapp> what does CBuffer do? 20:29 <@emostar> its used by ICQ and MSN to have a buffer to send data over the sockets 20:29 <@emostar> licq_buffer.h 20:29 < Rapp> yes, i found the file 20:29 <@emostar> having to calculate the size of the packet before you send it is.. a pain in the ass 20:30 < Rapp> hmmm 20:30 < Rapp> i don't know that much about these protocols. 20:30 <@emostar> well, just think of CBuffer as a std::string that you append text to always 20:30 < Rapp> couldn't the protocols just << into the buffer, and the buffer would count the bytes? 20:30 <@emostar> you dont have to know the size of the text beforehand 20:31 < Rapp> yeah, so why don't we use some stringstream instead? :-P 20:31 <@emostar> \0 is in the ICQ packets a lot 20:31 < Rapp> hm 20:31 <@emostar> strings tend to stop at \0 20:31 < Rapp> yes 20:32 < Rapp> ok, so no raw stringstream class. but something similar. 20:32 < Rapp> what methods does the buffer class need to have? 20:32 < Rapp> i see a lot pack/unpack stuff in there 20:32 <@emostar> the current methods are ok.. minus the ICQ specific ones 20:32 < Rapp> ok 20:32 <@emostar> that can be done by a CICQBuffer in the icq protocol.. but it is nice to have a base class to do the grunt work 20:33 <@emostar> yep.. thats all we do with buffers.. pack, unpack.. ;) 20:33 < Rapp> so what do the pack/unpack methods do? 20:33 < Rapp> is this pure read/write, or is there some magic happening? 20:33 <@emostar> the protocol builds packets.. and the packets are just a stream of data.. in Licq's case, CBuffer 20:34 <@emostar> some endian swapping is done with certain calls 20:34 < Rapp> hm, ok 20:34 <@emostar> I will post more details on licq-devel when i examine it fully though 20:35 < Rapp> ok 20:35 < Rapp> i wonder if it would suffice to build a CBuffer that only does >> and <<. and to have some class for byte-sex-relevant objects 20:35 < Rapp> so that operator<< would convert automatically 20:35 < Rapp> or if that would mean too much changes to existing code 20:36 <@emostar> there are some << >> operators i think.. just didnt use them cuz icq mixes big and small endian together 20:36 < Rapp> i think about something like buf << header << ip_address << message; and ip_address would get conversion automatically 20:36 < Rapp> it does?? omg... 20:37 <@emostar> yeah.. icq protocol is a nightmare these days 20:37 < Rapp> oh ... my ... god... 20:37 <@emostar> aol bought it.. and mixed stuff from aim half-heartedly into it 20:38 < Rapp> so... 20:38 < Rapp> well... 20:38 < Rapp> still we could use a stream manipulator to do that :) 20:38 <@emostar> yeah 20:38 < Rapp> buf << header << setConvert() << ip_adress << message; buf.send() 20:38 < Rapp> don't know if we want to do that 20:38 <@emostar> if you dont mind calling the manipulator constantly. 20:38 < Rapp> would make the code more readable, maybe 20:39 < Rapp> that again is also true... 20:39 < Rapp> so how to you write to the buffer at the moment? 20:39 <@emostar> PackUnsigned[Long|Short]{BE} 20:40 <@emostar> PackString 20:40 <@emostar> PackUnsignedChar 20:40 < Rapp> so that writes to the buffer, and increments the buffer-position? 20:41 <@emostar> yep 20:41 <@emostar> and you have to know the size of the buffer before you make the buffer 20:41 < Rapp> why that? 20:41 <@emostar> because it isnt dynamic.. 20:41 <@emostar> thats why i want to change it 20:42 < Rapp> but in principle there is no problem making it e.g. a std::vector? 20:42 <@emostar> as long as it supports \0 20:42 < Rapp> (the underlying data structure) 20:42 < Rapp> yeah, sure 20:42 <@emostar> my STL reference book comes here tmw.. i havent had a good STL reference since i left the US 20:43 < Rapp> what about the SGI STL reference? very concentrated, but ok, i think 20:43 <@emostar> yeah, i use that.. 20:43 < Rapp> oik 20:43 < Rapp> -i 20:43 <@emostar> but its not very detailed about what throws exceptions and stuff 20:43 < Rapp> yes, thats true. on the other hand, i rarely use exceptions... 20:44 < Rapp> so shall we but the CBuffer task into Stage 1? 20:44 <@emostar> for the plugin api, making it exception safe would be a good design decision 20:44 < Rapp> true 20:44 <@emostar> yeah 20:44 <@emostar> nearly anyone can code that.. it has no relation to how an IM client works 20:44 < Rapp> sure 20:44 <@emostar> just have a few requirements that i have to specify 20:45 < Rapp> yes. will you write a short page on that? i think everythings said here, right? 20:45 <@emostar> yep! 20:45 <@emostar> probably tmw.. i'm gonna be a bit busy tonight 20:46 < Rapp> otherwise maybe i just type in something. 20:46 < Rapp> next top? 20:46 <@emostar> lets start talking about design decisions on licq-devel more 20:46 <@emostar> the meetings arent often enough to speed things up 20:47 < Rapp> yes 20:47 <@emostar> mainly aimed for Joachim.. who isnt here 20:47 < Rapp> thats ok 20:48 <@emostar> other than this... i dont have anything 20:48 <@emostar> just a few minor things 20:48 <@emostar> and some work that will be done to get thigns going as a team 20:48 <@emostar> i have been doing some coding and will start comming to svn tomorrow or friday 20:48 <@emostar> currently it compiles and runs.. but doesnt do a damn thing ;) 20:49 < Rapp> :) 20:49 <@emostar> but its exciting 20:50 <@emostar> thinking of the ideal messenger ;) 20:51 < Rapp> (just put in a short summary of the CBuffer thread, http://wiki.licq.org/Re-engineer_CBuffer) 20:51 <@emostar> so, i guess i will cut it short and take the train to meet my girlfriend now 20:51 < Rapp> ok, so moving the meetings to tuesday will be announced on the list, right? 20:51 <@emostar> yep 20:52 <@emostar> and on the current events in the wiki 20:52 < Rapp> ok 20:52 <@emostar> alright, thanks for coming! 20:52 < Rapp> have a nice evening, and see you soon, then. 20:52 <@emostar> alright, bye.. I'll leave the log open if others join and start talking ;) 20:52 <@emostar> bye 20:53 * Rapp is also away from keyboard.... 22:06 -!- wwp [EMAIL PROTECTED] has joined #licq --- Log closed Wed May 10 23:14:50 2006
pgpKMxHVRDUM2.pgp
Description: PGP signature
