--- Log opened Fri Mar 16 00:00:08 2012
07:41 < conseo> mcallan: it is in the javadocs. it is just a timeline with blue marked areas. you ask before you schedule the job, if it is necessary and you add its covered timeframe afterwards
07:41 < conseo> thats it
07:47 < conseo> i rather have a problem with making one harvester handle all forums of a type. instead of one harvester per forum, but its the same harvester type. i don't see a performance problem for now (it is likely that the database is the eye of the needle)
07:47 < conseo> i have to map to forum in scheduler, history and harvester if i only have one instance of each, so i have to pass around forum types about the requested state then
07:56 < conseo> a central scheduler makes some sense
08:41 < conseo> i think i figured the scheduler out.
09:49 < conseo> mcallan: maybe i can figure it out
09:50 < conseo> don't worry
09:50 < conseo> mcallan: what is your status?
11:20 < mcallan> conseo: if you decide you want feedback, what i need to see is the algorithm for a single scheduled job (as i mentioned previously: 1 do this, 2 do that...).  i think that algorithm is all that matters for the harvester, because the rest depends on it
11:22 < mcallan> i am slowed down because i put the footings on the stage (hooked them into it), and it turns out they need cross-page (and cross-site) persistence of state.  i wasn't planning to deal with that till we started coding tracks, but i have to implement it now.  so i'm working on that
11:29 < conseo> ok
11:30 < conseo> is something new visible?
11:31 < mcallan> making it so...
11:35 < mcallan> http://havoc.zelea.com:8080/vdev/w/D?a=3265&b=3245
11:36 < mcallan> (click on draft at top, third from right)
11:37 < mcallan> then lock in one of the footprints, e.g. B1
11:38 < mcallan> note how fast you can navigate back and forth (using browser controls).  it should be instantaneous
11:39 < mcallan> (supports vector skin too... a lot of work, but i think it will pay off)
12:27 < conseo> if i click on draft i get a 403 forbidden response
12:27 < conseo> do i have to be logged in?
12:31 < mcallan> no, hmmm...
12:33 < mcallan> try now
12:48 < conseo> works. awesome. speed is impressive
12:49 < conseo> i wish i could share your progress more
12:49 < conseo> your work is still a bit opaque to me... i have to change that
12:53 < mcallan> my fault, i haven't posted progress for 2 weeks or something.  also much of this isn't visible to anyone who isn't working in the code.  you'll see it when you get to the client side
13:04 < conseo> yes, i need to start reading commit logs. i think this is the way to participate in foss projects
13:05 < conseo> yet you haven't comit yet (i have a small log on my desktop, but don't read the diff-patches usually)
13:05 < conseo> commit
13:06 < mcallan> it's not ready to commit, not till it works correctly
13:08 < mcallan> i work in a separate branch (i.e. repo) for each job, and i don't commit till the job is done.  but i can move to other jobs at any time.  (bit i think this is fairly standard practice with this kind of repo system)
13:08 < conseo> yes, i think so too
13:09 < conseo> only that i usually only work on one job
13:14 < mcallan> one bit of news is that gwt has been giving me lots of trouble, both its breaking of the bfcache, and bugs in the xs linker (which i have to use to fix the bfcache problem).  there's a 30% chance i'll ditch gwt and move the stage to another platform.  will decide when the mini-beta is nearly done; all but the new footing layout, which i can't do till there's a stable platform
13:15 < conseo> ok :-/
13:18 < mcallan> we're ok, it's easy to port the stage from gwt, because there's not a lot of code there yet.  the scenes can stay in gwt as long as they need to, because the stage can interface with other platforms no problem
13:19 < mcallan> (i like gwt, and i think it's a good place to park till something better comes along.  i won't abandon it till there's something safer to move to)
13:27 < conseo> ok, good
13:31 < mcallan> conseo: i just fixed my apache config so it's easier to control access.  please confirm the tool track (at top) looks ok and is styled black: http://havoc.zelea.com/w/User:Test-bg-ZeleaCom/Sys/p/sandbox
13:50 < conseo> yes
13:50 < conseo> it is styled like before
13:54 < mcallan> thx
15:31 < conseo> mcallan: have you done async http io yet? i am just looking into nio (as of jdk 1.4) to use a SocketChannel and run all requests asynchronly from one thread
15:31 < conseo> this is not a server socket, but a tcp client connection
15:32 < conseo> i will use this for now: http://docs.oracle.com/javase/1.4.2/docs/api/java/nio/channels/SocketChannel.html
15:39 < conseo> hmm: https://github.com/sonatype/async-http-client as of recommended here: http://stackoverflow.com/questions/1014528/asynchronous-http-client-for-java
15:42 < conseo> looks perfect
15:48 < mcallan> conseo: i guess you need async or the requests will logjam.  ok.  do look at the apache one too, because we already have httpclient.jar
15:48 < conseo> ok
16:04 < conseo> doesn't look like having async io
16:07 < mcallan> oh, one of the so responses said it did
16:07 < conseo> mcallan: i am looking into it
16:09 < mcallan> ok, i'll leave it you
16:18 < conseo> mcallan: httpcore (which we also have) seems to be the candidate: https://hc.apache.org/httpcomponents-core-ga/tutorial/html/preface.html
16:26 < mcallan> ok, i'll leave it you
16:37 < conseo> mcallan: interested in some issue? i'd like to discuss a bit with the tools, if you like to. i need a break and let all this io stuff settle (or at least i cannot concentrate on it for hours)
16:37 < conseo> i already now how to schedule the http requests, which is good for todays outcome
16:37 < conseo> know
16:39 < mcallan> what issue?
16:54 < conseo> ok, you are definetly not political :-P
16:54 < conseo> well, you are, but not like me
16:55 < conseo> never mind, i can figure something out. it is the wrong approach to start with a personal discussion maybe... although i pictured this as the ideal growing of a consent
16:57 < mcallan> if it's a problem, i always have time to discuss
16:59 < conseo> mcallan: it is not a problem, but i disagree with you on gmos and melinda gates foundation as well as issues of contemporary capitalism. it is not an easy thing, yet it is more or less the discussions we are targeting imo
16:59 < conseo> at least, i came because you drafted the best discussion tools imo
16:59 < conseo> now i only write java code and don't use them. i am a bit confused...
17:00 < conseo> (not a problem with you though, so not your business unlike you care about one of these issues)
17:00 < mcallan> sounds like a problem.  i'll turn on the skype box...
17:01 < conseo> mcallan: nope, sorry. it is not a problem, just that i have failed my own vision of progressing on the tools. i wanted to explore resource management, while this is on ice atm., i thought i at least would discuss
17:02 < conseo> likely i shouldn't do it with you, yet, it is a possibility
17:02 < conseo> i have just looked back on the last one and a half years and i have been so slow on my ideas, it is a pity
17:03 < conseo> i am also often distracted by political and social issues which i try to learn about and i need the tools to gain a better understanding of them imo
17:05 < conseo> using and developing the tools could help me to keep the motivation up and at the same time expand the content in the wiki
17:05 < conseo> but i won't do a monologue
--- Log closed Sat Mar 17 00:00:24 2012