--- Log opened Sun Jul 24 00:00:50 2011
11:00 < conseo> mcallan: java's io api is a bit crippled and meant to be fixed in 1.7. i have patched your FileX.renameFrom() routine to work with different mount points (mine are / and /home conflicting with renameTo):
11:00 < conseo> http://pastebin.com/sz43BfcL
11:00 < conseo> may i commit?
11:02 < conseo> with it the servlet works again for me and tomcat is no more spilling errors about being unable to move the JSON files to /home
17:21 < mcallan> conseo: i see only two places where FileX.renameFrom* are called, did you add another?
17:22 < conseo> nope
17:23 < mcallan> then it's the wiki cache
17:23 < mcallan> did you look at using this?
17:23 < mcallan> http://zelea.com/project/votorola/_/javadoc/votorola/g/io/FileX.html#renameFromDefaultToMv%28java.io.File,%20java.io.File%29
17:24 < mcallan> i had a similar problem recently, and coded that as a solution
17:44 < conseo> hmm that is a simpler, yet more fragile solution
17:45 < conseo> the mentioned stack overflow article references that transferTo can use system memory to make the copying more efficient than writing in a loop (aka using the operating system)
17:46 < conseo> mv does that as well, but will not work on non-unices
17:46 < conseo> if i had actually seen the routine i would have used it, but i were just tracking the problem not really reading all of the api, so i missed it
17:47 < conseo> also you have to spawn a different process
17:48 < conseo> mcallan: what do you think about moving the transferTo stuff to renameFromDefaultToMv replacing the seperate process?
17:49 < conseo> since java 1.7 has new improved io api, we might stick with an intermediate solution anyway, so i don't care too much
17:50 < mcallan> i don't know what transferTo is, new method you coded?
17:50 < conseo> nope java sdk method
17:50 < conseo> i have included the header
17:50 < conseo> java.nio.channels.* i think
17:53 < mcallan> it doesn't look like a rename/move method, it looks like a copying method
17:53 < mcallan> http://download.oracle.com/javase/6/docs/api/java/nio/channels/FileChannel.html#transferTo%28long,%20long,%20java.nio.channels.WritableByteChannel%29
17:54 < mcallan> since there are only two renaming calls, and one works correctly, i would only fix the one that's broken. As long as the fix works (a clean move) then it should be ok.
17:55 < mcallan> co: plz document tho, in both methods, that they do the same thing
17:55 < conseo> mv copies and deletes, too, if the files are on different filesystems
17:56 < conseo> its just an internal implementation of the copy and remove method
17:57 < conseo> i can fix the renameTo calls to renameFromDefaultMv and mark renameTo as obsolete (?)
17:58 < mcallan> that's simplest and safest i guess
17:59 < mcallan> conseo: hopefully it's fixed in 1.7, as you say. plz document that in the method
18:05 < conseo> ok
18:06 < mcallan> xfbot: time
18:06 < xfbot> mcallan: The time is now Sun Jul 24 18:06:20 EDT 2011
18:06 < mcallan> did you restart it c, or did it recover from the netsplit itself?
18:18 < mcallan> i figured out where we can get users.  it's been bugging me and i haven't been able to concentrate on coding,  need to get it out of my system.  i'm drafting an essay and will post to the list, maybe in a couple days
18:24 < conseo> ok
18:25 < conseo> do you want to talk?=
18:33 < mcallan> conseo: you still busy with school?  i think the solution is vote mirroring
18:35 < conseo> no, i am finished, though i think i have f*cked up the last exam :-(
18:35 < conseo> hopefully i have still passed
18:35 < conseo> but anyway
18:37 < mcallan> you were confident though, so i hope that if you just did worse than expected, it will still be a pass
18:42 < conseo> yep, if i don't i'll have to repeat in october, so for now i am here
18:44 < mcallan> cheer up, man.  nothing you can do now but enjoy the summer :-)
18:48 < conseo> http://paste.ubuntu.com/651444/
18:48 < conseo> i know, i am not really worried it is just annoying
18:48 < conseo> but no prob
18:48 < conseo> i am really happy to be finally here with you and start hacking away on our little magical project
18:49 < conseo> i am trying to get javadocs properly setup, but i keep getting these weird unicode issues
18:50 < mcallan> javadoc has a problem with the class file.  truly weird
18:50 < conseo> actually i can't paste more then the first entry, the other 99 errors get ignored because of the chars i guess
18:50 < conseo> i am using an english system here and only activate the german keyboard if i need it atm. so i am not sure why this happens
18:51 < mcallan> in truth i have no idea what encoding our source files are, plain ASCII i guess
18:52 < mcallan> but it's the class file, so never mind
18:52 < mcallan> google
18:57 < conseo>  hmm my editor claims it is utf-8
18:57 < mcallan> class file?  try clean build, see if you get 100 errors for *every* file :-)
18:59 < mcallan> ASCII *is* UTF8, but not vice versa
19:02 < conseo> ok
19:13 < mcallan> yes, it is magical. :-)  we will wave the wand soon, i should tell Thomas before he goes (if not too late).  he would like to know
19:16 < mcallan> well, wait till it passes critique
19:22 < conseo> :-)
19:28 < conseo>  i think we might get users if crossforum is roughly running with your planned layout
19:28 < conseo> in fact it looks more compelling than without the toolbar, since you can then immediatly get an idea of the actions you can take
19:29 < conseo> before the toolbar it rather looked like a simple mapping tool, nothing which involves the user really
19:33 < mcallan> I never thought much of "build it and they will come".  I know we have to go and get them.  Hanging out in the Air-L and Liberationtech mailing lists this last week has changed my perspective on how to do that.
19:35 < mcallan> I've gone back to a scary assumption...
19:37 < mcallan> scary assumption: Online voting will take off next month.  It won't be Votorola that succeeds, or any project that is geared toward user freedom.  Instead it will be a monopoly that fences off its users and exploits them for private gain.  Other projects will be abandoned in a general stampede to the winner.
19:41 < mcallan> Ryan (that fellow from BC) pointed me to this: http://blog.hyperarchy.com/2011/07/19/hyperarchy-democracy-for-the-digital-era.html
19:42 < mcallan> they are just one more, and not any more likely to succeed than the others, but they are *so* closed it's scary
19:47 < conseo> yep itis
19:47 < conseo> is
19:47 < conseo> but they have nothing special
19:48 < conseo> i think we need the crossforum ui to a degree where it is alpha and then we can search for users, but before is difficult, isn't it?
19:48 < conseo> you can't talk anybody into using a software, esp. when it does not exist completely yet
19:48 < conseo> so in a way users have also come to us
19:49 < mcallan> we need the ui, including resources, and none of that changes
19:50 < mcallan> what changes (for me) is who we're coding for, and how we measure our success
19:51 < mcallan> how are these users, identified? https://hyperarchy.com/questions/414/votes/612
20:00 < conseo> what do you think :-)
20:00 < conseo> ?
20:00 < conseo> i think they are identified by the provider
20:00 < mcallan> right, hidden.  no links to user pages or anything.  so the app is useless to anyone as it stands
20:00 < mcallan> except maybe: http://www.facebook.com/pages/Hyperarchy/186466154739532
20:01 < mcallan> if it's not a nonsense app, then we can mirror their little votes
20:02 < mcallan> do the same for *every* other serious voting app on the net
20:02 < mcallan> then (wave of the wand) steal their users
20:02 < conseo> become troll du monde :-D
20:03 < conseo> good plan that is
20:03 < mcallan> i'll argue in the essay that it's an effective counter monopoly measure
20:03 < conseo> ok
20:04 < mcallan> it'll be fun too, with lots of covert activity ;-)
20:04 < conseo> still if our tools do not provide a comparable level of comfort we will not easily attract the users i fear
20:04 < conseo> we still have some work to
20:04 < conseo> do
20:04 < conseo> :-)
20:04 < mcallan> definitely the UI is important for this to work
20:04 < mcallan> but we know how to make the UI serviceable, that was never a problen
20:05 < mcallan> _problem_
20:05 < mcallan> i posted to their blog, nice and friendly.  they deleted it. :-)
20:07 < mcallan> I posted: "An excellent intro page!  Similar ideas are being developed in experimental projects such as Adhocracy and Votorola (a project I contribute to).  Both Adhocracy and Votorola incorporate the transitive delegation that Hyperarchy is now adding."
20:07 < mcallan> "All of these tools (and others) have their own strengths and weakenesses.  I hope we can make them work together for the benefit of the voters."
20:07 < mcallan> and so we shall ;-)
20:11 < mcallan> there's no way (short of wrecking your app) to prevent vote mirroring.  and places like Liberationtech (a kind of technical support network for freedom) are full of experts ready to help anyone who confronts "oppressive authorities" online
20:11 < mcallan> legally too, voting data cannot be protected by copyright, especially when the form of the data is modified in translation
20:12 < mcallan> (shut up Mike, draft the essay)
20:18 < conseo> sorry, were getting sth. to drink
20:19 < conseo> yep
20:20 < conseo> i think you are right, yet we have to get there so we can finally get something started
20:20 < conseo> crossforum looks very promising and is definetly the best vote ui i have seen so far
20:20 < conseo> even bettermeans, only focusing on a small aspect of what we can potentiall do is not that sophisticated imo
20:21 < mcallan> (u can get a drink this late on a Sunday night?  Mannheim is a rocking town ;-)
20:21 < conseo> mcallan: really weird, i can  build javadoc on my virtual machine (debian squeeze), but on my laptop (fedora 15) i get these unicode errors
20:22 < mcallan> google c, it's something weird
20:22 < conseo> i am a bit lost, hex editors didn't show something bad like wrong end line characters and astyle fixing testwise hasn't helped either
20:23 < conseo> huuh building takes now 670 s on my vm :-S
20:23 < mcallan> (UI) i agree we need it, but we need to know which (U)sers we'll be (I)nterfacing to
20:23 < conseo> at least the project is not small then :-D
20:24 < conseo> astyle has found 70.000 lines of java code, although i have measured 25.000 once in the past (?)
20:24 < conseo> sure, but we have got something at least reasonable to the state you showed in the ascii draft
20:26 < mcallan> (unicode) ;-) I think it's because you broke Woltman's Law: Never program and drink beer at the same time.
20:27 < mcallan> like i said tho, i agree about the ui.  vote mirroring has to be in the ui (a kind of resource), and xf theatre is our ui
20:28 < mcallan> it's our users (candidates) who are going to steal the users (voters) from closed projects
20:29 < mcallan> we'll give 'em the cross-project tooling/ui to do that
20:29 < mcallan> very similar to crossforum extent as a resource, except it's cross-project extent (mirroring sources)
20:30 < conseo> :-D
20:30 < conseo> -> woltman's law
20:31 < mcallan> we/they will "cultivate" that kind of thing, because it's in the candidate's material interest
20:31 < conseo> yep
20:31 < mcallan> Woltman was a smart man
20:31 < conseo> +1
20:31 < conseo> :-)
20:31 < mcallan> :-)
20:32 < mcallan> material interest is the basis of the essay, very solid thing
20:32 < mcallan> i like what Weber said, where is it? ....
20:33 < conseo> Weber? about materialism?
20:35 < conseo> sorry although they are nuts about Weber in Heidelberg, I haven't read him yet
20:35 < mcallan> me neither, just *about* him
20:36 < mcallan> here it is: "Not ideas, but material and ideal interests, directly govern men's conduct. Yet very frequently the world images that have been created by ideas have, like switchmen, determined the tracks along which action has been pushed by the dynamic of interest."
20:38 < mcallan> we can be the switchmen if we do this right, and let the users push things in an open direction by self-interest
20:39 < mcallan> it's the same motive that causes them to stampede toward a monopoly, but just a different *track*
20:39 < mcallan> it follows that we have to move fast, and head off the stampede
20:40 < mcallan> we can do it, it's been done before
20:41 < mcallan> but i should go and work, and let you enjoy the rest of your evening
21:04 < conseo> have checked file encoding: definetly ascii. have run javadoc from commandline on WikiCache, which leads me to:http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6442982
21:04 < conseo> seems i miss an include or a jar file on my classpath
21:05 < conseo> mcallan: maybe you should read marx if you already name it like that :-p
21:06 < conseo> we can only open a way up for people to organize differently in their material behaviours, we cannot change matter, only enable social changes, so i think on this level we consent :-)
21:08 < conseo> but +1 for going back to work, i really like what you have worked out so far, now i need to get my javadocs setup
21:08 < conseo> if i can't get them building then i will simply copy them from the server
21:10 < mcallan> ok (certainly i must read Marx, maybe even before Max Weber.  but i do not speak of social issues at large, only the behaviour of users in tool adoption. is it to be an open toolset like email, or a closed one like micro-blogging (Twitter) or publicity markets (Facebook))
21:12 < mcallan> (but never mind, you take Marx metaphorically, as I take Weber)
21:13 < mcallan> they are here too, c: http://zelea.com/project/votorola/_/javadoc/index.html
21:13 < mcallan> time to go, enjoy the rest of your night c...
21:17 < conseo> mcallan: sure, i just wanted to have the docs offline on my disk so integrate them to quickly look things up
21:18 < conseo> i can of course use yours or mine from my server
21:18 < conseo> i have a feeling that this issue is not worth wasting anymore time on it
21:18 < conseo> i don't know where to go
21:18 < conseo> i simply got a drink from downstairs (and it was without alcohol) so i am pretty boring tonight
21:19 < conseo> i don't think i will go somewhere
21:22 < mcallan> i got a ton of results: http://www.google.ca/search?q=javadoc+%22unmappable+character+for+encoding%22&ie=utf-8&oe=utf-8&aq=t&rls=org.mozilla:en-US:unofficial&client=firefox-a
21:24 < conseo> yep, but the encoding is right
21:24 < conseo> in fact the files are from you
21:26 < conseo> calling javadoc directly gives me this error: java.lang.ClassCastException: com.sun.tools.javadoc.ClassDocImpl cannot be cast to com.sun.javadoc.AnnotationTypeDoc
21:26 < mcallan> confirm it is a clean build
21:26 < conseo> i have several times
21:27 < conseo> i have even wiped the repo and tried with a new pull from my server
21:27 < mcallan> it was a char encoding bug, i think the classcast is unrelated
21:28 < mcallan> classcast means your missing a jar in your direct call, that's all
21:28 < conseo> oh ok
21:28 < mcallan> what's the first error on clean build
21:28 < conseo> http://stackoverflow.com/questions/44853/why-am-i-getting-a-classcastexception-when-generating-javadocs
21:28 < conseo> here i got with the casting error
21:29 < mcallan> confirm, casting error on clean build using our scripts (which always worked)
21:33 < conseo> no casting error
21:35 < mcallan> just this? http://paste.ubuntu.com/651444/
21:35 < mcallan> BTW, you can just do: votorola/b/build clean javadoc
21:38 < conseo> mcallan: have a look at zelea.com:/home/c/javadoc-build-problem
21:40 < mcallan> ok, btw this is a better search term for goog: javadoc "class file" "unmappable character for encoding"
21:41 < conseo> ok
21:42 < conseo> i have had this error before btw. (also when using ubuntu in winter), i just haven't cared enough to track it down
21:42 < conseo> weird that works on the debian server and for you
21:42 < mcallan> yes, like that first google result.  looks like it's mixing up .java and .class files
21:42 < mcallan> https://bugzilla.redhat.com/show_bug.cgi?id=719118
21:44 < conseo> thx, fix postponed until the bug is fixed, i'll use the server generated ones for now
21:45 < conseo> mcallan: xfbot should reconnect if it gets thrown out of irc indefinetly, i hope it works
21:45 < conseo> did you have to restart it sometimes?
21:45 < mcallan> nope, looks like it connected itself
21:46 < mcallan> conseo: is everything building on the same host and fs?
21:47 < conseo> yep
21:47 < conseo> but fedora has selinux enabled
21:47 < conseo> i get some nfs log entries which i don't understand atm. from selinux
21:48 < mcallan> and you opened this file, and looked at the encoding? /opt/votorola-build-cache/votorola/Java-compile-out/votorola/a/WikiCache$1.class:1
21:48 < mcallan> and what is the source file encoding, too?
21:50 < mcallan> emacs should say, for me it is plain ASCII
21:50 < mcallan> (the source file)
21:57 < mcallan> conseo: those class files are binary, and have no character encoding it appears.  that would mean that the *source* file has the wrong encoding.
21:58 < mcallan> you say the source file is UTF8, but what is javac expecting?  it expects default encoding for platform
22:00 < mcallan> conseo: echo $LC_CTYPE
22:03 < conseo> nothing
22:03 < conseo> empty
22:03 < conseo> no i looked at the encoding in the java file
22:03 < mcallan> conseo: echo $LANG
22:04 < conseo> en_US.UTF-8
22:04 < mcallan> it still makes no sense, because those source files are ASCII (you maybe changed WikiCache, but not VoteServer right?)
22:05 < mcallan> hmmm, that would be OK
22:06 < conseo> server is de_DE.UTF-8 and works
22:06 < mcallan> did you modify VoteServer.java?
22:06 < conseo> i haven't changed any of them
22:06 < conseo> well i have WikiCache
22:06 < mcallan> hg status
22:07 < conseo> http://pastebin.com/Vb1FUNiZ
22:08 < mcallan> it's building in your source directory, did you know?
22:09 < mcallan> http://zelea.com/project/votorola/b/Build_config_example.pl
22:10 < mcallan> look at your build_cache_root, and out_dir
22:10 < mcallan> should not be same as your source dir
22:10 < conseo> i see
22:11 < mcallan> i still don't know why exactly that would get javadoc upset
22:13 < mcallan> conseo: you might want to be careful, because 'build clean' deletes those output dirs
22:13 < conseo> i can't find the problem in the setup, both settings are fine and don't point to /opt/WORK
22:13 < conseo> maybe i have messed up something else
22:13 < mcallan> hmmm, remove those files (! ones) from your source code
22:14 < mcallan> sorry, i mean ? ones
22:14 < mcallan> they are javadoc output
22:16 < mcallan> conseo: then run: votorola/b/build --verbose clean javadoc
22:18 < conseo> c@zelea.com:~/verbose-output
22:18 < mcallan> conseo: also do 'pwd'
22:19 < mcallan> hg status
22:20 < conseo> see hg-status in c's homefolder
22:20 < conseo> weird
22:21 < mcallan> weird
22:22 < conseo> the "1" file is my fault
22:22 < mcallan> where have i seen '.orig' before?
22:22 < conseo> maybe it has to do with my editor
22:23 < mcallan> your editor is busy doing something?
22:23 < conseo> build clean removes all those *.orig files again
22:23 < mcallan> nonesense, then it will remove your repo source
22:23 < conseo> nope, no change
22:24 < mcallan> conseo: build clean works by deleting the directories!
22:24 < mcallan> it cannot remove single files
22:25 < conseo> hmmm no idea
22:25 < conseo> editor closed, cleaned, files gone
22:25 < mcallan> are you using a union file system?
22:25 < conseo> error still here
22:26 < conseo> so the orig files are at least not the reason, they might be related though
22:26 < conseo> nope it is a plain fedora 15 install with some packages
22:26 < mcallan> all that garbage in your repo is the smoking gun
22:26 < conseo> i use luks->lvm->ext4, nothing tuned nothing special from my side
22:27 < conseo> hg status is clean now
22:27 < conseo> the error stays
22:28 < mcallan> they could be temporary files, somehow generated during build process
22:29 < conseo> yep
22:29 < conseo> mcallan: don't waste your time with me
22:29 < mcallan> conseo: remove write permission on the source directory votorola/a/ (where all that garbage was) and retry
22:29 < conseo> i have worked around it and will stick to it to get on
22:31 < mcallan> conseo: ok, but i'm curious what happens if u do -w on the dir
22:32 < mcallan> it's bizarre to the extreme that it would writing junk into your source code, and then magically recovering
22:33 < conseo> hm i have applied it recursively so i have all files with status "M" in hg status now, but the build javadoc returns still the same output
22:33 < mcallan> not the files c, the directory
22:33 < conseo> sry, i am a bit wasted
22:34 < mcallan> just this last thing (and take your time! :-)
22:35 < mcallan> it *cannot* write then, and i want to see what error it produces WikiCache should be OK
22:37 < conseo> i think the orig files come from kate (my editor) and happen if i change files (e.g. through mercurial)
22:38 < conseo> if i remove write access nothing changes on build output
22:38 < conseo> still the same 100 unicode related errors and 100 warnings
22:38 < mcallan> http://pastebin.com/Vb1FUNiZ
22:38 < conseo> ?
22:38 < mcallan> caompare with verbose_output
22:38 < conseo> i have
22:39 < mcallan> kate did not make all those changes, how could it?
22:39 < conseo> i closed it
22:39 < conseo> if they were backups then it might
22:39 < conseo> i am not sure, sounds weird anyway
22:39 < conseo> you are right it is weird
22:40 < mcallan> ah, maybe it pre-emptively makes backups even when there are no changes
22:40 < conseo> puuuh is it late already that bug sucked all of my time
22:41 < mcallan> well, maybe after a good night's sleep...
22:41 < conseo> maybe :-)
22:42 < conseo> you are atm. working on the wiki stuff?
22:43 < mcallan> wiki stuff?  account defs are done, and I posted asking what you think of them: http://mail.zelea.com/list/votorola/2011-July/001134.html
22:44 < conseo> yep i have read that mail finally today
22:44 < conseo> i think it is good, i wanted to dive into the wiki a bit more once i am setup again
22:45 < mcallan> atm, i am writing an essay about why we can get users by vote mirroring, and not worrying about getting flattened by a 600 pound, venture capitalized gorilla
22:45 < conseo> i will try to use it myself the next days
22:45 < mcallan> _not worry_
22:45 < conseo> ok
22:45 < conseo> for what audience to you write it?
22:46 < conseo> for these lists?
22:46 < mcallan> for us first, and the list.  then i'll post on Air-L, see if anyone finds a flaw
22:47 < mcallan> i can't focus on coding when it feels like we're at risk (that's why i've been so slow lately).  but this should only take a few days
22:48 < mcallan> (i should let you get some zzz's, c)
22:50 < conseo> ok, good reflections shall be with you :-)
22:50 < conseo> cu tomorrow
22:50 < mcallan> thank u :-), good night c
22:52 < mcallan> conseo: one sec!
22:53 < mcallan> damn, i forgot to ask you to confirm: WikiCache was still in the javadoc error output, when it lacked permission to write to that dir?
22:54 < mcallan> something to check tomorrow... good night again
22:56 < conseo> yep it is
22:56 < conseo> and good night :-D
--- Log closed Mon Jul 25 00:00:05 2011