--- Log opened Fri Nov 09 00:00:50 2012
12:57 < conseo> mcallan: why are the manifest files in the jars not sufficient to document version and license?
12:57 < conseo> i still see the usefulness to point out which code depends on the import.
16:21 < conseo> mcallan: why does Pollwiki.ConstructionContext has both setScriptLocation (redirect to setScriptURI) and setScriptURI?
17:21 < conseo> setScriptLocation is the right one
19:07 < mcallan> because it's config for Pollwiki.scriptURI().  http://zelea.com/project/votorola/_/javadoc/votorola/a/Pollwiki.html#scriptURI()
19:08 < mcallan> conseo: there's confusion because we have 2 Pollwiki classes.  also there was a javadoc bug.  both fixed in latest
19:10 < mcallan> re jars, i guess we could doc dependant code in associated .txt file, same as we doc dependant jars
19:19 < conseo> well maybe, but this can also be a lot of effort. i think one can gather that information through tooling
19:21 < conseo> i have also fixed the bug in maven (was not eclipse's fault) by including the jersey dependencies. i can now run the unit tests from inside eclipse and from maven command line
19:22 < conseo> i have also fixed the bug, so if the stage properly works for me, i will release
19:23 < conseo> but atm. i get the weird bug i posted to the list on my position page when navigating around in the votetrack to show the differences with julchen-68er
19:23 < conseo> http://zelea.com/w/User:4consensus_WebDe/Mann/p/urgar
19:23 < mcallan> (if you have proposal for better doc of jars, just let me know.  myself i am happy with current, so i won't take the lead on that) i will look at list ...
19:24 < conseo> ok
19:25 < mcallan> conseo: did u run in devmode, and look at trace?
19:25 < conseo> this is headless on the server
19:25 < mcallan> you are talking about this?  http://mail.zelea.com/list/votorola/2012-November/001497.html
19:28 < mcallan> (looks like personName is undefined in jso, but devmode will tell u)
19:28 < mcallan> (or it might not, because devmode is very forgiving of nulls)
19:29 < conseo> yes
19:31 < mcallan> u can repeat in devmode?
19:32 < conseo> in which jso does personName belong?
19:33 < mcallan> not sure.  devmode should tell u exact line #, right?
19:33 < conseo> i cannot start it without x server atm.
19:33 < conseo> have to retry locally
19:36 < mcallan> u saying u don't have a browser just now to test it.  ok
19:36 < mcallan> (loads ok for me, btw)
19:37 < conseo> i have a browser, i just can't run votorola/b/gwt-devmode there
19:37 < conseo> i could try to use ssh http://comments.gmane.org/gmane.org.google.gwt/67339 as it seems
19:41 < mcallan> why can u not run devmode?  i do not understand
19:43 < conseo> mmmh
19:43 < mcallan> u are using lynx browser :-)  ???
19:43 < conseo> i can run it locally with a remote tomcat? maybe i have done that once
19:44 < mcallan> back up.  c, are you sitting at a computer?
19:45 < mcallan> (yes)
19:45 < conseo> mcallan: i need to build it on v.polyc0l0r.net with this command, right: votorola/b/build xgwt --module='votorola.s.gwt.stage.StageInDev   votorola.s.gwt.mediawiki.MediaWikiInDev votorola.s.gwt.pollwiki.PollwikiInDev'
19:45 < conseo>  ?
19:47 < mcallan> first, is it possible to do debugging on your vdev, the local one, that uses havoc?  is that what meant earlier by "local"?
19:47 < conseo> yes, well, i first have to reproduce it with the sandbox list then
19:48 < mcallan> no, that's a valid reason to use v, if it's hard to replicate on vdev...
19:49 < mcallan> so yes, i think all u need do is ensure u have dev versions of gwt built...
19:49 < mcallan> since it is pollwiki page, all you really need is votorola.s.gwt.pollwiki.PollwikiInDev
19:49 < conseo> ok. i still have to replicate on the dev wiki for further testing. i think i will this the next days then
19:50 < conseo> ok, i am just rebuilding
19:50 < conseo> will consider it the next time
19:50 < mcallan> (yes, if u can replicate, far better to use vdev ofc.  people might be using v)
19:50 < conseo> (takes 20 mins or so)
19:51 < conseo> yes
19:51 < mcallan> conseo: just xgwt this one: votorola.s.gwt.pollwiki.PollwikiInDev
19:51 < conseo> i just wanted to release, never mind. i will finally get it done anyway
19:51 < conseo> mcallan: then i have to restart
19:52 < mcallan> (well, in future :-)
19:52 < conseo> is it dangerous?
19:52 < conseo> i guess so
19:52 < mcallan> breaking into xgwt?
19:52 < mcallan> no, it's safe
19:53 < mcallan> (i don't know what u meant by 'restart')
19:53 < conseo> well, but not intended for a productive environment, better only to enable what is necessary
19:53 < conseo> restart building
19:54 < mcallan> right, xgwt.  it's safe to interrupt it, and rerun
19:54 < conseo> ok doing so with PollwikiInDev only now
19:54 < mcallan> oh heavens, u mean safe to run PollwikiInDev?
19:54 < conseo> yes
19:55 < mcallan> yes, perfectly.  it just uses a different linker!  the only difference is BFCache behaviour, which has no security implications
19:55 < conseo> ok
19:58 < mcallan> (u can see in the gwt.xml modules.  only difference is linker http://zelea.com/project/votorola/s/gwt/pollwiki/PollwikiInDev.gwt.xml
19:58 < mcallan> xs linker (production) supports bfcache, but not devmode
19:59 < mcallan> xsiframe supports devmode, but not bfcache
19:59 < mcallan> that's all
20:04 < conseo> ok
20:18 < conseo> running in devmode...
20:23 < conseo> Wild exception: java.lang.IllegalStateException: overlap in event dispatch
20:23 < mcallan> !
20:24 < conseo> nothing in devmode console
20:24 < conseo> some error happened before
20:24 < conseo> should i recompile for the stage as well?
20:25 < mcallan> PollwikiInDev has all code needed for that page, incl stage
20:25 < mcallan> u say there was a *prior* error
20:25 < conseo> oohh, it lost the ?gwt.codesvr part
20:28 < conseo> http://paste.ubuntu.com/1346695/
20:30 < mcallan> dOldL.personName
20:30 < conseo> ?
20:31 < mcallan> u see it in the code?
20:32 < mcallan> somehow a LightableDifference is missing it
20:32 < mcallan> ...its personName
20:33 < conseo> yes
20:34 < mcallan> and i can't relicate at my end (loading same page), which is weird
20:36 < mcallan> one thing i can tell u, is the LightableDifference is a ShadowedDiff
20:38 < conseo> hmm it only happens when i am logged in as it seems
20:38 < mcallan> please trace.  if you look at ShadowedDiff code, you see it should be impossible for no personName
20:38 < conseo> have cleared the cache and tried to reproduce in devmode but failed, logged in to pollwiki and now it throws it again
20:42 < mcallan> careful you stay in devmode
20:43 < mcallan> conseo: there's no quick fix.  i can't replicate.  u need to trace it
20:44 < mcallan> odd that login state should affect it
20:46 < conseo> i will have a look into it the next days
20:46 < mcallan> ok
20:46 < conseo> need to go to bed and do stuff for university the next days
20:47 < mcallan> yes, ofc.  don't let this stuff worry u.  it's some tiny thing, and we'll find it
20:49 < conseo> wow, it crashed devmode :-D
20:50 < conseo> (jdk is a bit unstable on ubuntu 12.10 with multithreading still)
20:51 < conseo> have you had a look at groovy once?
20:52 < mcallan> as i say, looks fine for me.  here mouse is hovering over ur sister: http://zelea.com/var/tmp-public/urgar.png
20:52 < conseo> (it seems to happen after some period of time only)
20:53 < conseo> ok
20:53 < mcallan> (never looked deeply at groovy)
20:55 < mcallan> conseo: only one thought.  if login state really affects it, then that *is* your own position page.  there's a slightly difference path of execution because of that
20:57 < mcallan> (e.g. it knows you're on your own page, and disables links accordingly)
20:58 < mcallan> anyway, leave that for later
20:59 < conseo> ok
21:21 < conseo> gn8
21:28 < mcallan> n8 c
--- Log closed Sat Nov 10 00:00:07 2012