2009-11-22.log

--- Day changed Sun Nov 22 2009
-!- danielfalck [n=chatzill@pool-71-111-79-148.ptldor.dsl-w.verizon.net] has joined #hplusroadmap00:08
genehackerI'll bet you could charge $500:11
genehacker$2 for a hot dog?00:11
genehackersounds like a deal00:11
genehackeralso00:11
genehackerdon't mix coke and cofee00:11
genehackerit's always a bad idea00:12
katsmeow-afksedatives and caffine don't mix00:38
-!- flamoot [n=root@bas2-barrie18-1242454863.dsl.bell.ca] has joined #hplusroadmap01:39
-!- wrldpc2 [n=benny@ool-ad03fe34.dyn.optonline.net] has quit []07:04
-!- kardan_ [n=kardan@p54BE49B4.dip.t-dialin.net] has joined #hplusroadmap10:00
-!- kardan| [n=kardan@p54BE78AB.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]10:16
-!- strages [n=strages@c-76-29-231-141.hsd1.al.comcast.net] has joined #hplusroadmap10:53
-!- wrldpc2 [n=benny@ool-ad03fe34.dyn.optonline.net] has joined #hplusroadmap11:28
-!- rmond [n=strages@c-76-29-231-141.hsd1.al.comcast.net] has joined #hplusroadmap12:28
-!- strages [n=strages@c-76-29-231-141.hsd1.al.comcast.net] has quit [Read error: 110 (Connection timed out)]12:40
-!- zancas [n=zancas@c-71-201-189-153.hsd1.il.comcast.net] has joined #hplusroadmap12:41
-!- rmond [n=strages@c-76-29-231-141.hsd1.al.comcast.net] has quit [Read error: 110 (Connection timed out)]13:17
-!- strages [n=strages@c-76-29-231-141.hsd1.al.comcast.net] has joined #hplusroadmap13:26
fenndammit does nobody run unit tests?13:28
fennsince when does this not work?13:29
fennpackage = skdb.load_package('screw')13:29
ybitso no hatta?13:30
ybitwhat wiki to use?13:31
ybitfenn, kanzure 13:31
kanzurefenn: Package("screw")13:31
kanzureybit: god damnit, why don't you listen to either of us13:31
kanzurefdslkadsladl'dsa;l13:32
ybits/damnit/dammit13:32
ybitjklsdfajkldfjkasdf13:32
kanzuresorry, it's damnit13:32
kanzurei'm damning it13:32
kanzurenot damming it13:32
fenni can't even load a package without opencascade involved13:32
fennit's dammit13:32
kanzurefenn: not true..13:32
ybithttp://www.answers.com/dammit13:32
fennok well i can't load lego because of opencascade13:33
kanzureshow me13:33
fennwtf13:33
fennskdb.load_package('lego')13:33
kanzuresorry, how does that show me anything13:33
fennthen about six pages of traceback ending with:13:33
fennImportWarning: skdb.geom not loaded. load_CAD not available.13:33
kanzurethat should be a warning13:34
kanzurenot an assertion13:34
kanzure"raise ImportWarning, ..." 13:34
kanzuremy bad.13:34
kanzurethat should maybe be a print?13:35
kanzureybit: we're not using hatta13:35
kanzureybit: i don't know if you remember, but i was working on djangit for a while13:35
ybityeah i know, that's why i'm asking13:35
kanzureand then on ~/code/skdb/web/web.py13:35
kanzurebut apparently you've forgotten already13:35
ybityeah, i do, but we switched to cherrypy13:35
fennwell a warning shouldn't cause whatever i'm doing to fail13:35
kanzureweb.py still exists13:35
kanzurefenn: it's not a warning. it's a "raise"13:36
fennImportWarning isn't a warning?13:36
ybiti just didn't want to create a wiki from scratch -_-13:36
kanzurenope13:36
kanzureit says "raise ImportWarning"13:36
kanzureraise makes stuff stop13:36
fennam i supposed to use some other keyword besides raise?13:36
kanzureybit: oh please.13:36
kanzurefenn: i was the one who wrote that13:36
kanzurethis is my fault.. it shouldn't be "raise"13:36
kanzureit should probably just be "print"13:36
fenni dont understand; how do you do a warning without fucking everything up then?13:36
kanzurei don't think i've ever done that13:36
fennprint goes to stdout13:36
kanzurei guess you can do print >sys.stderr, "blah"13:37
fennwell, whatever. i'm sure there's a correct way to do it but i guess i don't care anymore13:37
kanzuresorry it's >>13:37
kanzureyou don't care?13:37
fenni'm saving my apathy for later13:37
fennto be savored as a fine delicacy13:38
kanzureshould i fix this or are you going to fix it?13:38
fennum, i'm just wondering what to say is  the "correct" way to load a package13:38
fennyaml.load isn't it13:38
kanzurePackage("name goes here")13:38
fennthat fails the same way load_package does13:39
kanzuregod damn it13:39
kanzuredidn't i just explain this?13:39
kanzureso i asked you13:39
kanzureare you going to fix this or should i?13:39
fenn  File "screw.py", line 48, in __init__13:39
CIA-33skdb: kanzure * r c9128c7 /core/part.py: because fenn is too lazy?13:39
fenn    if thread == None: raise ValueError13:39
fennnow i know i did something with that line, but you messed with it before that and it was even worse, and i dont even think it should check anything about thread?13:40
kanzurewhen i do pscrew=Package("screw") i get something about tensile_area in threads.py13:41
kanzure    assert Unit(self.pitch).compatible('mm/rev')13:41
kanzureAssertionError13:41
kanzurenot anything about a ValueError?13:41
fennand anyway why is thread == None?13:41
kanzure== checks if two things are equal13:41
kanzureit's not assignment13:41
fennum. then it really ought not to do that13:41
fennbecause something != nothing13:42
kanzure.. but it could be nothing13:42
kanzurehow long has it been since you've touched any code?13:42
* fenn hunts around for the elusive nothing13:42
kanzureit's very easy to "return None"13:42
kanzureand then if you have something assigned to that return value, that something is None13:42
kanzurehow are you getting the ValueError?13:43
fennPackage("screw")13:44
kanzurei get an assertion error about Unit(self.pitch).compatible('mm/rev') in threads.py13:45
kanzurewhen i run that13:45
kanzureare we running the same version?13:45
fenni'm running adl/master13:46
fennwow this is amazingly broken.. i dont know what's going on13:49
fennwhy is all this thread stuff in screw.py?13:49
* fenn notes all the lines saying "shouldnt this be in thread.py?"13:50
kanzurethen move it?13:50
kanzuredidn't you write this?13:50
fennwah. go away13:51
fennthis is giving me a headache13:51
kanzureseems too complicated13:51
kanzurehow did this happen13:51
kanzurei think we tried to cram too much yaml magic into metadata.yaml13:52
kanzure(tag hack?)13:52
fenndoes load_package even do anything with metadata.yaml?13:53
kanzureheh skdb.Package("screw") works on adl for me.13:53
kanzureyes13:53
kanzureloaded_package = load(package_file(name, 'metadata.yaml'))13:53
kanzureso the only reason for most of the complexity in Package is because of tag hack and the "template" attribute in metadata.yaml13:56
kanzureif the 'template' data is removed the point of the metadata is sort of lost. you'd /have/ to download the package to figure out if it's even remotely possibly compatible with something13:56
CIA-33skdb: fenn * r 4dd003d /core/part.py: plz at least try out yer changes13:57
kanzureyes i imported sys and forgot to commit13:57
fenni don't believe it; sys still wouldn't be defined in the module scope13:57
fennanyway13:58
kanzurein the module scope? huh?13:58
fenni guess you wouldn't get the error at all because OCC is installed13:58
kanzureactually i did get the error the first time because i was in code/skdb/core/ and so it was importing "skdb.py" when i was in bpython13:59
kanzuresilly mistake on my part13:59
fennwhat's up with git on adl14:02
kanzurehm?14:02
kanzurelibrary errors?14:02
fennsomething about libcurl14:02
kanzurei updated libcurl a few weeks ago for something unrelated14:02
kanzureif you know how to revert to a previous .deb of it, please go ahead14:03
fennsure you are running an up to date version on adl? i get the same error there14:05
fennof skdb*14:06
kanzurehuh? no i updated libcurl on adl, not on my laptop14:06
kanzureoh14:06
kanzureblargh? no i'm not14:06
kanzure /home/bryan/code/skdb/ is symlinked to /var/www/skdb14:06
kanzure(and that's what's in my $PYTHONPATH over there)14:06
kanzuregit: /usr/lib/libidn.so.11: no version information available (required by /usr/lib/libcurl-gnutls.so.4)14:06
kanzureok now i get the ValueError14:07
kanzurethat makes sense though because in the __init__ for Screw in packages/screw.py it sets thread to None14:07
fenni fixed git at least14:07
kanzurehow?14:07
fennapt-get install libidn1114:07
kanzure"GNU Libidn library, implementation of IETF IDN specifications"14:08
fennyeah, whatever that is14:08
-!- zancas [n=zancas@c-71-201-189-153.hsd1.il.comcast.net] has quit [Remote closed the connection]14:11
kanzurewill you clean up screw.py's __init__ re: thread shit?14:12
fennyah14:16
fennOCC is installed on adl.. so am i like the only person who doesn't have it installed?14:19
kanzurei think so yes :)14:20
-!- branstrom_ [n=branstro@c-381de055.438-1-64736c10.cust.bredbandsbolaget.se] has joined #hplusroadmap15:22
-!- branstrom_ [n=branstro@c-381de055.438-1-64736c10.cust.bredbandsbolaget.se] has quit [Remote closed the connection]15:22
-!- branstrom [n=branstro@c-171ce055.438-1-64736c10.cust.bredbandsbolaget.se] has quit [Read error: 104 (Connection reset by peer)]15:27
kanzurefenn: we have a few .png's in the git repository15:28
fennso?15:28
kanzurei was wondering if it would be terrible if i commit a .ico15:28
fennno, i'm mostly concerned about file sizes, that's all15:28
kanzurebinary data and such15:28
kanzureok15:29
kanzureit'll be really really tiny :)15:29
fennyeah they are like 1k15:29
-!- rmond [n=strages@c-76-29-231-141.hsd1.al.comcast.net] has joined #hplusroadmap15:49
-!- strages [n=strages@c-76-29-231-141.hsd1.al.comcast.net] has quit [Success]15:59
-!- flamoot_ [n=root@69.158.94.148] has joined #hplusroadmap16:04
kanzurei wish yaml had more mature path resolvers16:06
kanzureer, pyyaml16:06
-!- flamoot [n=root@bas2-barrie18-1242454863.dsl.bell.ca] has quit [Read error: 110 (Connection timed out)]16:16
kanzurehey doesn't github have their source up on the web somewhere?16:33
CIA-33skdb: kanzure * r c96c8b7 / (10 files in 6 dirs): added an email-based authentication system to web.py16:39
CIA-33skdb: kanzure * r a9b19e9 /web/web.py: converted from email-based authentication to username-based authentication17:03
fennyo kanzure can you re-enable wondershaper please?17:55
kanzuredone. dunno if it was running already or not though17:56
fennhmm must just be slow17:56
kanzuresam is filtering my emails to replab17:57
fenni think you totally missed the point of his essay18:08
fennit's like you figured you knew what he was going to say before reading it18:08
kanzurei'm kind of angry about replab.org if you haven't noticed18:09
fenni noticed; i think you're unfounded18:09
kanzureit's awesome they want to do this18:09
kanzurei feel ignored18:10
fenn'rfid powder' 0.05mm x 0.05mm18:10
kanzurehm?18:10
fennis that for real?18:11
kanzuredid i do that?18:11
fennno18:11
kanzurecontext?18:11
fennit's in charles collis' oekonus presentation18:11
fennoekonux*18:11
fennadciv.org/Slides  slide 818:11
fenndon't ask me why i'm reading this18:12
kanzureeff why eye: http://localhost:8081/package/lego/data/yaml18:13
kanzureneed to figure out how to do distribs18:13
kanzureer18:13
kanzurehttp://heybryan.org:8081/package/lego/data/yaml18:13
fennwhat does the cookie store?18:14
kanzurea session identifier18:14
kanzuredo you know how sessions work?18:14
fenni think it's like, whether you close the browser window or not18:15
kanzurebasically instead of giving you a cookie with all sorts of fun info, 18:15
kanzurei just give you one cookie, and cherrypy keeps track of some data for me18:15
kanzurein particular a User object18:15
kanzurewhich, for you, doesn't exist18:15
fennwhy are all the carriage returns stripped out? is that what happens when you dump text into html?18:15
* fenn mumbles something about <pre> tags18:16
kanzureif you view the source of the page you'll see the carriage returns are there18:16
kanzurethis is just a plain dump18:16
kanzuresowwy18:16
kanzurehtml doesn't respect \n18:16
fennoh. then it's a mime type problem18:16
kanzurerightright18:16
kanzurebut yeah, distributions18:16
fennum, sorry to be so clueless, but is this a new development?18:17
kanzureno18:17
kanzurebut user authentication is :)18:17
fennmeh18:17
kanzureso for distribs,18:17
fennwhat are "distribs"?18:17
kanzureon the web if joestupid edits the screw package,18:17
kanzurehe can screw up a lot of hardware18:17
kanzurefor this reason i was thinking of copying github's model of having users have their own git repos18:18
fennright18:18
fennthat was always the idea18:18
kanzure(either on the server or aggregated from other servers)18:18
kanzureright18:18
kanzureso how do we specify which 'distribution' to use for individual parts18:18
kanzureright now in the code base, in the metadata, we don't specify anything other than a tag/name18:18
kanzure"er, use cold rolling. there"18:18
fennum.. each package is its own git repo, right?18:19
kanzureyes18:19
kanzurebut referring to it by commit is not enough18:19
fennbecause it's not a URL?18:19
kanzurenonono18:19
fenni dont see why you cant refer to a commit18:19
kanzureyou can18:19
kanzureybit/screw/screw.yaml18:19
kanzurefenn/screw/screw.yaml18:19
kanzureybit's screw might have lots of changes18:20
fennbut i can pull from ybit18:20
fennthen fenn/screw/screw.yaml has changes too18:20
kanzurewell18:20
kanzureer18:20
kanzurethis feels upside down18:20
fennnow, version numbers would be a nice thing, and that's one thing i dont like about git18:20
kanzurei was thinking that you shouldn't have to download all of the hardware on the server18:20
kanzuresince you're already on the server18:20
kanzurein debian each package doesn't contain all other packages18:20
fennwhat server?18:21
kanzureheybryan.org:808118:21
kanzureer18:21
fennadl?18:21
kanzurefenn/screw/screw.yaml is your local changes to a screw, this is true18:21
fennsome hypothetical server?18:21
kanzurebut what if you want to use ybit's screw18:21
kanzurei don't think a copy of every single piece of hardware should be in fenn/18:22
fennfenn/ is my home directory?18:22
kanzurethe whole point of packages is that you can refer to them18:22
kanzurefenn/ is like http://github.com/kanzure/18:22
kanzureso yes(?)18:22
fennhuh18:22
fennwhat's wrong with having everything?18:22
kanzurewhy not just refer to the package18:22
fennpeople would only make mirrors of packages they have modified18:22
kanzureright now it sounds like you can _only_ mirror18:23
fennwell that's how github works18:23
kanzureyes but that's not hwo packages work18:23
fennwe can also give people access to various packages if they want18:23
fennwrite access*18:23
kanzureoh18:23
kanzureso here's what i was thinking of, maybe this is stupid18:23
kanzurei want to be able to say "use kanzure's wheel, commit blah"18:24
kanzureor "use wheel, commit blah" if you want the one from the global namespace18:24
fennmirrors can be merged, that's why dvcs rule18:24
fennjust say what commit and provide the url you got it from18:24
fennis that too much data?18:24
kanzureok so that's not the same thing as only allowing mirroring18:25
kanzureright?18:25
fenni cant think of a better way to do distributed version numbers18:25
fennwe can still do branches if it matters18:25
kanzurei don't think we're talking about the same thing18:25
kanzureand i'm still stuck18:25
fennno i'm confused, maybe you should flap your mouth and make noises or something18:26
kanzureon a user's hard drive, they should have a local copy of all of the .git's for packages, yes18:26
kanzure(that's their local cache of packages)18:27
kanzureon the server, the situation is different since the equivalent of /etc/apt/sources.list is basically "everyone on here"18:27
kanzureso there's namespace conflicts on "wheel" when there's fenn/wheel.git and ybit/wheel.git etc.18:27
kanzureso we need to be able to say /who's/ wheel18:28
fenni still don't get why the server has multiple copies of the same package18:28
kanzurewe approve some packages right? and others we might say "haha, no thanks"18:28
kanzurebut others might still use those packages if they are on the server, the ones that we didn't include in the main distribution18:29
fennwe don't have to host mirrors we dont like18:29
fennhere's how the linux kernel does it18:29
kanzureif a user doesn't have write access to the main distribution, then they can put it in their local namespace18:29
kanzurethis is how github does it :)18:29
fenneach dev hosts their own git repo. they send a git-format-patch to the linux dev mailing list18:29
kanzureright18:30
fennit automatically gets emailed to the correct 'lieutenant' who approves or deniesthe change18:30
fennthen the lieutenant forwards to linus for his veto18:30
fennthen linus does some massive git-pull from all the individual lieutenants18:31
fennthat's all there is to it18:31
kanzureok?18:32
kanzurenot what i'm talking about :(18:32
fennunfortunately we can't use that model because people are too stupid to host their own repo or send changes to a mailing list18:32
kanzureyes, this is just a repo hosting service layer thingy18:32
fennok i think that should be totally separate from everything else18:32
kanzuregithub.com/main/ will be the one that Master Linus commits to18:33
fennbecause it's one nad only function is to pretend like it's github18:33
kanzureso we shouldn't host other people's hardware?18:33
kanzureor by 'separate' what do you mean18:34
-!- branstrom [n=branstro@c-171ce055.438-1-64736c10.cust.bredbandsbolaget.se] has joined #hplusroadmap18:34
fenni just mean it's a totally different thing from "the database"18:34
kanzurethe what?18:34
kanzureyou don't use that word often.. 18:34
fennhum18:34
fennwell originally the idea was to have a select aggregation of hardware, compiled into a distribution (autogenix)18:35
kanzureyes18:35
fennbut lots of other people would be using the software for their own stuff, whether anyone cared or not18:35
fennthere's a lot of crap on thingiverse you notice18:35
fennbut we only want like 5% of it18:35
kanzureit's seeping out :(18:35
fennnow hosting the other 95% is ok but i just want to be clear what's what18:36
kanzureright18:36
kanzurethis is why i was saying 'distributions'18:36
kanzurenow, 18:36
kanzurewhat if someone wanted to make a new package18:36
kanzurethat depended on something from the other 95% of the stuff18:36
kanzurefrom the 'crapbin'18:36
fennmost thingiverse crap is not changes to other peoples' crap18:36
fennif you have good stuff that is derived from crap, then we can import the whole history of changes; that's why git is cool18:37
kanzurethere needs to be a way to specify a dependency in terms of which distribution it's from.18:37
fennno you just clone the crap18:37
fenn-_-18:37
kanzureit still doesn't make sense18:37
kanzurefor every user on thingiverse, each user should have a clone of every other package in their personal home folder on thingiverse?18:38
fennthe goal is to reduce data redundancy right?18:38
fennno, each use has a clone of whatever package they've made changes to18:38
fenn(unless they have write access to the package)18:38
kanzureok great so how does skdb know to use their screw rather than some other screw18:39
fennif it really actually in all honesty depends ona certain commit, then one should be able to specify that...18:39
fennbut i hope people are smart enough to not depend on such miscellanea18:39
fennthat's the point of OO, because people were digging in undocumented DOS registers and then it changed18:40
kanzurea dependency on autogenix/screw.git v. a dependency on ybit/screw.git18:40
fenninformation hiding, whatever (not OO)18:40
fennyou're not paying attention18:40
fennautogenix/screw.git is just a file location, not a version18:40
kanzureok true18:41
fennoh duh18:41
fennwe can put version numbers in the metadata file18:41
kanzureall i'm asking is that when a user makes a new package on the website, how do they say "use this version from this distribution"- *without* copying every single package in existence to their home dir?18:42
kanzures/copying/merging via git/18:43
fenn url and commit hash?18:43
fennsome urls have hashes embedded, but then you'd have to extract it if the url is bad18:44
fenncould be annoying if the url scheme changes18:44
kanzureso let's say i modify a non-autogenix package. this means i clone it and put it into my home dir on the server, etc.18:45
fennyou keep saying "the server"18:46
kanzuresomeone else then can depend on it by using a url/hash to github.com/kanzure/screw/4104814u90jfadsoiufh8934118:46
kanzureyeah a server18:46
fennany old server anywhere on the net, right?18:46
kanzurewell18:46
kanzurefor practical purposes, we might not want to allow that on ours18:46
kanzuresince importing a package means running python18:46
fennyeah well that's why we separate good from bad18:47
kanzureno i mean.. er.18:47
kanzureon our server, we "import skdb" in web.py18:47
kanzureand if you do load_package() on some foreign package, just to display a CAD model or something,18:47
fennffs make up some names please18:47
kanzureyou could be executing malicious code18:47
kanzurenames?18:47
fennall these servers i can't keep them straight18:47
fennjohnny fabonic uploads his changes to the tesla roadster to his free fabonautix account18:48
kanzuresally mcmistress makes a new package and depends on the johnny's tesla roadster18:49
fennelon reviews johnny's changes and decides to incorporate them18:49
kanzuresally also has a free fabonautix account18:50
fennelon does 'git pull fabonautix.com/johnny/tesla.git'18:50
kanzurewithout fenn knowing it, johnny included malicious code in tesla.git18:50
fennfabonautix goes out of business and johnny's mirror is gone forever18:50
kanzureweb.py is programmed to "import skdb" and run packages to make certain views on the website18:50
fennsally uploads her changes to mcmistress.com18:51
kanzurewhen elon and sally view tesla.git over fabonautix's web view, fabonautix is fucked18:51
fennwhat's fucked?18:52
kanzuretesla.git depends on badhacker.com/malicious-code.git18:52
kanzureweb.py would have to download that dependency in order to, say, render a CAD model for sally and elon when they view the page on fabonautix's site18:52
fennum, no, that's retarded18:53
fennall files necessary for the package should be included in the package18:53
kanzureso then you _don't_ like URLs for dependencies?18:53
fennpackages depend on other packages18:53
kanzureit's no different from them hosting on fabonautix, badhacker.com, etc.18:53
fennright so we have tools to manage dependencies18:53
fennwhen elon pulls the changes it should notify him of the external dependency, 'you want to download? [y/n]'18:54
kanzureok so how would it run on fabonautix?18:54
kanzurefabonautix would only run packages that have dependencies hosted on fabonautix? big deal.. they can just upload malicious code18:54
fennhow would what run, the apt-get style tool?18:54
kanzurethe package/code18:55
kanzurePackage("lego") <-- python is being executed from the lego package when you run that18:55
fenni dont think i ever figured out a good solution for that18:56
fennanyway, all the dependcies are in metadata.yaml18:57
fennno python involved18:57
fennunless its some quantitative thing like 'power output > 50W'18:57
kanzureok maybe web.py will distinguish between approved packages and questionable packages. "approved" packages will have more on-web functionality. 18:57
kanzure"questionable" packages (that haven't been reviewed) will still exist on the server, but the python won't be executed except for the metadata18:58
kanzurethis (might) mean no CAD preview for "questionable" packages 18:58
kanzure(especially for python-based code that generates a CAD model)18:58
fennyou know about virtualization right?18:58
fennfor sandboxing18:58
kanzurein general.. but not specifically for this18:59
fenneach fabonautix user could get their own virtual server.. :(18:59
kanzurea sandbox on each page loaad?18:59
kanzureblah18:59
fennok how about only static data for web preview19:00
fennscrews.py gereates a bunch of step files, but that only gets called manually19:00
kanzureso when you download elon/tesla.git it might have a dependency on sally/something-or-other.git which is on fabonautix. 19:00
kanzurewhat i'm wondering is how the dependencies should be written down19:00
fenni dont understand why you wouldn't download the dependency19:00
kanzureshould they say {username or distrib name}/blah blah blah19:00
kanzurewell you might have multiple versions on the hdd19:01
kanzurei guess git merge cna handle that19:01
kanzure*can handle that19:01
fennyou should be thinking about it in terms of branches, not username19:01
fennlike sally's branch is the hot pink paintjob branch19:01
kanzureok so you need to specify a branch and a commit then. 19:02
kanzurecan you selectively git clone/pull/whatever from particular branches?19:02
fennyeah19:02
kanzureok so all branches (specified in your dependencies) must be on fabonautix already?19:03
kanzurei mean, if you want things to work19:03
fennyeah19:03
fenndisk space is cheap btw19:04
kanzuredo you sign your commits with a key?19:04
kanzure(you personally)19:04
fenni dont know how to do that, but yeah the 'good' branch would be signed after it was reviewed19:04
kanzurewas just wondering because i dunno how github does it 19:04
kanzurein its display of commit messages it links to my username19:05
kanzurebut for you it doesn't do anything19:05
kanzurebut what if someone registers 'fenn' on github?19:05
fennoh you19:05
fenn're just worried about username conflicts?19:05
kanzurethis is a new issue. not worried, just wondering19:05
fennall this signing and reviewing is a lot of work, so i don't expect day to day development work to involve all this crap19:07
fennprobably only once per minor release19:08
kanzureno i mean username conflicts in terms of the users registered on fabonaughter versus the usernames in the commit messages19:09
fennmeh19:09
fennthis isn't AIM19:09
fennwe grownups can use real mathematics19:09
kanzurewhat does math have to do with it?19:09
fennpublic key signing19:10
kanzureso we do want individual commits to be signed? or what?19:10
kanzureor at least allow it as an option19:10
fennonly release commits i think19:10
fenn(bear with me a sec)19:10
fennsay i make five changes, then decide it's good enough for a while and make a release19:11
fennit gets uploaded to autogenix.org, i sign the commit hash with my key and paste that number in some document in the central autogenix.org repo19:12
fennthe commit i signed depends on the other four commits so they must be good too right?19:12
kanzureyes19:12
kanzurehttp://heybryan.org:8081/package/lego/data/edit 19:23
fenna 'submit' button would be nice19:24
kanzurebah! blasphemy19:24
kanzurei wonder if you can set a cookie through ajax19:29
kanzurethat would be nifty because right now an unregistered user can't edit a page19:29
kanzure(but i think we should let them edit and be able to register to save their changes, without losing their progress)19:29
fennhm19:30
kanzurei already tried on genebay to pass the content of their POST to the login form, but that didn't work19:30
kanzureand after discussion with the cherrypy peeps i gave up on that19:30
kanzureso ajax it is.. sorry lynx users :(19:30
fenni think that would be confusing, to edit something and then realize you have to register19:30
fennbetter to just not make the edit button show up if not registered19:31
kanzuredoesn't that lose the wiki-like appeal?19:31
fennif it were wiki-like we'd just let people edit whatever, which i'm still totally open to doing19:31
kanzureok cool19:32
kanzurebranch naming conventions plz?19:44
fennanthropomorphic assembly robots: http://youtube.com/watch?v=ZqLe0aWz8mg19:47
-!- wrldpc2 [n=benny@ool-ad03fe34.dyn.optonline.net] has quit []19:51
-!- Juul [n=Juul@85.235.22.9] has joined #hplusroadmap20:23
fennthere is no dana20:24
fennthere is only juul20:24
Juul:P20:24
Juulwhere does the name for the channel come from?20:26
fenna webpage on heybryan.org somewhere, which may no longer exist20:26
Juulok20:27
fenndo you know about transhumanism?20:27
kanzuresam thinks i'm worthless apparently?20:28
Juulyes20:28
Juulhey kanzure, just read up on the skdb project. much greatness. i like the ambition of it20:29
fennkanzure: that's why he was filtering your mail? because it's worthless?20:29
fennerm, you're worthless i mean20:29
kanzurefenn: he claims skdb is not interesting and isn't related to replab at all20:30
kanzure"unless you're highly autistic"20:30
fennwell, is that supposed to be an insult?20:30
kanzureno there's a smiley face after that20:30
fennwell who elected sam to be god20:31
kanzuremarcin20:32
fennmarcin loves skdb20:32
kanzuresince when20:32
fennsince a few days ago20:32
fennwhen i explained it to him20:32
fennJuul: tell me more about yourself, what got you interested in skdb?20:34
Juuli actually looked at skdb briefly a few days ago, but it wasn't until i read the slides that i had any clue how it would work. and the slides are easy to miss20:34
fennwhat slides?20:35
kanzureJuul: do you know what "apt-get" is?20:35
Juulfenn, http://adl.serveftp.org/lab/presentations/updates-from-austin.pdf20:35
fennok, i think that presentation is awful but apparently it does something for people20:36
Juulkanzure, sure, but that doesn't actually tell you how much work has already been done. it seemed very much like an idea with no implementation yet before i saw the slides20:36
kanzureoh what about the code in the repository?20:36
fennnobody looks at code, and even if they did they wouldnt know what they were looking at20:36
Juulexactly20:36
Juuli just thought "huh, well guess i'll check back later when there's more info"20:37
fenni'm afraid it's really hard to make pretty pictures of this sort of thing, or i would have done it long ago20:37
kanzureif you're not a developer maybe that's a good idea?20:37
kanzureer20:37
kanzurei mean, everyone is appreciated20:37
kanzurebut to be honest developers are what we need most20:37
fennanyone can learn to write a yaml document in 20 minutes20:37
kanzureok, a developer or someone with half a clue about engineering :)20:38
Juuli am actually a developer20:38
kanzureneat20:38
kanzureare you sam?20:39
Juulfenn, to answer your question: i'm booting a hackspace/fablab in Copenhagen right now, Denmark right now. IT engineer and studying biotech engineering now20:39
Juulnope20:39
Juulwow, i screwed that sentence up20:39
fennhuh i was just talking about copenhagen, about why legos are expensive (taxes in denmark)20:40
Juulheh, i remember some cheap clone-legos on the market here about 10 years ago20:40
Juulthey were horrible20:40
genehackerthat's why?20:40
fennsupposedly.. i think it's actually because lego groups i a bunch of greedy fucks20:40
kanzureJuul: ok sorry to confuse you. i guess you're fresh blood then20:40
genehackerfresh meat20:40
genehacker is the more correct term20:41
Juulkanzure, no problem, i'm new here yeah. only recently got on the open manufacturing mailing list20:41
Juuli think lego definitely charges more because they have the patents, but the high minimum wages and taxes in denmark are also a factor20:42
Juulit's not like someone can compete though, so i guess they just try to estimate how much people are willing to pay20:43
Juulso what are your backgrounds / approximate locations?20:45
kanzurebackground: internet :). i'm in austin20:46
Juulah cool. is "mojo's daily grind" still there?20:46
kanzureer, one sec20:48
fennlego factory is completely automated, no humans actually do anything in the factory20:49
fennso 'high minimum wages' is hardly an excuse20:49
Juulhehe, they still have employees doing packing and trucking, but yes, it should have minimal effect20:50
fennso i can get a usb composite video input dongle, and a usb to vga output dongle, but no such thing as a usb to composite output20:54
genehackercompletely automated you say?20:55
genehackerheheheheh20:55
fennthat's what rauchwerk said at least20:56
genehackerif only it was capable of automatic resource requisitioning20:56
fenn'more plastic needed'20:58
genehackerie robots getting more plastic automatically21:01
fennoh you mean stealing?21:01
fennyou're such an immoral person21:01
genehackernot necessarily21:01
Juulthe factory has evolved an AI and is secretly building a lego robot army by ordering 1% more plastic than needed for the normal lego production21:02
genehackerheh21:02
genehackerin b4 lego self-replicating machines21:03
Juulbeen done --> http://tekartist.blogspot.com/2007/02/lego-mindstorms-automated-car-factory.html21:06
fennwow this is terrible.. apparently marcin doesn't know what "recursion" means: http://openfarmtech.org/index.php?title=Technological_Recursion21:06
Juuleither that, or he's explaining it very poorly21:15
Juulno, you're right21:16
-!- Juul [n=Juul@85.235.22.9] has quit ["sleep"]21:21
kanzurehm maybe that was erik21:30
fennerik is belgian (?)21:35
fennapparent juul is a danish name21:36
* fenn needs a better euphemism for 'google says'21:37
fenn"If you already know what recursion is, just remember the answer. Otherwise, find someone who is standing closer to Douglas Hofstadter than you are; then ask him or her what recursion is."21:42
kanzuredid you come up with that?21:44
fennno it was on wikipedia21:44
kanzureha21:44
kanzurei think that's even better :)21:44
kanzurehttp://samba.org/~jelmer/dulwich/21:48
kanzure"Dulwich is a pure-Python implementation of the Git file formats and protocols.21:48
kanzureThe project is named after the village in which Mr. and Mrs. Git live in the Monty Python sketch."21:48
kanzurehead explodes21:48
-!- phwooenix [n=chatzill@99-190-134-198.lightspeed.austtx.sbcglobal.net] has joined #hplusroadmap21:49
kanzurehello phwooenix 21:50
phwooenixheyo21:50
kanzurehm an austinite21:50
phwooenixits david21:50
kanzuretreadwell?21:50
phwooenixor should i have let you keep guessing?21:50
phwooenixnope21:50
kanzurelee?21:50
phwooenixyessir21:51
kanzurei was ready to go through about 20 others21:51
kanzurehi. what's up?21:51
phwooenixnothing much just decided to visit21:51
kanzurei'm reading over http://samba.org/~jelmer/dulwich/tutorial/ at the moment21:51
phwooenixgit is that version control stuff?21:53
kanzureyep21:53
kanzurethis is, in particular, for a website i'm working on, to let users use git without knowing it21:53
phwooenixhow do you mean?21:54
kanzurehttp://heybryan.org:8081/21:54
kanzurehardware on this website is going to be stored in git repositories21:54
kanzureso links/buttons need to do stuff with git. that tutorial ishopefully going to be useful to me in this endeavor21:55
phwooenixhmmm your site takes extremely long to load21:56
kanzureit's being hosted on my residential connection :(21:56
phwooenixoh i see21:57
-!- flamoot [n=root@69.158.94.148] has joined #hplusroadmap22:01
-!- flamoot_ [n=root@69.158.94.148] has quit [Read error: 60 (Operation timed out)]22:01
-!- phwooenix_ [n=chatzill@99-190-134-198.lightspeed.austtx.sbcglobal.net] has joined #hplusroadmap22:11
kanzurewb22:13
-!- phwooenix [n=chatzill@99-190-134-198.lightspeed.austtx.sbcglobal.net] has quit [Read error: 60 (Operation timed out)]22:20
fennMarc Juul <marcjc@gmail.com>22:27
kanzurefrom where?22:27
fenncc'ed to "Fablab inventory and beyond" recently cross posted to hell and back22:28
kanzureare you still raped by cross-posts?22:28
fenni havent really figured out how to deal with it, no22:28
kanzurei wonder why marcin linked to http://openfarmtech.org/index.php?title=MIT_Fab_Lab_item_list rather than our list22:29
kanzurehaha he said "C02"22:29
kanzurei mean the formatting sucks much more on the wiki page22:29
fennis that even the same list as on mit's page?22:30
fenngod i can't stand any more of this wiki stuff22:30
* fenn used to like wikis :(22:31
kanzurei'm pretty sure they went "copy-paste"22:31
fennyes, but, there is stuff i don't remember22:31
* kanzure checks http://openfarmtech.org/index.php?title=MIT_Fab_Lab_item_list&action=history22:32
kanzuresam slapped me on the wrist for making an anonymous edit on openfarmtech.org. i didn't even know i was anonymous22:33
kanzurebut apparently that's bad22:33
fennthat was bad anyway, your huge copy-paste with no explanation22:33
fenndon't do that22:33
kanzureok22:33
fennyeah there is stuff that wasn't on the list before; the luxo lamp for instance22:35
fenn> Can you help me convince him that I wish him and skdb no ill will? I                               22:37
fenn> wasn't around for whatever made him upset but he seems bound and                                   22:37
fenn> determined to consider me an enemy.22:37
* fenn wonders what you said22:37
kanzurei thought we cleared that up22:37
kanzureuh, well, i told him i was upset22:38
kanzuremostly about toe-stepping and how they're ignoring us22:39
kanzureif they want to do something different that's cool, but they don't22:39
ybiti think my computer is on its last few volts of energy22:43
kanzurei can't figure out how to traverse a tree with dulwich22:44
ybitit has raped gmail over the past two days now, and it has only fetched 4.3gb22:44
ybitand it has taken about two hours to convert 20% of KSRM to audio22:45
kanzurehttp://www.diybioforums.com/projects/mccorkle_tomkins-tinch_microchannels.png22:45
-!- phwooenix_ [n=chatzill@99-190-134-198.lightspeed.austtx.sbcglobal.net] has quit ["ChatZilla 0.9.85 [Firefox 3.0.14/2009082707]"]22:45
kanzureoops i meant to link to http://diybioforum.org/projects/mccorkle_tomkins-tinch_microchannels.png22:47
genehackerwhat's that22:52
genehackeris that for shrinky dink microfluidics22:53
genehackeralso about that monitor 360 thing, what sort of questions did they ask you?22:53
kanzurethis was months ago, i guess you could go look it up on the web if you want22:53
kanzureit's for a laser cutter22:53
genehackeroh22:54
genehackerdid it work out?22:54
kanzuremonitor 360 interviews: http://groups.google.com/group/diybio/browse_frm/thread/446cff057ec54d3d/f4e4507819776d10?lnk=gst&q=monitor+360#f4e4507819776d1022:54
genehackerwhich was a month ago?22:54
kanzuremonths ago22:54
kanzurefebruary?22:54
genehackerok22:55
kanzureaha: http://diybioforum.org/projects/mccorkle_tomkins-tinch_microchannels.svg23:05
fennis that 'using computers to mimic biology' actually worth watching?23:17
fennor is it another trigger happy bryan moment23:17
ybitbastards at ustream have made it near impossible to cipher through the garbled text to download the video23:18
fennbut ybit aren't you permanently wired into the net?23:18
ybitwas planning on downloading and playing the audio at high speed tomorrow at work, but that probably isn't an option23:19
fennyou should have ustream receptors implanted in your eyeballs23:19
ybitheh, no thanks you can have adobe files flying through your body if you want though23:20
fennoo can it do silverlight too?23:20
ybitof course, it was made my microsoft23:21
ybitwtf is xulrunner-stub23:21
ybitit's taking half of my system's resources23:22
ybitthe other half going to festival 23:22
fennappears to be something firefox uses23:22
fennto communicate between ff and thunderbird, or something23:23
ybitwhy? i have only static pages loaded, damn this bloat, damn it to... to, uh..23:23
fenndll hell?23:23
ybityes! dll hell!23:23
ybiti feel much better now23:24
fennthis is intersting the vuzix VR920's come with magnetometer built in: http://www.pabr.org/wxhmd/doc/wxhmd.en.html23:24
fennnot worth an extra $250 though23:25
fenni love his bit-banging video output23:27
ybitthat's kind of neat23:28
ybitfenn: how often are you using your myvu now?23:29
* fenn wonders whether to tell the truth..23:30
kanzurei think i've been using it more23:30
ybit:P23:30
fenni ran out the batteries today, supposedly it goes 4 hours23:30
fenni was playing this terrible video game "malice"23:30
kanzurei still think there's a beaver in there somewhere, or something23:30
fennbecause another terrible video game also named "malice" has a cool soundtrack, which i have, and got sort of mixed up23:31
fennare you thinking "crash bandicoot"?23:31
kanzurenope23:31
kanzurebut that, sir, is also a terrible video game23:31
ybitdiddy kong racing ftw23:32
ybitwow, i'm out of the loop, i didn't realize they still make crash bandicoot games23:33
kanzuremy first exposure to crash was with spyro the dragon. crash bandicoot was a demo game on the same disc23:33
* ybit hasn't played a videogame in ..how long Phreedom? grep the logs!23:34
fennbatten the hatches! swab the lens!23:34
ybitwow.. 7 months maybe?23:34
ybitis this what happens when you go from geek to nerd?23:35
fennare gamers geeks?23:35
fenni mean it's correlated but not causative23:35
ybiti'd classify them as such23:35
ybitthinkgeek thinks so as well23:35
fennof course thinkgeek would like you to think you're a geek23:35
ybittrue, they would call operah a geek if they knew she would buy their stuff23:36
ybitoprah*23:36
fennoprah is a geek.. her name is an anagram, or something23:36
kanzureophar?23:37
kanzurei fail23:37
fennit's "harpo" spelled backwards :\23:37
kanzure"Harpo (born Jan Harpo Torsten Svensson, 5 April 1950 in Bandhagen, Stockholm), is a Swedish pop star known under the stage name Harpo. ..."23:37
ybitArthur Adolph "Harpo" Marx (November 23, 1888 – September 28, 1964), born Adolph Arthur Marx, was the second-born of the Marx Brothers and a unique comic performer whose style was influenced by clown and pantomime traditions. 23:38
ybitcertainly geeky ;)23:38
ybitharpo studio runs the oprah show or something like that23:38
kanzurethe what show?23:39
ybitsome show on t.v. that older women tend to watch23:39
ybitminus katsmeow-afk 23:39
genehackerharpo studies runs oprah?23:41
genehackernever knew23:41
fenno teh joys of knowledge23:41
ybitneither did and i still don't, it was a quick google scan 23:41
genehackerthe vuzix has a magnetometer?23:41
ybithow cool is that23:42
fenn<- not impressed23:42
genehackerwhy?23:42
fennbecause it should have a fucking gumstix built in too!23:42
genehackerwhat the heck does it use the magnetometer for?23:42
fennso you know what direction you're facing23:42
fennaccelerometer only tells you which way is down23:43
genehackerdoes it show north?23:43
genehackerin the goggles?23:43
fennjust think about it for a while23:43
fennok got it?23:44
genehackerwell I thought vuzix only made stuff for ipods and things like that23:44
genehackerI didn't know they were actually doing head tracking23:44
fennah but this is the "VR" model23:44
genehackeroh23:44
genehackerno wonder23:44
fennit's totally lame, don't worry too much about it23:44
genehackerI should have known23:44
genehackerit's not totally lame if you get the head tracking output from it and hook it up to an RC plane with a camera on it actuated by a servo controlled by the head tracking output23:45
fennok but why do they have to use HIDRAW and recompile the kernel just to read the tracking data :(23:46
genehackerI don't know23:46
fenni mean that just screams serial port23:46
genehackerbecause it's proprietary?23:46
* fenn shrugs23:46
genehackeranyway, something called a monotiltrotor would make a perfect delivery UAV23:47
genehackerhttp://www.youtube.com/watch?v=lFW7HEupkCM23:48
fenn"The systems draws 1 A with no power optimizations. This is acceptable since nobody would want to spend more than a few minutes with two pulsed microwave RF transmitters, an overheating lithium battery and eye-straining optics strapped to their forehead anyway."23:48
genehackerof course that's what it's being touted as23:48
genehackerhahaha23:48
genehackerhow's the eyestrain on the myvu?23:49
fenni dunno, it's certainly not supposed to fit over your glasses so hard to say what it would be like if it werent so obnoxious23:50
fennman i just want to smack the industrial design department http://www.slashgear.com/kopin-golden-i-gen-2-wearable-computer-on-course-for-2010-launch-2257525/23:53
genehackerwhy?23:54
fennbig red blobs of plastic!23:55
fennand obnoxious logos, on my face!23:55
ybitsed -e :a -e 's/<[^>]*>//g;/</N;//ba'23:55
ybitnow that's nice23:55
genehackerthat looks optimal for not falling off your head and not vibrating23:55
genehackernot falling off your face is important in the applications where wearables are used23:56
genehackerIE plant inspection23:56
genehackererrr...23:56
fennuh huh whatever23:56
genehackerare they used in plant inspection yet?23:56
genehackerbut yeah awful23:56
ybitregex isn't difficult, and it's quite fun/useful.. 23:56
genehackerwait that's a whole computer?23:57
fennit's the same processor as the beagleboard or gumstix23:57
genehackerso not much power23:58
fenncortex can render 720p video.. good enuf23:58
genehackerwhat do you mean by render?23:59
genehackeris it enough to play doom?23:59
fenndisplay on screen23:59
fennof course it can play doom.. wtf23:59
genehackerok23:59

Generated by irclog2html.py 2.15.0.dev0 by Marius Gedminas - find it at mg.pov.lt!