Friday, 2009-03-13

*** tpb has joined #melange00:00
*** dmitrig01|afk has joined #melange00:01
*** ChanServ sets mode: +v dmitrig01|afk00:01
*** dmitrig01|afk is now known as dmitrig0100:03
*** dmitrig01 is now known as dmitrig01|afk00:34
*** SRabbelier has joined #melange01:32
*** ChanServ sets mode: +o SRabbelier01:32
*** SRabbelier has quit IRC01:59
*** dmitrig01|afk has quit IRC02:12
*** penyaskito has quit IRC02:31
*** SRabbelier has joined #melange02:39
*** ChanServ sets mode: +o SRabbelier02:39
*** _raz_ has joined #melange04:08
*** penyaskito has joined #melange04:17
scorcheso, an org admin uses the same link_id for the org application as he mistakenly did for his own profile...will this cause issues?04:57
*** osgeo_wolf has quit IRC05:04
*** osgeo_wolf has joined #melange05:07
*** mithro has joined #melange05:08
*** ChanServ sets mode: +v mithro05:08
mithrohey lh - you still about?05:44
* mithro didn't think so, but thought he would ask anyway05:51
mithroSRabbelier: you about?05:51
SRabbeliermithro: pong06:07
SRabbelierscorche: yes, that would indeed be an issue06:07
SRabbelierscorche: file a new org app with the proper link_id06:07
mithroSRabbelier: pmsg06:08
scorcheSRabbelier: should i wait for Pawel to come around and delete the user or the user and the org app?06:08
SRabbelierscorche: you could do that as well, but I don't know if he'll be around in time06:08
scorcheSRabbelier: i think i will wait for Pawel then, and if he doesnt show by 30-minutes-till, another app will be submitted then06:11
scorche(to be clearer, it is the admin's id that is the issue rather than the org's)06:14
_raz_exactly the same problem as for me - interesting that I'm not the only one doing that mistake06:16
SRabbelierscorche: aah, so the link id of the admin is that of the org, instead of something that indentifies the user?06:16
scorcheyes...he mistakenly thought it was for the org itself06:16
SRabbelierAnd that while there was this huge notice telling you what to pick as link_id...06:16
* scorche shrugs06:17
SRabbelierscorche: I'm really amazed by how badly people do not read help text06:17
_raz_SRabbelier: nope06:17
SRabbelier_raz_: nope?06:17
_raz_SRabbelier: when one clicks on apply as org (being logged in on google) and gets some 'sign in' with something named link_id, it's slightly confusing06:18
_raz_especially as one would expect to create the org params here06:18
SRabbelier_raz_: I see, so yyou're saying the "sign up as org" link shoulnd't show unless you created your Site Profile?06:18
SRabbelier(that's how it was before)06:18
scorcheSRabbelier: it was the differentiation between user and org that seems to be tripping people up for some odd reason..06:19
SRabbelierscorche: yes, I can imagine them getting the org link_id wrong, as that wasn't properly documented06:19
_raz_Either that or rename the link_id to something that explains what it means in which context06:19
SRabbelierbut I find it hard to udnerstand the other way around06:19
SRabbelier_raz_: Yeah, we will do that :)06:19
SRabbelier_raz_: see issue... 24706:20
_raz_SRabbelier: the main problem (in my opinion) is the plain "sign in" note, if you haven't created an apphost acccount yet - it would be fully sufficient to write something like "sign in as _user_ first" or so06:25
SRabbelier_raz_: ok... that can be fixed easily06:26
SRabbelier_raz_: dude06:29
SRabbelier_raz_: it says:  'Please create <a href="/user/create_profile">User Profile</a>  in order to view this page.'06:30
SRabbelier_raz_: how more obvious can it be that it's about a User Profile, not an Org profile?06:30
_raz_SRabbelier: yep, and when I asked in #gsoc about it yesterday, people explained how to use link_id and that stuff for the org application - so it is a problem source, if you are absolutely new to that app and the org link_id is titled the same way as a user link_id06:32
SRabbelier_raz_: yes, but the problem is that the user filled in the org link_id instead of their own06:33
SRabbelier_raz_: so either they did not read, or they did not understand what the text said06:33
_raz_right - in my case I explicitly asked for that and people got it wrong, too06:34
SRabbelier_raz_: either way, I don't see a way to improve the situation so that users don't use the org app's link_id for their own06:34
SRabbelier_raz_: explicitly asked for what?06:34
_raz_what to use there as link_id06:35
* SRabbelier mumbles something about Reading Documentation06:44
SRabbelierdanderson: ping07:05
SRabbelierdanderson: is it ok with you if I write down some assumptions about the release script in the Release Document from which you can then determine it's requirements for the Melange release process?07:06
dandersonSRabbelier: not sure what you mean, but write and I'll review07:07
SRabbelierdanderson: like this: "The name of the branch should be in the format "$major.$minor.$year$month$day", the release script will enforce this format and require you only to specify the correct major and minor version."07:08
dandersonSRabbelier: it's actuallt with '-' as separators07:10
SRabbelierdanderson: correct, my bad07:10
dandersonand the version format also support a 'pX' suffix07:10
dandersonfor a patch level07:10
SRabbelierdanderson: nevertheless :P07:10
dandersonbut, that's fine, yes07:10
SRabbelierdanderson: awesome07:10
SRabbelierdanderson: (will fix the format decsription)07:10
WinterMuteSRabbelier, nobody reads documentation until something goes wrong :p07:11
_raz_even then customers usually call the support hotline and blame them *ducks and covers*07:12
SRabbelierWinterMute: Maybe those that do not bother to read the documentation should not be part of GSoC :)07:12
dandersonSRabbelier: I've always said that07:12
SRabbelierWinterMute: If you can't take the effort to do at least that, I don't want to waste my time on helping you07:12
* WinterMute deletes his backup admin :p07:12
SRabbelierdanderson: :)07:12
dandersonI'm of the extreme opinion that if you post a FAQ to the mailing list, you should be immediately barred from SoC for 1 year07:13
SRabbelierdanderson: *grins*07:13
dandersonno warning, no second chances07:13
dandersonyou've annoyed 5000 people, that's what you get07:13
SRabbelierdanderson: Lol :D07:13
SRabbelierdanderson: tempting, but a tad too harsh :P07:13
SRabbelierdanderson: the point is to get new students into Open Source07:13
WinterMutebe funny though07:13
SRabbelierdanderson: not to send them running for the hill :D07:14
WinterMutewouldn't fancy dealing with the fallout you'd get from barring 90% of new users though ;)07:14
*** James--Crook has joined #melange07:21
dandersonSRabbelier: 90% is much I think07:22
dandersonon list we see a noisy minority07:22
James--CrookWhat problem does a user link_id being the same as an org link_id actually cause?  Confusion or something worse?07:23
James--CrookSo far we have been thinking about people accidentally using the same LinkId.07:24
dandersonofftopic awesome attack:
tpbTitle: YouTube - Shinedown - Devour - (Official Video) (at
James--CrookDoes it cause a problem if some time after orgs have been chosen I create myself as a user with LinkID apache?  I'd have thought not...07:25
SRabbelierJames--Crook: it won't07:26
SRabbelierJames--Crook: user and org space do not interact07:26
SRabbelierJames--Crook: you could even create a user called 'gsoc2009' and it would not matter07:26
James--CrookThat's what I thought.07:26
SRabbelierJames--Crook: documents have a prefix to differentiate between users, orgs, etc.07:26
SRabbelierJames--Crook: and all other places directly refer to their scope by type07:26
SRabbelierJames--Crook: :)07:27
James--CrookSo we don't HAVE to fix up people who make mistakes either.   The only issue is sorting of orgs in the org list07:27
James--Crookwhere they get sorted by LinkID rather than by name.07:27
SRabbelierJames--Crook: we won't have to, but I reckon some people won't like having 'apache' as their link_id07:27
SRabbelierJames--Crook: some other org was complaining about a user having their org name as link_id, which is copyrighted07:28
James--CrookYeah, OK, guess it make sense to want to delete that.  Or "inappropriate content in linkIDs".07:29
SRabbelierJames--Crook: right07:29
James--Crookclear to me now.  thnx.07:30
SRabbeliernp :)07:30
James--Crookdanderson: re 'awesome attack'.  I was most disappointed.  I was expecting something like the Boston Metro hack. :-)07:31
dandersonI require awesome music to wake up07:32
dandersonthis track may just make it to my wakeup playlist.07:32
SRabbelierdanderson: I am waking up to "Ghost Division - Sabaton (Art of War)" at the moment :D07:35
SRabbelierdanderson: that song totally blasts you out of bed :P07:35
SRabbelierdanderson: The one you linked is pretty awesome indeed! :D07:36
*** James--Crook has left #melange07:39
SRabbelierdanderson: what e-mail should I share the document with?07:52
SRabbelierdanderson: your gmail?07:53
danderson[email protected]07:53
dandersonand/or [email protected]07:53
SRabbelierdanderson: hehe, what's natulte?07:54
SRabbelierdanderson: oh?07:56
SRabbelierdanderson: what's the story behind that? :D07:57
dandersonvariant on the Quenya word meaning "he/she who is strong, who supports (morally/in battle)"07:57
dandersonthe story is long, complicated, and uninteresting :)07:57
SRabbelierdanderson: hehee, ok :P07:57
SRabbelierdanderson: shared it with your gmail account :D07:57
SRabbelierooooh, you have a blog there!07:57
SRabbelierhow cute :P07:57
SRabbelierdanderson: I therefore need to write a Motorola 68k emulator generator in Python.07:59
SRabbelierdanderson: sweet :P07:59
dandersonkeep reading07:59
dandersonit gets better/worse07:59
SRabbelierdanderson: you totally need to write a generator for Brainfuck, you'd an award for that, I'm sure of it08:00
SRabbelierdear god...08:02
* SRabbelier had an entire semester of writing in Scheme08:02
SRabbelierdanderson: that's totally awesome man :)08:02
dandersonand if that's your reaction to lisp, maybe you need to use it a bit more :P08:02
SRabbelierdanderson: lisp is like, _the_ language for AST stuff08:02
dandersonit took me a year of on and off fiddling, and reading Practical Common Lisp to finally get what the fuss was about08:02
dandersonSRabbelier: more than that, it's a great language for language-oriented design08:03
SRabbelierdanderson: right, exactly :)08:03
SRabbelierdanderson: we had loads of fun with Scheme :)08:03
dandersoninstead of modelling the problem domain etc. etc., you create a new programming language in which the domain can be readily expressed08:03
SRabbelierdanderson: I even wrote lists in our small langauge that had only functions08:03
dandersonso that's how I ended up defining a language to describe 68k cpu opcodes08:03
SRabbelierdanderson: (that is, I created a lambda function that returned the first value if you pass it 0, or the second value if you pass it 1)08:04
SRabbelierdanderson: (and you create a list by nesting these lambda functions)08:04
dandersonalong with a subset of lisp functions acceptable for use in the implementation of the opcode (let, +/-/*//, that kind of stuff)08:04
dandersonand then, epic code generators08:04
dandersoneating through the AST and constructing C or Java :)08:04
SRabbelierdanderson: the most ineteresting part would be the conversion of Common Lisp to Java I think?08:05
dandersonthat part is actually relatively easy08:05
*** Lennie has joined #melange08:06
*** ChanServ sets mode: +o Lennie08:06
dandersona lispy AST maps reasonably well to Java code08:06
SRabbelieroh really?08:06
SRabbelierdanderson: ok nice :)08:06
dandersonthe showstopper I hit was lack of unsigned ints in the JVM08:06
dandersonwhereas the 68k is all about unsigned ints08:06
SRabbelierLennie: see for some fun, you'll be going "AAAARGH" before the end of the article :D08:06
tpbTitle: Natulte::Blog (at
dandersonwell, not a showstopper exactly08:06
SRabbelierdanderson: yup, true08:06
SRabbelierdanderson: you just have to hack around it :p08:06
dandersonmost ops are the same for signed/unsigned08:06
SRabbelierdanderson: I'm sure the 68k did not have unsigned long longs :P08:07
dandersonthere's just stuff like division and comparison that suddenly become slow, because you have to squint at bit patterns08:07
dandersonrather than letting the target cpu take care of it08:07
* SRabbelier nods08:07
Lennie"But Java sucks"08:07
SRabbelierdanderson: So what's up with all the french?08:07
dandersonLennie: but Java is the only current target on Android08:08
dandersonwell, aside from native code, which I guess I could hack up from within google08:08
dandersonor outputting dalvik bytecode directly08:08
SRabbelierdanderson: that would be even more fun08:08
SRabbelieranyway, I have a presentation to hold08:08
LennieI like Java08:08
SRabbeliertoodles :)08:09
Lenniegood luck08:09
SRabbelierthanks :)08:09
Lennieethics ftw :P?08:09
* SRabbelier is afk now08:09
Lenniehi _raz_08:36
*** Merio has joined #melange08:38
Lenniehi Merio08:39
_raz_hi Lennie08:39
Lenniestill in need of link id change?08:39
_raz_sort of, I mailed pawel already and am waiting for a reply08:40
WinterMutedanderson, you're a nutter ( the good kind )08:43
Meriohi Lennie!08:47
Lenniehi Merio :)08:47
MerioNeed to go away, be back at 15:00/15:30 :)08:49
*** Merio has quit IRC08:50
*** solydzajs has joined #melange09:19
*** ChanServ sets mode: +o solydzajs09:19
Lennie_raz_ meet solydzajs09:19
Lenniesolydzajs meet _raz_09:19
Lennieyou two can fight over the chaning of his link_id ^_^09:20
solydzajsyo yo finally09:20
solydzajs_raz_: ping09:20
solydzajsLennie: going thru emails now09:20
solydzajs_raz_: what's the problem ?09:24
_raz_the user link_id being the same as the org application link_id, which's not that nice at all - it'd be great if the user (who's also the org application submitter) could be deleted09:26
solydzajs_raz_: you are the one who submitted it ?09:27
solydzajs_raz_: what org ?09:28
* _raz_ takes on the hat'o'blame09:28
_raz_Sledge: pygame09:28
_raz_errr solydzajs09:28
haoyuis appspot down?09:28
solydzajs_raz_: ok, if we delete your user account we will also delete your org application and you will have to submit it again09:29
solydzajshaoyu: works from me09:29
Lenniehaoyu, seems up but it is sluggish today09:29
Lenniethe datastore latency is high :)09:29
solydzajsLennie: yes correct09:29
_raz_solydzajs: no problem - there's enough time left so I can do it right after anything's cleaned up09:30
Lenniedid you create any documents under your user account _raz_?09:30
_raz_documents published on docs or under the appspot?09:31
haoyucan't be accessed from China09:31
Lennieif only we had access to docs :D09:31
_raz_I just created the org application, if that's what you mean09:31
solydzajs_raz_: ok good09:31
Lennieit's under appspot :D09:31
haoyuI tried both my machine and a remote machine (also located in China)09:31
Lenniehaoyu, interesting09:31
solydzajs_raz_: let me check your user account and org app09:32
haoyuoops, traceroute show appspot is BLOCKED by the GFW (aka. Great FireWall)09:35
LennieGreat FireWall of China :p09:36
haoyuwell, it is not a joke09:37
LennieI know09:37
haoyuhow about Chinese GSoC participants if appspot is BLOCKED?09:37
Lenniethey like to block the BBC :)09:38
Lenniehaoyu, yeah that sucks. I hope they'll be able to contact the people from google09:38
_raz_solydzajs: okay09:39
solydzajs_raz_: marcusvonappen ?09:39
solydzajs_raz_: ok I got your email too09:40
solydzajs_raz_: it's done please test09:43
solydzajs_raz_: let me know if it works09:45
_raz_looks good - new user acccount created successfully09:45
_raz_thanks a lot09:45
solydzajs_raz_: ok no problem09:45
Lennieheading home09:56
*** Lennie has quit IRC09:56
*** penyaskito has quit IRC10:00
*** solydzajs has quit IRC10:20
*** Lennie|Andy has joined #melange11:05
*** Merio has joined #melange11:22
*** AlexUA has joined #melange11:25
*** brlcad has joined #melange11:26
* brlcad wonders what the intent of the (required) Public Email is .. and what to put for a project that effectively has several or doesn't provide a generalized public e-mail11:28
*** Lennie|Andy has quit IRC11:31
*** MatthewWilkes has joined #melange12:33
*** brlcad has left #melange12:57
*** Lennie has joined #melange12:59
*** ChanServ sets mode: +o Lennie12:59
*** AlexUA has quit IRC13:00
MerioLennie: pong13:00
Lenniepm ^_^13:00
*** Lennie is now known as Lennie|Gone13:12
*** Lennie|Gone is now known as Lennie13:12
*** Lennie is now known as Lennie|Food13:13
*** dberkholz has left #melange13:31
*** wsfulton has joined #melange13:39
wsfultonhi, in my application I've used my own personal linkid for the linkid. Is this id meant to be the id for the project? If so, how can I change it or should I just delete the application and submit a new one?13:39
wsfulton is the current 'public' view13:40
tpb<> (at
wsfultonso wsfulton needs changing to swig13:40
ajaksuthat really needs a better wording in UI, my first document also had my personal linkid instead of a useful one :/13:47
lhajaksu: it's in the documentation. i realize the tool tip could have been better and there is an issue on file to improve it, but folks also need to read. :)13:49
haoyuI think link_id should be kinda of implemetation details and should not exposed to final users13:51
ajaksulh: IIRC, the UI just says 'linkid', where 'doc linkid' or 'new linkid' would be much more effective (to avoid this particular error) than any docs or tooltips could13:52
*** Lennie|Food is now known as Lennie13:52
haoyuLennie, ping13:52
lhajaksu: update the tool tip bug?13:52
tpb<> (at
Lenniehaoyu, wassup?13:53
haoyuI think the link id in everywhere shoud be replaced to 'User id' 'org id' etc13:53
haoyuLennie, wsfulton have problem with submitted a appliaction with wrong link id13:54
Lennieyeah we know13:54
haoyuis there way to fix this?13:54
Lenniethere might be13:54
lhhaoyu: see the issue report13:54
haoyuor maybe just delete the old one and create a new one?13:54
wsfultonI don't mind making a new application, won't take long if it is much trouble13:55
lhthat's probably much quicker than asking the devs to delete stuff for you13:55
Lenniewe'll have to delete the application anyway13:56
* haoyu agree with lh13:56
Lenniesince the system wont let him :)13:56
wsfultonokay no probs, I'll make a new application13:56
Lenniejust tell me the link id of the old one?13:56
lhLennie: as you wish :)13:56
haoyulh, I have see the issue report, so why not fix this? maybe it is too late to fix this?13:57
Lenniehaoyu, the biggest problem is that appengine uses the label set in the super class13:57
Lennieso we can't just overwrite it13:57
lhhaoyu: because by the time we do the next release of melange, we will no longer be taking organization applications. :)13:58
Lenniestudents dont have to fill it in :)13:58
haoyuyes indeed13:58
Lenniewe tried to fix it, but we havent found a suitable solution13:59
Lennieand we also have more pressing issues :)13:59
haoyuLennie, all these things are Linkable object and so they share the same link_id, right?13:59
*** Merio has quit IRC14:00
ajaksulh: done :)14:00
tpb<> (at
haoyuSo how about a new 'link_id_name' param when rendering template14:01
lhajaksu: tyvm14:01
haoyuthen set link_id_name='User id' for profile page and link_id_name='Org id' for org page14:02
Lenniewsfulton, send an email to [email protected] telling him what the link_id of your application is, so he can delete it :)14:02
Lenniehaoyu, it doesnt work that way14:02
Lennieif it did we would have fixed it already14:03
*** solydzajs has joined #melange14:03
*** ChanServ sets mode: +o solydzajs14:03
wsfultonLennie, will do, thanks14:04
Lenniehi solydzajs ^_^14:04
solydzajsHi Lennie :-)14:05
lhsolydzajs: hail!14:06
solydzajslh: hello :-)14:07
wsfulton<Lennie> just tell me the link id of the old one? > it is wsfulton, the same as my personal id. I've emailed pawel14:15
solydzajslooking into that right now14:16
Lennieshould be deleted soon14:16
solydzajsLennie: it's weird GAE console showed me more than 1000 users before in data viewer14:16
solydzajsLennie: now it doesn't14:16
Lennielol k14:16
Lenniedoesnt do it here either14:17
solydzajsit's weird14:18
solydzajssomething did break datastore ?14:18
Lennieguess its just GAE :S?14:19
solydzajswsfulton: can you check if your account still exist on socghop ?14:19
solydzajswsfulton: user profile, not org application14:20
solydzajsok got it14:21
solydzajsit's gae problem14:21
solydzajswsfulton: ok I just confirmed your account14:21
solydzajswsfulton: you want me to delete SWIG org application ?14:22
solydzajswsfulton: I need confirmation cause the application deadline is soon and I want to make sure you will resubmit it in time14:22
wsfultonseems to be there, I've been using it14:24
solydzajswsfulton: ok14:24
solydzajswsfulton: do you want me to delete SWIG org application ?14:24
solydzajsor do you want to change the link id of the org to swig ?14:25
wsfultonI've submitted a new one: The old one does need deleting:
tpb<> (at
solydzajswsfulton: ok deleting now14:27
wsfultonmy personal account seems to be working okay... I logged off and back on again14:27
solydzajswsfulton: the old has been just deleted, please confirm14:28
wsfultonseems to have gone... I can only see the one application now (the one I want)14:29
solydzajsok good :-)14:29
solydzajsso issue solved :-)14:30
wsfultonindeedy, great stuff, many thanks14:30
solydzajsno problem14:30
*** dhaun has joined #melange14:32
*** solydzajs has quit IRC14:43
*** schultmc1 has joined #melange14:43
*** schultmc has quit IRC14:45
*** schultmc1 is now known as schultmc14:45
*** EllenKo has joined #melange15:03
*** ajaksu has quit IRC15:23
*** ajaksu has joined #melange15:24
*** Lennie is now known as Lennie|Gone15:42
*** penyaskito has joined #melange16:17
*** eeanm has left #melange16:24
*** Alex_UA has joined #melange16:37
*** Lennie|Gone is now known as Lennie16:58
*** _raz_ has quit IRC17:25
*** MatthewWilkes has quit IRC18:01
*** MatthewWilkes has joined #melange18:03
*** dhaun has quit IRC18:19
*** WinterMute has quit IRC18:30
*** Lennie has quit IRC19:09
dandersonrefactoring of release tool progresses19:13
dandersonthere is now a logging subsystem that logs (with colorization) to the terminal, but also logs *everything*, including subprocess stdout/err, to a transcript log19:13
SRabbelierdanderson: ping19:16
dandersoncommence speaking19:16
SRabbelierdanderson: you import threading but then don't use it?19:17
dandersondamn, where?19:17
dandersonoops, got it19:17
tpb<> (at
SRabbelierdanderson: hehe, ok :)19:17
dandersonif you have other comments (on this or previous changes, please use the code review tool from google code19:19
dandersonI'm off to locate food19:19
SRabbelierdanderson: I will :)19:19
SRabbelierdanderson: just figured it was such a small change :)19:20
dandersonand if you can figure out a more elegant way of doing input than my sys.stdin.readline/terminal_echo combo, I'm all ears19:20
dandersonthe problem being that with the normal stdout redirected, I have to manually recreate raw_input()19:20
dandersonand yet I still want to log the prompts/responses to the transcript file19:21
danderson(but not to the terminal, because it's already there since we prompted the user)19:21
SRabbelierdanderson: Hmmm, this is always a problem I think19:23
SRabbelierdanderson: (even when doing redirection in bash)19:23
dandersonyeah, but I managed to work around it with a hack19:24
dandersonI'm just asking for help to make the hack a bit more elegant :P19:24
*** penyaskito has quit IRC19:25
dandersonbut I'm sure tlarsen will have an elegant solution :P19:25
*** penyaskito has joined #melange19:25
*** solydzajs has joined #melange19:51
*** ChanServ sets mode: +o solydzajs19:51
solydzajsdanderson: +1 :-)19:51
solydzajsdanderson: we will take care of that soon I hope19:51
SRabbelierGoing to bed now20:10
SRabbeliercya'll later :)20:10
*** solydzajs has quit IRC20:14
*** SRabbelier has quit IRC20:14
*** Alex_UA has quit IRC21:29
*** EllenKo has quit IRC21:40
*** serphen`` has joined #melange21:48
serphen``heya :)21:48
*** MatthewWilkes has quit IRC22:27

Generated by 2.13.1 by Marius Gedminas - find it at!