All times shown according to UTC.

Time Nick Message
00:00 paulproteus It took me some niggling with details to make it work, but eventually it did.
00:01 willkg: The big problem for me is time-frame.
00:01 I have all this fairly-urgent MC work to do.
00:02 I can try to spend 1-2h on the extraction and documentation of "apilib".
00:02 ajonas left #miro-hackers
00:02 willkg i was thinking we could just tweak the api so that auto_scrape took an http-handler that we could slide in for urllib2.  but it'd do callback/errback stuff.  then miro would create an eventloop/libcurl based one, we could create a sync/urllib based one, adn mc could do whatever you want.
00:03 paulproteus Now you've lost me.
00:03 willkg i'm totally game for doing the work on this.  you're only one man working on mc all by yourself with some hefty deadlines.
00:03 paulproteus Okay, that would rock my socks.
00:03 willkg so auto_scrape is the thing that mc and miro will be calling.  i figured i'd just pass it an http-handler object.  then vidscraper would use that instead of urllib2.
00:03 paulproteus In that case, can I suggest a next step for you, that is to read the OpenHatch profile_importer.py file and customs/management/commands/customs_twist.py until you understand them?
00:04 Perhaps with me on the phone as you say "wtf?!??" a few times and I explain things? :P
00:04 willkg i could do that...
00:04 paulproteus I could read flashscraper to sanity-check my claim that I have the answer.
00:04 Can you give me an e.g. git link?
00:04 I'm happy with having that as a next step for me.
00:05 willkg i wouldn't read flashscraper.  it's icky.  mostly it's just a bunch of work linked together by callbacks and errbacks.
00:05 paulproteus That's fine, but I still want to see it!
00:05 willkg mostly it's gross.  so gross, webber looked at it and said, "ew, gross" and wrote vidscraper.
00:05 oh, he's even online.
00:05 paroneayea: how gross is flashscraper in 5 words or less?
00:05 paulproteus http://git.pculture.org/miro/t[…]per.py?h=Miro-3.5 here we go
00:06 willkg https://git.participatorycultu[…]b/flashscraper.py
00:06 the one in git master is the same as the one in the Miro-3.5 branch.
00:06 i dislike touching it.
00:06 paulproteus You might also want to read the (icky) test suite for the OpenHatch stuff, in mysite/customs/tests.py -- the Github one is a good example file to look at.
00:07 Your flashscraper.py is surprisingly (and pleasingly) short.
00:07 willkg wait...  is this a sneaky way to get me to read openhatch site code?!
00:08 paulproteus Honestly, no. (-:
00:08 willkg you should use this in one of your "sneaky ways to reduce the barriers to entry" slides.
00:08 paulproteus "Distract people from their current problem by forcing them to solve yours"
00:08 s/forcing/tricking/
00:08 You know.
00:08 willkg heh.  :)
00:09 paulproteus Anyway, I think there's years of "how not to do it" wisdom in flashscraper.py and the old, synchronous OpenHatch profile import code, and if we are reasonable, we can expose that as a library and the world will be a better place. Mostly importantly, our worlds.
00:09 willkg ok.  i've got to jet.  baby crying.  i'll read through these tomorrow morning and get back to you with either a bunch of wtf?! questions, a bunch of patches for openhatch, suggestions for vidscraper, or something like that.
00:09 paulproteus bd
00:09 willkg true that.
00:09 paulproteus I'm totally game for being on the phone as if walkie talkies so you can just speak up
00:10 willkg actually, the bad part of flashscraper is that it's hard to debug because "deferreds" in miro don't tell you a lot about where they came from.
00:10 paulproteus willkg: go jet.
00:10 willkg ok.  rock on everyone!  see you all tomorrow!
00:10 paulproteus (-: You too!
00:10 willkg left #miro-hackers
00:15 bob999 left #miro-hackers
00:41 bendk left #miro-hackers
01:29 maggie_s left #miro-hackers
02:15 Odysimus left #miro-hackers
02:34 Odysimus joined #miro-hackers
02:40 bendk joined #miro-hackers
02:43 bendk left #miro-hackers
03:15 kcw left #miro-hackers
03:30 ajonas joined #miro-hackers
04:23 ajonas left #miro-hackers
04:27 ajonas joined #miro-hackers
05:33 ajonas left #miro-hackers
05:55 JasonWoof left #miro-hackers
05:56 JasonWoof joined #miro-hackers
06:08 ajonas joined #miro-hackers
06:36 kcw joined #miro-hackers
06:38 ajonas left #miro-hackers
06:59 CarlFK left #miro-hackers
08:43 DGMurdockIII left #miro-hackers
09:17 janetPCF1 left #miro-hackers
09:50 janetPCF joined #miro-hackers
10:47 kcw left #miro-hackers
14:24 willkg joined #miro-hackers
14:47 z3p good morning
14:55 willkg hi!
14:58 z3p paulproteus: don't know if you're here, but ping me if you still need help getting your tabs to show up correctly
14:58 paulproteus I am quite here.
14:58 That would be very helpful.
14:58 I am cold, though; brb thermostat!
14:59 z3p paulproteus: it's all CSS, based on the ID and CSS class on the <body>
14:59 ajonas joined #miro-hackers
15:00 willkg know what i want?  i want clothes to be described using css properties.  that'd make it a lot easier to buy stuff online without worrying about whether the color on my monitor is quite right.
15:00 plus i could automate clothes-buying with a script.
15:00 paulproteus You can already do both of those things!
15:00 Just stop worrying at all!
15:00 z3p: Okay, let me see what the other pages do.
15:01 tty in 3 min
15:01 z3p paulproteus: check out admin.css:671
15:02 paulproteus Got it.
15:14 janetPCF left #miro-hackers
15:22 arpu left #miro-hackers
15:31 paulproteus ajonas: I made some more tiers-y changes.
15:31 The dev server will take ca. 10 minutes to notice.
15:31 I'm going to work on your points about the admin count stuff now.
15:31 ajonas paulproteus: awesome, i will check them out
15:31 paulproteus Hah, you can't! :P:
15:31 Whoa, more punctuation than I wanted. Ignore one of those ":"s. (-:
15:39 willkg afk for cooooooffffeeeeee.
15:41 paulproteus paroneayea: So I don't really live the Emacs Dream.
15:41 Two reasons.
15:42 1. Syntax highlighting for Python goes wrong for me sometimes.
15:42 2. "It starts slowly"
15:42 3. Configuring it seems like a bother.
15:42 I wonder if I'm worth converting.
15:42 paroneayea 1) not sure if I've had that problem, would be interested in examples (maybe that was the old python-mode, rather than the newer builtin one?)
15:43 2) Keep the same instance open for a month, just use emacsclient (possibly with -t or -C) to open up new files from the command line
15:43 z3p paulproteus: 1. editing something around the break fixes it for me
15:44 paulproteus paroneayea: I'm basically sure it's the old python-mode.
15:44 paroneayea 3) just configure it when you need it, and don't use M-x customize if you can avoid it imo :)
15:44 paulproteus It's the '''Look, more apostrophe's caused the highlighting to end''' issue.
15:45 I think I've had that problem since 2001.
15:45 And now nano on Debian comes with syntax highlighting by default that's even better than my nanorc.
15:45 * paulproteus apt-get removes python-mode
15:45 paroneayea paulproteus: yeah use the newer python-mode for emacs
15:45 the one that comes bundled
15:45 muuuuuch better
15:45 paulproteus Oh, and I can never remember how to edit my ~/.emacs with Emacs and cause it to live-reload it, so I always quit and restart it.
15:46 paroneayea M-x eval-buffer
15:46 or, when you add something new
15:46 go to the end of the newly added paren group and C-x C-e to evaluate it
15:46 paulproteus Okay, just did server-start.
15:46 paroneayea I usually add new things one by one and evaluate them
15:46 paulproteus Eval-buffer sounds nice enough.
15:47 But your way is okay too!
15:47 paroneayea so another thing, once your ~/.emacs gets big enough, and you want to use the same emacs setups on multiple machines that might not all have the same packages installed
15:47 here's how I do my emacs configuration
15:47 paulproteus Whatever, for now I have one computer, my desktop. (-:
15:48 paroneayea git repository in ~/.config/emacs/, in there a file called general.el and all the different mode customizations in modes/org.el and etc
15:48 and include the ones you want for that computer in your ~/.emacs like
15:48 (load-file "~/.config/emacs/modes/erc.el")
15:48 paulproteus la la la la I'm ignoring the terrifying even if powerful complexity you are describing
15:48 paroneayea hahaha
15:48 okay
15:48 sorry
15:48 paulproteus (-:
15:49 paroneayea: Also, find-replace in Emacs is weird.
15:49 And there's no coy, only cut.
15:49 No copy!
15:50 Unless there is, and I've been missing it for the past 10 years or so. Actually I remember learning about it and then forgetting it.
15:50 I used to rely heavily on the graphical clicky things in the Emacs GUI.
15:50 paroneayea there's a copy
15:50 paulproteus M-w or something.
15:50 paroneayea M-w
15:50 yeah
15:50 emacs has ancient terminology though that sounds like you're battling zombies
15:50 paulproteus Also, why is regex search-replace so hard?
15:50 It's so easy in nano.
15:50 paroneayea killing and yanking text from the kill-ring
15:51 yeah regex search-replace is kind of irritating.
15:51 I always forget how much I need to escape things
15:51 so I use regexp-builder
15:51 paulproteus AKA http://emacsblog.org/2007/03/1[…]e-tip-re-builder/ ?
15:52 paroneayea yup
15:52 paulproteus a) Nice, XEmacs has perl-compatible regular expressions.
15:52 b) Wait, GNU Emacs doesn't anywhere in there?
15:52 c) Wait, XEmacs is still alive?
15:52 paroneayea :)
15:52 yeah, all 5 devoted users
15:53 paulproteus brb, calories.
15:53 paroneayea also, http://www.masteringemacs.org/[…]tion-to-ido-mode/
15:54 bendk joined #miro-hackers
15:55 willkg paroneayea: some day, i would love love love it if you wrote up a blog entry with a list of links to good emacs-fu resources.
15:56 janetPCF joined #miro-hackers
15:57 paulproteus paroneayea: Wait, emacsclient opens up the file in the editor I started originally?
15:57 I wanted an editor in the terminal I am typing in.
15:57 paroneayea emacsclient -t
15:57 willkg: ok, I'll do that at some point!  Though usually I just search emacswiki.org :)
15:58 paulproteus http://pastebin.com/6TADAQHF # AKA invalid option
15:58 I'll do some googling
15:58 paroneayea you're using an old version of emacs I bet!
15:58 pre-multi-tty
15:58 M-x emacs-version ?
15:58 paulproteus Twenty two point three point one.
15:59 * paulproteus upgrades and gets eamcs23
15:59 paroneayea I think 23x is the multi-tty version
15:59 yeah multi-tty means you can have an X session of emacs open, and ssh into your machine and still open up console interfaces to it!
15:59 and it'll be the same session!
15:59 paulproteus Okay.
16:00 That doesn't actually sound all that amazing, but okay.
16:00 paroneayea paulproteus: willkg: btw, here's my list of absolutely mandatory .emacs things. http://paste.lisp.org/display/118396
16:00 it was a huge deal to me, considering how much I keep open in emacs at once, that I could connect to my main session at any time
16:01 paulproteus Okay, wow, the fonts in Emacs 23 are even nicer than in 22.
16:01 * willkg nods.
16:01 paulproteus I am suddenly more interested in using GUI emacs rather than locking it up inside a terminal.
16:01 paroneayea oh yeah, the anti-aliasing
16:01 willkg i keep meaning to put my emacs stuff in my gitorious thingy.
16:01 paroneayea yeah I've been wondering if I should push my emacs repository public
16:02 I have all my private stuff outside of the directory, so it wouldn't be "unsafe"
16:02 paulproteus Weird, the command line highlighting for diffs is more colorful than the GUI one.
16:06 ajonas: For the upgrade workflow with PayPal --
16:07 When you click "Upgrade", should it take you straight to PayPal?
16:07 ajonas paulproteus: "Upgrade" on a particular tier, y?
16:07 paulproteus Instead, it could take you to a confirmation page that says "Look at this awesome stuff you're going to get, and again, here's how much it'll cost".
16:07 Yeah -- each tier has its own "Switch to this" link.
16:07 ajonas paulproteus: seems to me if they click on a particular tier it shoudl be clear, and paypal will show them the amount again before they confirm
16:08 paulproteus I'm thinking through it because we'll definitely need that page for downgrading.
16:10 ajonas: Okay -- so I'll make that page for downgrading, but not for upgrading.
16:10 ajonas paulproteus: is that still needed with "switch to this" ?
16:10 i thought that was just for when we had a downgrade link
16:11 paulproteus: or is it that we wanted to warn people what they were losing?
16:11 paulproteus Well, whatever the link is called, if you have 1000 admins and you're only allowed 5 in the tier you "Switch to" or "Downgrade" to, -- yeah, to warn people
16:12 I think we should warn people just before the downgrade, and actually give them a chance to cancel it if it will change the way their site works.
16:12 ajonas paulproteus: ok, i think a page is fine, and once they do it we should also send them an email
16:12 paulproteus If it won't change the way their site works, I'm totally cool with jumping right to the new tier (with paypal first, if it's a paid tier).
16:12 Great, warning page and email coming up.
16:13 paroneayea: One thing I frequently do is "git grep -l $PATTERN" then open said file with nano.
16:13 What's the emacsism for that?
16:14 z3p paulproteus: M-x rgrep
16:14 paulproteus But since it's not accelerated by git pack files it'll probably be sad and slow.
16:14 * paulproteus tries
16:14 z3p paulproteus: oh, git grep
16:15 paulproteus: I use it on the Miro codebase all the time without a problem
16:15 willkg ok...  time to move the binary kits.
16:15 paulproteus Yeah, seems pretty fast here actually.
16:15 Click to jump to result is pretty exciting.
16:17 paroneayea M-x grep
16:17 yeah
16:17 oh
16:17 git grep
16:17 hm!
16:17 paulproteus It's "okay" that it's not git grep. I can live for now.
16:17 paroneayea but yes M-x grep *is* very exciting
16:17 paulproteus In the OpenHatch directory trees, that won't be okay probably, but for here it's fine.
16:18 willkg paulproteus: woah--i'd never seen git grep before.  that's superb!
16:18 paulproteus (The buildout-generated parts and eggs will cause false positives and slow the world down.)
16:18 willkg i have to go dark to get work done....  afk for a bit.
16:18 paulproteus bd
16:21 paroneayea: Can I get focus-follows-mouse in Emacs buffers?
16:21 That would be a dream come true..
16:21 z3p paulproteus: yeah, there's a setting
16:21 paulproteus http://stackoverflow.com/quest[…]-buffers-in-emacs # whoa
16:22 z3p paulproteus: (mouse-autoselect-window t)
16:22 paulproteus b
16:22 Oh MY GOD!!! Yes!
16:22 This is so awesome!
16:22 paroneayea wow scary (and cool)
16:22 paulproteus Okay, now emacs is worth using.
16:22 z3p paulproteus: just don't tell people you're using the mouse with emacs; they'll freak out :)
16:23 paroneayea btw if you're gonna middle-mouse-paste
16:23 (setq mouse-yank-at-point t)
16:23 I think is mandatory
16:23 otherwise it doesn't paste at point, and often pastes where you don't want, and is hard and irritating to control
16:24 paulproteus paroneayea: already did that.
16:24 paroneayea :)
16:25 paulproteus Heck, if I get focus follows mouse, I could stay in emacs all day.
16:25 It's like focus follows mind. I love it.
16:25 paroneayea focus-follows-gaze
16:26 oh nice
16:26 speaking of themes
16:26 looks like themes are finally officially part of emacs in current bzr trunk
16:27 paulproteus I'm using M-x shell to SSH somewhere, and I want to open a file for editing.
16:27 * paroneayea can finally toss out his old color-theme.el!
16:27 paulproteus I might as well use that convenient emacs-only sshfs-type-thing to do it.
16:27 Is there something reasonable I can do/configure to make that smooth?
16:28 paroneayea tramp!
16:28 /ssh:staging.creativecommons.org:/var​/www/staging.creativecommons.org/cc.e​ngine_stage/cc.engine/cc.engine.ini
16:28 paulproteus TRAMP, that's the sshfs thing.
16:28 paroneayea not sure about in combination with M-x shell though, which I find to not be very good with ssh
16:30 I hear http://www.emacswiki.org/emacs/ClusterSSH or http://www.splode.com/~friedma[…]s-lisp/src/ssh.el is better for M-x shell
16:30 though honestly
16:30 I M-x ansi-term
16:30 and ssh in
16:33 paulproteus What's the best way to get a tramp directory listing?
16:34 willkg paulproteus: just open the directory.
16:34 paulproteus All I really do with SSH is cd around, run ls, and sometimes edit files.
16:34 b
16:34 z3p paulproteus: mx dired
16:35 paroneayea yeah dired works pleasantly well with ssh
16:35 z3p (getting lazy because pidgin keeps converting M-x to AM-x)
16:35 paulproteus No prob
16:35 So now I have a bajillion buffers. Plague of buffers!
16:35 I want to close most of them.
16:36 It would really be nice to say, "Close all the ones I opened in the past 2 minutes."
16:36 Or maybe set a M-x bookmark-buffers, ... open a bunch ... , M-x close-all-since-bookmark
16:36 paroneayea I usually clean things out with C-x C-b when necessary but
16:36 honestly
16:36 paulproteus Then I go back to the state at bookmarktime.
16:36 paroneayea I don't really close buffers very often.
16:36 z3p just leave them open
16:36 paroneayea leaving them open is the emacs way :)
16:36 * paulproteus gulps.
16:37 paroneayea if you have ido-mode enabled navigating between buffers is easy
16:37 paulproteus I'll do that, you nutcases, just for now.
16:37 I know a guy named Ido.
16:37 I don't have a mode.
16:37 Okay, enabled.
16:37 I'll read about it.
16:37 z3p right now I've got 64 buffers; I'm sure paroneayea has more
16:37 paroneayea 242 buffers
16:37 paulproteus WHOAAAAAAAAA
16:37 ido mode is crazy awesome
16:37 paroneayea yeah
16:37 it is :)
16:37 paulproteus Jesuuuuuuuuuuuuuus!
16:38 paroneayea http://www.masteringemacs.org/[…]tion-to-ido-mode/ has some good tips
16:38 paulproteus Where have you people BEEN all my life?
16:38 paroneayea the main thing I think to know is that C-s cycles through the results
16:38 so if what you really want is just the next one over
16:38 just C-s
16:38 paulproteus Got it, cool.
16:45 bendk1 joined #miro-hackers
16:49 bendk left #miro-hackers
16:51 * paroneayea updates his localtv install!
16:55 paulproteus paroneayea: :D
17:01 bendk1 left #miro-hackers
17:04 paulproteus paroneayea: When I scroll with the mouse wheel, I want to leave the cursor (AKA "point", I think) where it is.
17:04 Is that possible?
17:04 So I can scroll, read some text above, and then keep typing back where the cursor was.
17:05 paroneayea paulproteus: not afaik, but you could C-space to set the mark, scroll, then C-x C-x to get back there
17:05 paulproteus Oh, yeah, I radically don't understand "the Mark".
17:05 * paulproteus tries this
17:05 paulproteus What? But now stuff is selected.
17:06 paroneayea the region is active, yeah
17:06 paulproteus That sounds terrifying.
17:07 paroneayea you can probably disable transient-mark-mode's visibility of the region somehow
17:07 not sure
17:07 at the least, you can do this
17:07 C-space, C-g (to cancel the "selection"), scroll, C-x C-x, C-g
17:07 paulproteus Jesus Christt.
17:07 I'll try it.
17:07 paroneayea haha
17:08 paulproteus Jesus, that "worked".
17:08 And here I thought Emacs was customizable.
17:08 I guess that's a meaningless non-sequitur.
17:08 Anyway.
17:08 paroneayea sure, script your own solution! ;)
17:09 but anyway, to answer the question, the "point" needs to be within visibility
17:09 of your current window
17:09 so you can't scroll without it moving
17:09 paulproteus That's what I was afraid of.
17:09 So isn't that, like, quite annoying for you also?
17:09 paroneayea no
17:09 paulproteus With nano, that's also true, but I forgive nano.
17:09 paroneayea so usually what I actually
17:09 do
17:09 is not do what I said.
17:10 usually I C-x 2
17:10 which splits the window
17:10 look for what I want
17:10 then C-x 0
17:10 funny, I didn't realize that's what I always do until I said this!
17:10 I just do it :)
17:10 paulproteus Interesting. Now my relative window sizes have been re-arranged.
17:10 Other than that, it's workable.
17:11 paroneayea I do a *lot* of window-splitting
17:11 my window setup is rarely the same between every couple of minutes
17:11 * paroneayea very jumpy
17:12 paulproteus So I just apt-got magit.
17:12 I hear you don't have to restart Emacs all the time to change things.
17:12 Is there another way to cause it to notice magit?
17:13 [12:07:30] <Freenode:#miro-hackers:paulproteus> I hear you don't have to restart Emacs all the time to change things.
17:13 [12:07:35] <indymedia:#techfed:Clint> people are trained to throw money at problems
17:13 Hmm.
17:13 paroneayea hm, for installing new libraries with apt, it won't notice it on the load path itself, so you'll have to fine out where it added it to, (add-to-list 'load-path "/path/to/magit/")
17:13 then (require 'magit)
17:14 if you don't want to restart
17:14 usually when I'm doing something throw-away like that I do it in *scratch*
17:14 brb
17:14 paulproteus b
17:14 <3 for your hand-holding.
17:25 z3p http://whatthecommit.com/
17:26 janetPCF left #miro-hackers
17:36 paroneayea z3p: awesome
17:39 paulproteus "No changes made". That's a winner!
17:43 paroneayea fixed the israeli-palestinian conflict
17:43 that's my favorite
17:43 whew, conflict resolved
17:44 janetPCF joined #miro-hackers
17:50 willkg i just moved binary kits to a new happy place.  i sent an email to develop about it and updated the docs.
17:51 paulproteus (-: willkg
17:53 ajonas: w/r/t your latest comment on http://bugzilla.pculture.org/s[…]_bug.cgi?id=14796 -- if today I am paying, and tomorrow I decide I want to keep my site but stop paying, I think there should be a "Switch to basic" option that I can click.
17:53 ajonas paulproteus: yep, that's what should replace the "upgrade" button in the basic bar
17:53 paulproteus Oh, right-o, yeah.
17:54 I already kind of did that, but the dev server didn't update.
17:54 Let me refresh there and show you the latest.
17:54 ajonas paulproteus: cool
17:55 willkg afk for a much needed mid-day meal.
17:55 paulproteus Also, oh my effing god, the PDB integration is a mind-blower.
17:55 in Emacs.
17:59 ajonas paulproteus: i'm going offline for lunch and hopefully to install a new hard drive, then i'll check out the changes
17:59 paroneayea oh yeah
17:59 the pdb tracking in M-x shell?
17:59 gr8 stuff
17:59 paulproteus ajonas: Okay -- tty then
17:59 ajonas left #miro-hackers
18:00 paulproteus So I'm going to have, like, up to test<13> if I don't start closing buffers.
18:00 tests.py<13> rather.
18:00 It'd be nice if it picked a different naming convention. I guess that's probably actually customizable.
18:01 paroneayea paulproteus: okay, so :)
18:01 there's an awesome solution to that!
18:01 * paulproteus gulps.
18:02 paroneayea (require 'uniquify)
18:02 (setq uniquify-buffer-name-style 'reverse)
18:02 (setq uniquify-separator "/")
18:02 (setq uniquify-after-kill-buffer-p t) ; rename after killing uniquified
18:02 (setq uniquify-ignore-buffers-re "^\\*") ; don't muck with special buffers
18:02 run that stuff
18:02 and you'll *always* have useful names
18:02 none of that <number> stuff
18:02 * paulproteus runs arbitrary code from some person in an irc channel.
18:02 paroneayea some person!
18:03 paulproteus SWEET
18:03 paroneayea :D
18:04 paulproteus No one told me that *my* Emacs could also be a very nice place, on my terms! I thought only bearded weirdos got to have that.
18:04 paroneayea paulproteus: can I quote you on that
18:04 paulproteus I'm kind of serious, but I now regret calling people weirdos.
18:05 I guess so, sure. (-:
18:07 The big problem with Emacs before was that the fonts were too big or too ugly.
18:07 I guess things have changed.
18:09 paroneayea antialiasing ftw
18:10 paulproteus: I badly attempted to quote you twice on identi.ca
18:10 * paulproteus searches on collecta
18:10 paroneayea the first time I got your username wrong, the second time I didn't put quotes around it because your username is so long
18:17 paulproteus http://pastebin.com/ud5JgFz2
18:17 With this output in the shell, can I jump to that file:line ?
18:20 bendk joined #miro-hackers
18:21 z3p paulproteus: I think there's an emacs-friendly unittest format
18:23 paulproteus Okay, that one I'll deal with later (since after some Googling, nothing obvious shows up)
18:23 z3p paulproteus: I know trial does
18:23 paulproteus I wish it were easier to get unittest through Django to put you into a pdb at the exception, I suppose.
18:24 z3p: (-:
18:24 re: trial
18:26 paroneayea paulproteus: running nosetests?
18:26 --pdb --pdb-failures
18:26 paulproteus No, the Django test runner.
18:26 paroneayea ah
18:26 paulproteus I could use a non-default one, but gah mumble grumble.
18:26 paroneayea paulproteus: btw, I'm really impressed with your mc tutorial
18:27 paulproteus ?
18:27 paroneayea even if I didn't know how django worked and stuff, I could set this up
18:27 https://develop.participatoryc[…]ro_Community_site I mean
18:27 paulproteus :D
18:27 I used to not really know how Django works. Have you *seen* the OpenHatch code? (-:
18:27 paroneayea heh
18:28 paulproteus But yeah, I'm also writing it vaguely as if https://openhatch.org/blog/201[…]-the-buildhelper/ existed.
18:28 But also -- thanks! I appreciate you saying that!
18:29 ajonas joined #miro-hackers
18:29 paulproteus ajonas: So the "Switch back to basic" button already has that text.
18:29 But there's a big "Upgrade now" image that appears on *top* of the text.
18:29 (Note: it really says 'uprade now' [sic])
18:31 ajonas paulproteus: ok, so can you just take out the image? or will it need more smoothing over from matt?
18:31 paroneayea paulproteus: django.core.exceptions.ImproperlyConfigured: You must define the HAYSTACK_SITECONF setting before using the search framework.
18:31 paulproteus Zing!
18:31 paroneayea nothing in the guide teaches me how to set this!
18:31 paulproteus Sounds like a bug in the example project settings.
18:31 paroneayea ah I see
18:32 I should just copy that in
18:32 I'll fix the wik8i
18:32 er wiki
18:32 paulproteus Sweeeeeeeeet.
18:33 paroneayea hm
18:33 django.core.exceptions.ImproperlyConfigured: ImportError haystack: No module named search_sites
18:33 paulproteus ajonas: Looking into that now...
18:33 paroneayea paulproteus: I'm not familiar with haystack, know what this error is?
18:33 z3p paroneayea: you need the search_sites.py in your project
18:34 paroneayea: should be able to just copy it from the example project too
18:34 paulproteus You can probably find it in example_project/ in the repo.
18:34 paroneayea oh heh
18:34 paulproteus Also, we should move the example_project to be within localtv as localtv.example_project.
18:34 ajonas paulproteus: we should just make a list of things for matt to look at, this seems liek one of those
18:34 paulproteus That way it won't have depressing namespace collisions.
18:34 ajonas: That'd be even better by me!
18:34 paroneayea oic
18:35 so we need to also add to the wiki that we need to add those variables
18:35 but also change them to localtv.foo instead of example_project
18:35 paulproteus I don't quite understand which variables you mean, but if you think you're right, do it and hit save and then show me the delta.
18:36 paroneayea k
18:39 so additional problem:
18:39 kcw joined #miro-hackers
18:39 paroneayea Not synced (use migrations):
18:39 - localtv
18:39 - localtv.playlists
18:39 (use ./manage.py migrate to migrate these)
18:39
18:39 but also:
18:39 File "/home/cwebber/env/localtv/src/miro-community/lo​caltv/migrations/0004_remove_openid_fields.py", line 56, in forwards
18:39 db.alter_column('localtv_openiduser', 'user_id', orm['localtv.openiduser:user'], ignore_constraints=True)
18:39 TypeError: alter_column() got an unexpected keyword argument 'ignore_constraints'
18:39 after trying to do manage.py migrate
18:39 so the doc says south is used but actually no not really
18:40 but the database isn't really set up after a manage.py syncdb
18:40 and manage.py migrate doesn't work as it is
18:40 paulproteus Have to bump up South to 0.7.3
18:41 paroneayea I am!
18:41 paulproteus But, but, ignore_constraints works for me!
18:41 paroneayea >>> import south
18:41 >>> south.__file__
18:41 '/home/cwebber/env/localtv/lib/python2.6/site-pa​ckages/South-0.7.3-py2.6.egg/south/__init__.pyc'
18:41 paulproteus Maybe I added it to some migrations where I shouldn't have. But I'd be surprised by that, but it's possible.
18:42 z3p paroneayea: it's not required, but if you have it installed you have to use it
18:42 paroneayea all these migration projects are "necessary" but also "don't work"
18:42 ah
18:42 I'll just disable it then
18:42 z3p paroneayea: if it's not installed, the django syncdb DTRT
18:42 * paulproteus sighs.
18:43 paulproteus south 4eva!
18:43 But sure, paroneayea, feel free to skip it for now.
18:43 Lunch o'clock.
18:44 willkg paroneayea: OH MY GOODNESS!  that buffer naming thing helps _so_ much.
18:44 paroneayea willkg: :D
18:44 willkg paroneayea: previously, i was using some buffer name renaming thing you wrote a year ago or so.  but it required me to do explicit things which i don't usually do.
18:44 paroneayea yeah
18:44 I was using that too
18:44 the rename-buffer-with-directory command I wrote
18:45 willkg paulproteus: i'm working on vidscraper stuff now.  about to go read through all your openhatch code.
18:45 paulproteus willkg: /me gulps
18:45 paroneayea I heard about uniquify and was like, "well I'm sure that's not as good as my function because there's no way it's as good as mine, because I'm sure it doesn't have tramp support"
18:45 but it does have tramp support
18:45 and i was wrong.
18:45 willkg: what's vidscraper being used for other than localtv these days, btw?
18:45 paulproteus uniquify is like a setting on a blender.
18:45 paroneayea or is it just localtv
18:46 willkg paroneayea: i'm about to overhaul miro to use vidscraper instead of flashscraper.
18:46 paulproteus I'm going to prep lunch, willkg, so if you want me with much urgency I might be afk for a bit.
18:46 z3p paroneayea: usubs
18:46 paroneayea :D
18:46 paulproteus You could ring me to get my attention.
18:46 willkg paulproteus: that's fine.  i'll chat with you some day about this stuff.  :)
18:46 paroneayea ok, database set up
18:46 I'm also going to lunch
18:46 thx for the help z3p, paulproteus ;)
18:46 bbiab
18:46 paulproteus I think that really there's some better-than-what-openhatch-d​oes-but-similar-in-many-ways abstraction we should arrive at.
18:47 willkg z3p: thank you soooo much for fixing 14332.  i kept meaning to get around to it, but it was low on my list.
18:47 z3p: but it was annoying as all hell.
18:47 z3p willkg: did it actually work?
18:47 willkg z3p: i thought you tested it...  ;)
18:48 z3p: i'll test it now.
18:48 z3p it compiles.... :)
18:48 willkg: but I couldn't actually get a torrent that would display the behavior
18:48 willkg i figured you tested it because you checked in a fix and then marked the bug as fixed.
18:48 z3p willkg: the commit message is 'blind fix for #14332'
18:49 willkg ...  oh, you're right.
18:49 ok, i'll test it now.
18:49 z3p but you're right, I should have mentioned it in the bug
18:49 willkg no biggie.
18:50 is clearbits the new legaltorrents?
18:50 z3p willkg: I think so
18:51 willkg the downloading tab is still horked.  boo....
18:52 z3p willkg: is there a ticket for that?
18:52 willkg: I think I figured there was, but don't actually see one
18:53 kcw z3p: it's 15754
18:54 I broke it.... I'll have it fixed soon
18:54 willkg kcw: but that's marked as a duplicate of the one paul just fixed.
18:54 kcw: cool--thanks!
18:54 z3p kcw: no, that's different
18:54 willkg: what problem are you having?
18:54 willkg the downloading tab doesn't show any items.
18:54 z3p willkg: okay, yeah, that's my problem too
18:54 willkg the tab has an orange counter, so miro thinks it's downloading things (i think), but there's nothing listed in the view.
18:56 z3p git bisect, here we come!
18:56 willkg z3p: i think kcw said he was working on it.
18:56 z3p willkg: no, I think he means 15754, which I fixed
18:56 willkg: and which is a different issue
18:57 willkg kcw: ^^^ ?
18:59 kcw I think this one's still me
19:04 ajonas paulproteus: all the dev sites seem to have custom theming turned on - maybe we can turn it off for one (floyd dev or anne2 would be fine) to test that?
19:04 back in a few
19:07 bendk okay, I'll try to fix everything I broke yesterday with signals
19:07 I wasn't expecting so much interesting things to happen with those changes
19:08 willkg bendk: fun fun fun when you pull back the carpet and look at what's underneath!
19:17 bendk there's something really baffling about our python stack on cocoa
19:18 that bug report says we're nesting the signal emmission, but I only see it being emitted once on the stack trace
19:19 okay fine, there's cocoa callbacks involved and whatever, I guess that could make sense.  But then I added code to print out a stack trace whenever we emit size-allocated, and I get a printout of about a dozen calls where traceback.print_stack() prints out nothing at all
19:20 willkg weirdness.
19:20 phantom emissions!
19:21 bendk well, I tracked it down anyways, I just don't understand why the print_stack() is failing, there's definitely a stack involved when I'm calling it
19:25 so the issue is that for list view, the handler for size-allocated sets the columns to their initial widths.  But changing the column width causes invalidate_size_request(), which causes signals to propate for a while and eventually ends up in another size-allocated call
19:26 there's actually code in do_size_allocated to avoid the infinite loop there
19:26 I can think of 2 fixes, I'd love some feedback if they make sense or not
19:26 willkg bendk: can you point out the file and line?
19:26 bendk itemlistwidgets.py in do_size_allocate
19:27 1) don't call invalidate_size_request when the column width changes.  I don't think that resizing the columns should actually make the table request a bigger/smaller width
19:28 2) make invalidate_size_request defer it's action for a while, I think this is good in general because we might do several things that make the size request invalid
19:28 line 361
19:29 willkg i'm puzzled.  does this mean that the "if not ..." block is getting executed twice?
19:29 bendk well the if gets executed twice
19:30 but it immediately sets _set_initial_width=True, so the second call doesn't cause an infinite loop
19:30 willkg ok.  i get what you're getting at.
19:31 i think i'd be +1 on making invalidate_size_request defer its action.  i'm not sure why resizing columns shouldn't affect the table width--i think it should.
19:31 though, i don't use the list view, so i don't know how it works.
19:32 z3p willkg: it depends on how you're resizing them
19:32 willkg er, how it works behaviorally.
19:32 bendk hmm, now that you mention it, I guess it should in theory
19:32 if I had an hbox with a table on the left and a label or something on the right, and I resize a column bigger, then i guess the table should get more room
19:33 actually, now that I think of it more, it should always call invalidate_size_request
19:39 paulproteus ajonas: Cool, I'll turn it off for floyd-dev and anne2.
19:49 bendk wow, sweet the linux error is a real bug
19:49 willkg i'm not sure.
19:49 i can't tell if it's a problem in miro or a problem with some torrents.
19:49 i can't reproduce it with the democracy now torrents--those all work fine.
19:49 i can't reproduce it with the torrents listed in the bug--those work fine.
19:49 but i can reproduce it with a couple of torrents i found on clearbits.
19:50 where "it" is miro starts downloading the torrent, then peters out and then the torrent goes back to a "starting up" phase.
19:50 wireshark sees continued bittorrent handshakes, so it looks like libtorrent is still trying to do things.
19:52 bendk one thing I noticed is that there is sometimes confusion between torrents starting up and 0bps downloads
19:52 that's from my memory though, I'm not sure if it's still true
19:53 paulproteus ajonas: floyd-dev and anne2 are both set to check the tier for custom template permission.
19:53 * willkg nods.
19:54 willkg bendk: i'm adding some debugging code so i can figure out what the downloader is doing when things happen.
19:54 ajonas paulproteus: cool, ty
19:54 willkg bendk: depending on how that goes, arvid said there's libtorrent logging.  so i'll look into turning that on so libtorrent can tell us what's going on.
19:54 man...  today has been one diversion after another.
19:55 afk a smidge for coffee....
20:03 paulproteus Good idea.
20:05 paroneayea paulproteus: another comment from having just installed the tutorial
20:06 it doesn't teach the user that they have to set up a sitelocation
20:06 paulproteus They don't!
20:06 Isn't it automatic?
20:06 It should be on the latest master, anyway.
20:06 paroneayea hm, well if I go to the root I get
20:06 paulproteus It's one SiteLocation per database now, suckas!
20:07 By "should be" I only mean that "it's a bug if it's not"
20:07 paroneayea Caught VariableDoesNotExist while rendering: Failed lookup for key [sitelocation] in u'[
20:07 paulproteus Eep.
20:07 Oh -- right.
20:07 paroneayea is what I get while visiting /
20:07 paulproteus You have to enable a piece of middleware.
20:07 That's new and not documented.
20:07 paroneayea :)
20:07 so what's the middleware
20:07 paulproteus localtv.SiteLocationMiddleware # should be in example_project/settings.py
20:08 P.S. I appreciate your patience!
20:08 paroneayea oh, that's enabled
20:08 still not working!
20:08 time to look at the code I guess :)
20:09 paulproteus D'oh.
20:09 Yeah, guess so.
20:10 paroneayea okay, so this middleware
20:10 what should it do
20:10 add request.sitelocation I'm guessing?
20:11 paulproteus Yeah, and also create an object if the object doesn't exist in the DB.
20:11 paroneayea ok
20:11 so also
20:11 paulproteus There should be a comment along those lines.
20:11 paroneayea so yeah, there's nothing in the context to provide sitelocation
20:11 so I think it's correct in having that error?
20:11 since it's trying to cache sitelocation and that doesn't exist
20:11 also, cache_invalidator apparently doesn't exist either?
20:12 bendk willkg: fixed that downloader unittest failure
20:12 about to check in it once I make sure the rest of the suite passes
20:12 willkg bendk: you rock!  thank you!
20:13 paulproteus paroneayea: Hmmph!
20:13 Is f658f13a2ba700aeb1648e5d8efcb245746dc872 an ancestor of your current git commit?
20:13 paroneayea yup
20:13 paulproteus I'm going to try to finish something, then talk to you in ca. 15 min.
20:13 paroneayea paulproteus: sounds grood
20:14 paulproteus If you want faster service, it'll have to be self-service. (-:
20:14 paroneayea I'll keep looking
20:16 paulproteus ajonas: If a user has custom CSS, but they switch to a CSS-less tier, what happens to their old CSS? Some possible answers:
20:17 1. We still serve it out, but we refuse new changes. (my preference)
20:17 2. We delete their old CSS from the database (and we warn them, first!) (also easy)
20:17 paroneayea paulproteus: okay, that one was pebkac, sorry!
20:17 forgot to set up the context processor variable
20:17 paulproteus paroneayea: what was the pebkac error?
20:17 Oh, okay.
20:17 Context, cool.
20:17 ajonas paulproteus: i think we went through and documented this somewhere, let me look - i think #1 is too lenient, it'll likely be more like #2
20:18 paulproteus ajonas: Yeah, sorry to be a little dense about keeping these decisions straight.
20:18 ajonas: 3. We keep their CSS in the database, but refuse to *show* it.
20:18 ajonas paulproteus: np, i definitely am as well
20:19 paulproteus: that would be the ideal, i think
20:19 paulproteus I like 3 the most, but I'll need to check that it's actually reasonably easy.
20:19 Yeah, I'll give that a quick look now.
20:20 willkg paulproteus: i'm not sure i can get to vidscraper stuff today.  it's been a day filled with side-tracks and deviations.
20:20 paulproteus willkg: yup
20:20 willkg paulproteus: it's not a big rush, though.  and looks like you're busy, too.  so...  can we push this off until monday or something?
20:20 paulproteus Sure, willkg.
20:20 willkg paulproteus: we can rename this the "kick-the-can-down-the-road async overhaul" project.
20:21 paulproteus: and i'll reply to your email.
20:21 paulproteus b
20:21 willkg speaking of email....
20:21 bendk: i fully plan to get to your "stop the corruption!" email.  i just haven't had time to work through it, yet.
20:21 bendk: i'll try to do that soon.
20:22 paulproteus ajonas: #3 turns out to be easy -- doing that!
20:22 bendk okay, no problem
20:22 I took a while to get to it
20:22 ajonas paulproteus: perfect
20:23 paulproteus: can we leave it in the DB, but not show it to the admin, as well? as in, besides just not doing what the css says, no longer having the custom css area be visible?
20:23 paulproteus ajonas: Right now, we grey it out, rather than not show it.
20:23 ajonas paulproteus: that's better, great
20:23 paulproteus I slightly prefer that, and it even says (upgrade for it to be available)
20:23 Great (-:
20:27 paroneayea paulproteus: I have a stupid question!  I think I even originally put a lot of these calls in place, and now I can't remember :)
20:27 paulproteus Sure, fire away!
20:27 paroneayea get_static_url, that's not a django-specified tag, where does it come from?
20:27 grep doesn't return a definition
20:28 paulproteus It's whatever uploads things to S3 for us.
20:28 z3p paroneayea: uploadtemplate
20:28 paulproteus Oh, huh, okay.
20:30 paroneayea z3p: oohhhh now I see what happened
20:30 z3p: so iirc that was a custom tag originally in localtv when I left the project
20:30 and now it's in its own module
20:30 django-uploadtemplate
20:30 z3p paroneayea: yeah
20:30 paroneayea which is why I was confused while grepping for it :D
20:30 paulproteus git log -p -Sget_static_url
20:31 Hmm, that doesn't do much good.
20:31 (Whatevs.)
20:32 paroneayea that's another one for the doc
20:32 to copy over from example_project
20:32 MEDIA_URL
20:32 I was trying to figure out why CSS /JS only worked on /
20:32 :)
20:32 I'll add these, I'm taking notes for now
20:33 willkg bendk: so it looks like the torrent download works for about 15 seconds and the rate increases over that time.  then suddenly it drops to 0 and remains there.
20:33 bendk: status says there are 62 seeders and 4 leechers.
20:33 bendk weird
20:34 willkg yeah.  no kidding.
20:34 paulproteus ajonas: CSS enforcement fixed!
20:34 bendk so one error is that 0bps downloads shouldn't display "starting up"
20:34 but that's not the root of the problem
20:34 * willkg nods.
20:34 bendk okay, but let's get on the call
20:34 ajonas paulproteus: awesome
20:34 paulproteus I'm headed out for a bit. paroneayea, if you need me, I'll be here sometime! (-:
20:34 paroneayea paulproteus: cool.  I think I have everything set up now :)
20:35 later paulproteus !
20:35 thanks
20:35 for the help
20:35 paulproteus bd!
20:36 ajonas: p.s. so much for my "this week"!
20:36 But we're closer...
20:36 ajonas paulproteus: definitely
20:37 paulproteus: am i right in thinking http://bugzilla.pculture.org/s[…]_bug.cgi?id=14771 is FIXED ? at least provisionally?
20:41 paulproteus Yes!
20:41 ajonas paulproteus: cool, will mark it as such
20:44 Seganok joined #miro-hackers
20:51 mattjohnson left #miro-hackers
20:57 paroneayea paulproteus: http://develop.participatorycu[…]ro_Community_site is updated with my observed additional bits
21:00 mattjohnson joined #miro-hackers
21:17 DGMurdockIII joined #miro-hackers
21:30 willkg bendk: a while back i wrote a script that builds a database on whatever it can recover from an old database.
21:30 not sure what i did with it.
21:32 afk a smidge for coffee...  then back to testing that folder problem.  then the torrent problem.
21:33 Seganok left #miro-hackers
21:43 ajonas left #miro-hackers
21:44 ajonas joined #miro-hackers
21:44 janetPCF left #miro-hackers
21:45 ajonas left #miro-hackers
21:45 ajonas joined #miro-hackers
21:46 willkg bendk: you on?
21:46 bendk I'm here
21:46 willkg so i can detect when we've started downloading a torrent, but the rate is now 0.
21:46 i was thinking of returning an activity message of "download stalled" or something like that.  is that a good idea?
21:46 bendk I think that's reasonable
21:47 willkg ok.  i'll tweak that change now, then.
21:47 bendk or just report 0kbs/s speed
21:47 I think that's how transmission does it
21:47 * willkg nods.
21:47 willkg it's funky.  the message that gets shown is in two places in the code.
21:49 also, it doesn't look like the frontend knows how much we've downloaded so far.
21:54 janetPCF joined #miro-hackers
21:59 willkg mmm... scratch that.
22:01 Seganok joined #miro-hackers
22:01 paroneayea left #miro-hackers
22:02 willkg bendk: so it looks like what happens is that the torrent starts up, it starts downloading data, the total_wanted_done starts going up.  but then it suddenly drops to 0 and the rate drops to 0, too.  then libtorrent starts over with handshaking.
22:02 weirdness.
22:08 bendk gotta run, I have the item insert performance email almost done, but I'll finish it later
22:08 bendk left #miro-hackers
22:08 paroneayea joined #miro-hackers
22:10 ajonas left #miro-hackers
22:12 Seganok left #miro-hackers
22:25 ajonas joined #miro-hackers
22:35 willkg afk a bit.  baby crying.
22:52 paroneayea left #miro-hackers
22:53 paroneayea joined #miro-hackers
23:52 janetPCF left #miro-hackers

← Previous day | Index | Server Index | Channel Index | Today | Next day → | Atom Feed | Search | Google Search | Plain-Text | plain, newest first