All times shown according to UTC.

Time Nick Message
01:20 CarlFK joined #miro-hackers
02:05 DGMurdockIII left #miro-hackers
02:11 CarlFK left #miro-hackers
02:11 CarlFK joined #miro-hackers
06:45 Odysimus left #miro-hackers
07:51 CarlFK left #miro-hackers
08:05 CarlFK joined #miro-hackers
08:33 CarlFK left #miro-hackers
08:54 CarlFK joined #miro-hackers
09:02 Odysimus joined #miro-hackers
09:05 Dawnrell joined #miro-hackers
09:26 TSK joined #miro-hackers
09:57 maggie_s joined #miro-hackers
10:02 maggie_s left #miro-hackers
10:04 maggie_s joined #miro-hackers
10:05 maggie_s left #miro-hackers
10:12 Dawnrell left #miro-hackers
10:45 TSK left #miro-hackers
10:57 glee1 joined #miro-hackers
10:59 maggie_s joined #miro-hackers
11:00 glee left #miro-hackers
11:35 maggie_s left #miro-hackers
12:05 pythonian4000 joined #miro-hackers
12:23 pythonian4000 Yo miro-hackers!
12:30 I have Atom feed capability in SNAILBot now.
12:34 Go to http://irclogs.jackgrigg.com/atom.pl to generate a feed URL.
12:36 One query - at the moment the nick of the person making the comment goes in the "Author" field, but that doesn't seem so noticeable. Should I add, e.g. "willkg - " to the front of the title (which is currently the first 50 chars of the IRC message)?
12:45 willkg joined #miro-hackers
12:46 willkg hi!
12:47 pythonian4000 Hi willkg!
12:47 Check the logs ^_^
12:47 And to answer my own question, I added the nicks first and it looks better.
12:50 willkg whoops.  i saw pythonian4000 and thought "go look at logs for openhatch", but then i didn't see anything interesting there.
12:51 give me a second now that i'm more awake.
12:52 pythonian4000 Hehe.
12:52 willkg pythonian4000: nice!
12:52 pythonian4000 Figured I would post here since it was a Miro initiative that spurred me on.
12:52 willkg pythonian4000: give me a minute to toss this in the planet getmiro planet thing and see how it works out.
12:53 pythonian4000 If you don't specify a feed length it defaults to the last 200 messages that match your nick or test filters.
12:53 (I should probably put that somewhere on the generator page)
12:56 * willkg nods.
12:56 pythonian4000 Hmm, on thought, maybe the generated feed's link should point to the channel page...
12:56 willkg i did 30.  30 is plenty.
12:56 pythonian4000 paulproteus recommended 200 by default
12:57 willkg looks good.
12:57 so i have a question....  our planet script kicks off every hour.
12:57 if it does a pull on an atom feed every hour, is that going to make your server crazy?
12:59 pythonian4000 Well there is no caching as yet, but the generation is pretty simple - just select the last N entries from MySQL and create a feed. So shouldn't be too heavy =)
12:59 * willkg nods.
13:00 willkg i'll ping the other miro people and see how they like it.  another possibility we could use is to write another script that grabs a day's worth of deardiary stuff and generates a single entry.
13:00 then that would go in the planet.
13:00 that script would run once a day.
13:00 pythonian4000: thank you so much for this!
13:00 pythonian4000 I can add caching, but then I would need overhead to check if the cache needs updating, so it's a case of whether that overhead is less than the  actual generation of the feed...
13:00 No problem! Was fun to break out the Perl again.
13:01 Could you re-generate and re-add the link again?
13:01 (I've changed the feed link to point at the channel page rather than the index)
13:02 (and I want to check that it carries through properly)
13:02 willkg pythonian4000: looks good to me.
13:03 pythonian4000 Same here. Sweet!
13:03 willkg left #miro-hackers
13:04 willkg joined #miro-hackers
13:04 willkg rrr....  sorry about that.
13:04 i'm trying to switch from xchat to quassel and having serious difficulties with it.
13:05 pythonian4000 Np.
13:15 arpu joined #miro-hackers
13:16 Dawnrell joined #miro-hackers
13:52 willkg left #miro-hackers
13:52 willkg joined #miro-hackers
14:38 pythonian4000 is now known as pythonian4000afk
15:10 CarlFK left #miro-hackers
15:24 CarlFK joined #miro-hackers
15:28 CarlFK1 joined #miro-hackers
15:28 CarlFK left #miro-hackers
15:43 CarlFK1 left #miro-hackers
16:28 willkg glee1: i've got the unit tests running in GMT.  but i can't get the testgenerator to run in GMT and produce the expected output.  it's so weird.
16:30 glee1: bah.  got it working.  SO WEIRD!
16:31 glee1 willkg: All's good?  Probably just a transient error.
16:34 willkg glee1: nope.  turns out if i import unittest, my test generator does the right thing.
16:34 glee1: if i don't import unittest, then feedparser in the test generator uses EST instead of GMT.
16:35 so ridiculously klugy.
16:37 glee1 Bizzaro.  i thought the TZ environment variable applies program-wide, wierd that importing would make a difference.  I wonder what the value of TZ is in the test generator if you didn't import unittest
16:39 willkg it looks like TZ was working program-wide, except for some reason feedparser was still generating local datetimes.
16:39 when i import unittest, then it suddenly worked fine.
16:40 glee1 Hrmph.  I think something in there is a bit dodgy, but I'm not sure what it may be.
16:40 willkg there's totally something dodgy.  i skimmed through the feedparser code, but didn't see anything that jumped out at me.
16:41 glee1: want to do a git fetch and see if the unittests pass for you?
16:41 glee1: i set TZ in the os.environ, so you shouldn't have to do anything different to run the tests.
16:41 glee1 Sure, one second.
16:42 willkg no rush.
16:42 i just wanted to make sure the issue is fixed before i mark the bug as fixed.
16:45 glee1 willkg: No joy… I still get those errors.  More tests are actually failing but I think it's the same issue here.
16:45 I get stuff like:
16:45 AssertionError: Strings differ at index 3113:
16:45 , tm_hour=22, tm_min=20, tm_se
16:45 , tm_hour=8, tm_min=20, tm_sec
16:46 willkg did you delete dist/build first?
16:46 and how'd you run the tests?
16:48 * paulproteus waves.
16:48 glee1 willkg: So, I ran ./clean.sh, which should have nuked everything.  As for running: ./run.sh —unittest
16:49 willkg: I checked the changeset for good measure and from first glance my method of running the unittests should pick up the TZ thingy.
16:50 willkg mmm...  you're seeing a different delta between the hours of the two outputs.
16:50 glee1 Yeah, intriguing.  Somehow the timezone isn't doing what it is supposed to.
16:51 willkg so it looks like it's having some effect.  but it's still wrong.
16:51 i ran the tests on my laptop and my macbook.  so the changes i made continue to work in my timezone.
16:51 how do you change the timezone on a macbook?
16:52 paulproteus: hi!
16:52 paulproteus: pythonian4000afk implemented deardiary search atom feeds.
16:52 paulproteus Yup, I saw! (I suggested it to him on #openhatch I admit)
16:53 If they're truly going to be on Planet, maybe having them be batched into one per day, or one per two days, would be nice.
16:53 glee1 willkg: System Preferences -> Click on Date/Time under System
16:53 willkg paulproteus: also, i decided i don't like quassel at all.  the ui is funky.  i can't tell if that's because of how ubuntu has things set up or not.  but total ick.
16:53 paulproteus willkg: Yeah, the UI isn't great. The reason I like it is for the quassel *core* + quassel frontend.
16:53 glee1 Or you can access it by clicking on the time on the menu bar on the top right hand corner of the Mac.
16:53 willkg paulproteus: i was thinking of batching them, too.  amongst other things, i can fix the formatting.
16:53 paulproteus Cool.
16:54 willkg glee1: i'll try fiddling with the timezone.
16:55 glee1: if i can't get it to work without much more effort, then i'll create a new tag and limit running those tests to EST/EDT.
16:56 mmm...  either that or tweak pprint so that datetimes get represented differently as strings.
16:57 glee1 willkg: Sounds good to me.  I'd prefer to keep the bug open or perhaps open a new one so someone who's keen can take a look at it.  I mean, I don't think anything's broken and we run tests daily in the what happens to be the correct timezone
16:58 willkg yes, i agree that these tests should get fixed.  right now i'm mostly concerned with getting them not to fail or not to run for now.
16:58 having said that, i have exactly this problem with pyblosxom and i've never been able to figure out how to fix it.
16:59 paulproteus Store all dates in UTC! Vive la revolution!
16:59 willkg paulproteus: it's not us--it's fucking feedparser.
16:59 * paulproteus doesn't know quite what you're talking about, though.
17:00 willkg i wrote a bunch of tests to verify that the output feedparser gave us before is the same output that feedparser gives us now.
17:00 paulproteus Oh, I see.
17:00 willkg the purpose of that was to allow us to run regression tests to know whether the data is the same if we were to switch feedparser versions and when we make changes to feedparser.
17:00 paulproteus Yeah.
17:02 willkg mmm...  the more i think about this, the more i like my idea of removing datetime objects from the FeedParserDict.
17:03 glee1 I see one or two timezone related things in the feedparser, though I'm not quite sure if it has anything to do with it.  There's stuff like, rf822.mktime_tz().  There's a whole bunch of stuff in feedparser and I'm not quite sure what it's trying to do (yet).
17:04 Date/time handling is always …. interesting.
17:05 paulproteus see also Jonathan Coulton song, "Christmas Is Interesting".
17:18 Odysimus left #miro-hackers
17:45 willkg glee1: i just pushed another fix for 15953.  this one changes time.struct_time instances to "DATETIME" in the feedparserdict.  so then we're comparing "DATETIME" to "DATETIME".
17:46 glee1: and i removed all the TZ silliness, too.
17:46 glee1: want to give it a try and see if that fixes the issue for you?
17:47 glee1: it works for me with my macbook in GMT.  so i think it'll work for you, too.
17:47 glee1 willkg: ah, fantastic.  All tests ok.
17:48 willkg glee1: awesome.  thank you for helping!
17:50 glee1 No problem, all good!  I think we can close the bug.
17:51 willkg done.  :)
17:52 afk a bit to do baby things.
17:53 deardiary: fixed a timezone-related issue in the unit tests.  pushed my header/details changes to the extensions prefpanel.  going to try to finish up extension changes today.
18:25 paulproteus willkg: rock re: datetime
18:37 DGMurdockIII joined #miro-hackers
19:08 Odysimus joined #miro-hackers
19:15 pythonian4000afk is now known as pythonian4000
19:20 CarlFK joined #miro-hackers
19:36 paulproteus deardiary: I successfully got pythonian4000 to write the search RSS feeds for deardiary (-:
19:37 pythonian4000 paulproteus, I believe they are Atom feeds, you know, since I wrote them =P
19:38 paulproteus pythonian4000: er, right (-:
19:38 deardiary: accidentally forgot that RSS feeds and Atom feeds aren't the same thing.
19:39 pythonian4000 I'm quite pleased with how it works! No idea what it will do to my server in the long run, but then my only real monitoring of it is that I use the same server for devel ^_^
19:56 Odysimus left #miro-hackers
20:36 Dawnrell left #miro-hackers
20:37 Dawnrell joined #miro-hackers
20:56 DGMurdockIII left #miro-hackers
21:11 CarlFK left #miro-hackers
21:13 kcw left #miro-hackers
21:15 DGMurdockIII joined #miro-hackers
21:24 CarlFK joined #miro-hackers
21:30 CarlFK left #miro-hackers
21:32 CarlFK joined #miro-hackers
21:37 CarlFK left #miro-hackers
22:24 CarlFK joined #miro-hackers
22:25 CarlFK left #miro-hackers
22:25 CarlFK joined #miro-hackers
22:29 CarlFK left #miro-hackers
22:29 CarlFK joined #miro-hackers
22:36 CarlFK left #miro-hackers
22:51 CarlFK joined #miro-hackers

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