All times shown according to UTC.

Time Nick Message
00:02 bendk nope, just regular funkiness
00:02 sending the fix in now
00:07 zanoi so when should _set_currently_playing be used?
00:09 oh, not at all..you took it out completely
00:10 (or rather renamed it)
00:10 * zanoi is tired
00:15 zanoi bendk: thx for fixing
00:24 hmm..the miro download page on getmiro.com still has a banner for the line adoption even though adoption is closed
00:28 deardiay: shuffle/repeat code had to be reorganized/moved one last time and I got sidetracked with other bugs, so now it works less than this morning...
00:31 maggie_s left #miro-hackers
00:40 ajonas joined #miro-hackers
02:39 kcw left #miro-hackers
03:38 odenthaa joined #miro-hackers
03:53 odenthaa left #miro-hackers
04:02 odenthaa joined #miro-hackers
05:14 ajonas left #miro-hackers
05:15 odenthaa left #miro-hackers
07:26 janet_ left #miro-hackers
09:20 maggie_s joined #miro-hackers
09:41 Odysimus left #miro-hackers
09:41 Odysimus joined #miro-hackers
10:03 Odysimus left #miro-hackers
10:03 Odysimus joined #miro-hackers
10:42 maggie_s left #miro-hackers
12:03 willkg joined #miro-hackers
12:18 willkg hi!
12:52 Jarrhead joined #miro-hackers
13:14 z3p good morning
13:15 willkg z3p: morning!
13:19 paulproteus: so the windows build box cdrom drive does work.  i'm puzzled--i had a hell of a time getting it to do things a year ago or so.
13:19 paulproteus: but...  that makes this memtest thing much easier!
13:20 paulproteus: doing it because the windows box hung again last night.
13:22 deardiary: running memtest86 on windows build box which hung again last night.
13:23 z3p are we supposed to be saving the filter state for tabs?
13:23 willkg z3p: i don't think so.
13:23 z3p: at a minimum, i haven't seen a feature request for that, yet.
13:24 z3p willkg: okay; I'm filing a bug then because it's annoying me
13:25 * willkg nods.
13:25 willkg yay for bugs!
13:36 zanoi morning!
13:44 willkg paulproteus: all memory tests passed.
13:49 yay!  it's snowing.  again.
13:49 afk...  baby stuff.
14:09 zanoi deardiary: working on bug 9440 again
14:23 z3p left #miro-hackers
14:24 zanoi bendk: moving shuffle/repeat to PlaybackPlaylist caused more problems than it solved :-S
14:24 bendk: but i'll leave it at that and work around the problems now
14:24 bendk sorry about the grief
14:24 are there still issues?
14:24 willkg deardiary: windows build box is running again.  going to spend the rest of the morning cleaning out my inbox.
14:26 zanoi bendk: nothing big, just small issues
14:47 Odysimus left #miro-hackers
14:48 bendk willkg: the build box is a physical box right, not a VM?
14:48 maybe hard drive issues?  I have no clue, but it seems like it could be hardware related in some way
14:54 willkg bendk: yes, it's a physical box.
14:54 bendk: i think it's related to the cron thing kicking off.  seems to hang at around that time.
14:55 bendk that's what gave me the idea of disk issues, maybe cron is hitting some section of the disk that's failing??  Or maybe it's just windows being flakey.  I just wanted to shoot out that idea
14:55 willkg it's hard to know for sure, though.  and the box is hanging--it's not bluescreening.
14:56 i think if it was hitting a disk that was failing, that'd produce a more frequent problem.  this seems to happen really sporadically.
14:56 z3p joined #miro-hackers
15:05 zanoi yay, all problems fixed
15:08 z3p left #miro-hackers
15:08 z3p1 joined #miro-hackers
15:10 zanoi does anyone know if there is any way for a user to play the previous track without manually skipping there?
15:12 z3p1 zanoi: I'm not sure what that means
15:12 z3p1 is now known as z3p
15:13 zanoi well for the next track you can either progress by skipping to eat or because of your current track ending
15:14 *to the next track
15:14 but for the previous track i assumed you could only get there by skipping there
15:15 z3p zanoi: I think that's reasonable
15:15 zanoi but i think yesterday i saw some code that seemed to assume it was possible to get there without skipping
15:16 but i'm not sure, i didn't really look at it
15:16 z3p zanoi: could just be defensive programming
15:16 zanoi ok, great
15:37 willkg z3p: so...  the external drive thing works, but it's taking miro a couple of minutes to figure everything out.
15:38 z3p willkg: yeah; it scans the drive for media
15:38 willkg z3p: is it possible to do a "figuring things out ..." progress bar somewhere?  right now the video and audio views are empty, but i can see from the logs that it's working hard.
15:38 z3p willkg: hmm; I'd been treating it watched folders, where things just show up eventually
15:39 willkg: but I suppose we could throw up a 'processing' dialog if you hit the view while we're searching
15:39 willkg well, i've got a gazillion things on my external drive.  so "eventually" is a long time and the ui isn't telling me what's going on.
15:40 z3p willkg: which mimics existing behavior, but I agree it's not ideal
15:40 willkg well, no it doesn't.  we didn't handle external drives before.  so this is a new situation.
15:40 z3p willkg: if you added your external drive as a watched folder, I think the same thing happens
15:41 willkg right, but we didn't support that at all.  prior to 4.0 we didn't support network folders or external drives.
15:42 afk ...  baby.
15:45 ajonas joined #miro-hackers
15:46 willkg z3p: plus i'm getting a lot of errors when trying to do things while miro is searching and now metadataizing all that content.
15:46 z3p willkg: what's the error(s)?
15:47 willkg z3p: i've got 7 now.  do you want all the crash reports?
15:47 there are a variety.
15:47 z3p willkg: sure
15:47 willkg alternatively, i can wait until miro stops working on things and then send them to you.  it's possible i'll run into more.
15:51 z3p willkg: either way is fine
15:51 willkg i'm up to 15 crashes now.  i haven't check to see if there are duplicates--there probably are.
15:52 i'll wait until it's done importing or whatever it's doing.
15:58 Jarrhead left #miro-hackers
16:10 willkg i think i've overloaded miro--it can't handle my external drive.
16:14 also, we have to do something about these thumbnails.  most of the ones i'm seeing are 300K.  that's kind of silly and the thumbnails alone are taking up 800mb on my drive now.
16:14 that may be a linux-only issue.  i'll look into it.
16:23 oh wow!  so thumbnails are taking a ton of space on my external drive plus they're getting added to my regular drive, too.
16:24 18 crash reports since miro noticed my external drive.
16:38 paulproteus: so doing a dput with the other ppa's libtorrent .changes file doesn't work because i have to sign it anyhow.
16:39 so...  i think i'm going to go back a step, grab libtorrent 0.15.4 from maveric, do a proper update on the package, and push that.
16:41 but...  i'll wait to do that until miro is done thrashing on my system.
16:41 afk for a while to go eat.
17:28 bendk one idea about the external drive (and watched folders to): we should progressively add new items if possible.  Maybe send a blank ItemList message, then a bunch of ItemsChanged messages as we find things
17:30 z3p bendk: that would certainly be more useful for will
17:31 willkg i'm not sure.  it's still churning on metadata requests and update_finished() things and the Video and Audio tabs aren't working.
17:31 it's been 2 hours now.
17:31 z3p willkg: how big is your external drive?
17:31 willkg z3p: 250gb.
17:31 z3p: but it's not all media.
17:32 z3p willkg: yeah
17:32 willkg: I bet a lot of it is running moviedata on stuff
17:32 willkg rhythmbox says i have 2333 songs on there.
17:32 i think that there are other problems.  i'm seeing multiple thumbnails for the same video.
17:33 so i think that the metadata/moviedata things are running multiple times for the same item.
17:33 i'm seeing 11 thumbnails for one .mov file i have.  all generated in the last 3 hours.
17:33 z3p does moviedata generate thumbnails for image files?
17:34 I wonder if it's in a loop, where it's scanning the .miro directory and generating thumbnails for the things in there
17:34 willkg i'm not seeing evidence of thumbnails for image files.
17:34 z3p willkg: hmm, okay
17:35 willkg 800mb of thumbnails on my external drive and 1.1gb of thumbnails on my hard drive now.
17:35 most of that is from this run.
17:35 i don't know why it's putting images in both places.
17:36 on my external drive they're in .miro/ .  on my hard drive, they're in ~/.miro/icon-cache/extracted/ .
17:36 z3p: does miro create the thumbnail, store it in the iconcache directory and then additionally copy it to the external drive?
17:37 z3p willkg: looks like yes; it's set to copy so we don't destroy thumbnails for existing items
17:37 * willkg nods.
17:37 z3p willkg: but I should be checking for extracted thumbnails and moving them instead
17:37 willkg ok.  so that explains that behavior.
17:41 z3p willkg: just checked in a fix for that
17:42 willkg i wish i could get a handle on how much more work miro plans on doing.
17:44 z3p does mutagen detect if things are 'unknown'? or are we running moviedata on everything if mutagen can't figure it out?
17:45 willkg i don't know.  kaz would know.
17:53 z3p looks like no
17:53 so we're running moviedata on every one of your files
17:53 I can imagine that takes a while
17:55 willkg it's better--i'm pretty sure we're running moviedata/metadata on all of my files _multiple times_.
17:55 for some loose definition of "better".
17:57 z3p willkg: do you have symlinks?
17:57 willkg z3p: i don't know what you mean by that.
17:58 z3p willkg: on your external drive; are there symlinks, which would create multiple paths to the same file?
18:00 zanoi afk for dinner, will be back later
18:00 kcw joined #miro-hackers
18:00 willkg so this is not only my media drive, it's also my backup drive.  so some of the files will have at most 2 symlinks.  but i'm seeing 10+ thumbnails for those files.
18:00 maggie_s joined #miro-hackers
18:00 willkg z3p: http://dpaste.com/375229/
18:01 the three bottom ones in the backup directories--one of those is real and the other two are symlinks.  all of the thumbnails are real and new as of the last 3 hours.
18:05 z3p hmm
18:05 willkg is it possible that clicking on the item view can cause a bunch of requests for metadata to go out for things that already have requests pending?
18:07 z3p willkg: I think that's what's happening; at least, that's what I'm looking into atm
18:07 * willkg nods.
18:07 willkg if that's the case, then having a big progress bar along the lines of "hold yer horses!  i'm importing!" would help though it doesn't address the actual problem.
18:18 z3p willkg: I just pushed up a fix
18:18 willkg z3p: ok.  mmm....
18:18 z3p willkg: although it won't deal with all the existing thumbnails
18:18 willkg so i think i'll kill miro, clear things to an original state, and then see if i can reproduce the issue again.
18:18 z3p willkg: sounds good
18:19 unfortunately, I don't have a big drive with me here at the café but I can see how the code I changed would have been a problem
18:19 willkg it's ok.
18:19 also, i can send you all these crash reports.
18:20 zanoi back
18:22 willkg z3p: sent.
18:22 z3p willkg: thanks!
18:23 paulproteus willkg: Just debsign it yourself, then.
18:23 Morning, all.
18:23 willkg paulproteus: pretty sure my signature has to match the last entry in the changelog.  at that point, i'm doing my own package.
18:23 paulproteus willkg: You should let it run ca. 7 times, fwiw re: memtest
18:23 It doesn't have to, willkg; but okay.
18:23 willkg paulproteus: why?
18:23 paulproteus: it does for the PPA.
18:24 paulproteus: the changelog entry, my signature, and the ppa account email all have to match.
18:24 paulproteus Huh, PPAs are weird, then. (But you can just bump debian/changelog... but sure. whatevs.)
18:25 re: memtest: Memory failures can be subtle, and some bad RAM I just RMA's last week only saw failures after the third run of the test suite. (This is why memtest86 goes on a loop.)
18:25 So I guess my advice is, leaving it doing that overnight.
18:28 willkg well, next time it hangs, i'll try that.
18:45 janet_ joined #miro-hackers
18:53 DGMurdockIII joined #miro-hackers
19:06 z3p deardiary: threads are making my brain hurt
19:06 willkg z3p: it seems to be behaving much better now.
19:06 z3p willkg: glad to hear it!
19:07 willkg i need to look into why thumbnail files are so huge, though.
19:07 z3p willkg: I checked in a bunch of other changes, which should fix everything except the JSON dumping error
19:07 willkg z3p: awesome!
19:07 z3p: did the crash report files help?  are there things we can add to those files to make them more useful?
19:07 z3p willkg: and the JSON dumping error is going to be a huge pain in the ass, but that's alright
19:07 willkg i have to say, zipping up a bunch of files and sending them to you was EASY.
19:08 z3p willkg: to be honest, they were mostly errors I'd seen once or twice, but not often enough to be sure what was happening
19:08 * willkg nods.
19:09 willkg i'm asking less about the specific reports and more about the file structure, information therein, ...
19:09 z3p willkg: what I mean is that I just skimmed the traceback, and ignored all the other stuff
19:09 * willkg nods.
19:10 willkg ok.
19:10 so one thing i noticed is that channel guides don't regenerate their icons if you wipe the cache.  i'm not sure that's a big deal, but it's weird looking.
19:11 stores did.  and podcasts did.  channel guides are the only thing that look weird.
19:12 takes a half-second for each update_finished() to execute.
19:12 z3p: when an update_finished() executes, does the json file get rebuilt each time?  or does it batch updates?
19:12 z3p willkg: some things batch update; I don't believe that's one of them
19:13 * willkg nods.
19:13 willkg we might want to change that.  it'd make the initial load faster.  maybe change it just for the initial load?
19:13 z3p willkg: update_finished() isn't called during the initial load
19:14 willkg: it's only called when we interact w/ moviedata
19:14 willkg i'm using "initial load" differently.  miro goes out and finds all the files that are on the drive, then queries metadata/moviedata on each one of them.
19:14 z3p willkg: ah, I see
19:14 willkg given that the files themselves aren't very useful without the metadata, i consider that whole operation to be the "initial load".
19:14 z3p willkg: those steps happen at different times
19:14 willkg oh, i know.
19:15 but because each update_finished() causes the json file to be re-generated and stored to disk and that happens 2333 times with 2333 times, that's a lot of extra work until the data is useful.
19:16 it looks like miro is done fetching all the metadata--it's just calling update_finished() which takes between .5 and .8 seconds each call.
19:16 z3p willkg: hmm
19:17 willkg bendk: how do we handle this sort of thing with the database?  do we write to the database for each individual changes or do we queue changes up into batches?
19:17 bendk we queue things
19:18 z3p willkg: yeah, I think I can queue the changes a bit
19:18 willkg even queuing 3 at a time would help a lot.
19:19 z3p willkg: I can use the eventloop to keep things together a bit
19:19 bendk with the database I think we schedule writes every X seconds (I forgot what X is), then also write out the data before we close
19:20 willkg one of the things about external drives is that, especially the case with thumbdrives, they have limited writes.
19:22 arpu joined #miro-hackers
19:23 willkg z3p: as a side note, if i'm getting boring, let me know.  i can take this to bug reports or aggregate thoughts into an email or push this off until another day.
19:24 z3p: it's way better than miro 3.5.1--that's for sure.
19:24 z3p willkg: I don't think that's much of an issue anymore, and will continue to be less of an issue
19:24 * willkg nods.
19:24 z3p willkg: but having it take forever is an issue
19:24 willkg: SSDs as main hard disks are becoming more common, and they're going to have a ton more writes
19:25 willkg i plugged my drive in at 13:44.  so it's been doing the "initial load" since then.  it's been doing update_finished() things since 13:58.
19:25 so it took 14 minutes to pull all the metadata and 22 minutes (so far--it's still going) just to save and update miro with it.
19:27 and kcw's metadata changes make pulling metadata _way_ faster than it used to be.  that's really nice.
19:28 z3p willkg: okay, so clearly something needs to happen with saving
19:29 paulproteus willkg: The "limited writes" thing is mostly a myth, I understand.
19:29 All consumer grade flash drives have done wear leveling since the dawn of time.
19:31 See also http://superuser.com/questions[…]ar-a-real-problem
19:31 willkg paulproteus: i've never heard that it was a myth and there was a real uproar when ubuntu switched filesystems (or something like that) and there were additional writes being generated.
19:32 paulproteus I did the math a few years ago, and satisfied myself.
19:32 morganknutson left #miro-hackers
19:32 paulproteus It's so not-a-big-deal that Microsoft ships a feature of Windows to do swap on flash.
19:32 http://windows.microsoft.com/e[…]atures/readyboost fwiw
19:33 morganknutson joined #miro-hackers
19:33 willkg ok.  so you said "since the dawn of time", but i'm not seeing evidence of that in this reference at all.
19:33 seems to indicate recent ones are fine, but it doesn't say all flash drives ever are fine.
19:35 also, there's a significant lack of references or tests in there.  seems like it's either "common knowledge" or anecdotal data.
19:36 paulproteus From 2003: "Maybe wear leveling is standard today and there's no need to advertise it anymore" http://www.pcreview.co.uk/foru[…]ves-t1941484.html
19:36 The real problem is that people who don't know any better assume the properties of NOR/NAND flash will propagate up to the block device level.
19:37 This is due to a lack of knowledge of the firmware on the block-device-level things, like USB flash keys.
19:37 ajonas paulproteus: note that http://bugzilla.pculture.org/s[…]_bug.cgi?id=15428 has two emails, unless the top one is covered elsewhere we should make sure they're both doen before closing it
19:38 paulproteus willkg: Even in 2003, the firmware on cheap USB flash keys would actually emulate a FAT filesystem and do block moving magic behind your back.
19:38 This actually meant it was a bad idea to mkfs.vfat on those.
19:38 This argument has been going on since 2001 or so, and every day new people come to it not knowing anything about the firmware on these devices and aware of the rewrite scare factor.
19:39 willkg i'm skeptical.  i think the only way i would become less skeptical is if someone showed me a real myth-buster article that walked through the myth and the misconceptions.
19:39 paulproteus Sure, fine by m.e
19:39 I'll work on that one day, but I guess not today.
19:39 willkg does such a piece exist?  if not, doesn't that lend credence to the fact that there's truth to the myth?
19:40 z3p: it's done at 14:28.
19:40 z3p willkg: holy cow, okay
19:40 willkg z3p: just filling out the data--not meaning to nag you or anything.
19:40 paulproteus willkg: Honestly, with all due respect, I think the situation is that people who understand this just don't care to get into the debate with people who are terrified their NAND/NOR will fail on them at any moment omg.
19:41 Such old things as https://secure.wikimedia.org/w[…]wiki/CompactFlash talk about wear leveling.
19:41 I'll see if I can find a good article, but that's what I've gleaned.
19:41 z3p willkg: I guess there's a lot more to signal_change() than writing the DB, but still that seems crazy
19:41 paulproteus willkg: Also no one you've ever heard of has had a part of their flash drive get bad sectors, if you're like m.
19:41 z3p willkg: but I suppose it's also moving 800MB of thumbnails
19:41 paulproteus ajonas: Yeah, that's right; only the second email is done iirc
19:42 ajonas paulproteus: *nod* also, since it's relevant for so many of the links in the emails, note http://bugzilla.pculture.org/s[…]_bug.cgi?id=16008
19:42 i'm trying to purge the word "tiers" from any user facing things i can
19:42 z3p I've got a possible fix, but I'll try it out later when I can test it
19:43 willkg paulproteus: i've gone through two thumbdrives.  it's not like it tells me what died on the drive, so it's not clear to me what the problems were.
19:43 paulproteus Yeah, but they usually catastrophically fail rather than grow bad sectors.
19:43 ajonas: yeah, I know you are, but I wasn't sure what word to use; so I thought I'd leave it to you.
19:43 ajonas paulproteus: totally, i'm trying out "Account" in most places
19:43 z3p willkg: the failure I always see is that my computer no longer recognizes the drive being plugged in
19:44 willkg: not random filesystem errors
19:44 * willkg shrugs.
19:45 paulproteus Skepticism is basically always the right approach, willkg. You shouldn't be swayed by me, either, just because I am a charming, good-looking person who knows how to sound authoritative on IRC.
19:45 (-:
19:45 z3p anyways, the important thing is that I'm packing my computer up to walk home
19:46 so I will talk to you all about flash drives another time :)
19:46 willkg z3p: good luck with that.  hope you brought your snow shoes.
19:46 z3p willkg: I'm wearing my boots, and the snow has calmed down a bit
19:46 but thanks!
19:50 willkg oh, wow.  miro picked up all the records i had converted to mp3.
19:51 z3p left #miro-hackers
19:57 z3p joined #miro-hackers
19:57 paulproteus back in a bit
19:58 willkg back to torrenty stuff.  but first, afk for coffee.
20:10 zanoi ok, i have shuffle repeat ready
20:11 (except for some corner cases and per display)
20:11 willkg zanoi: nice!
20:11 zanoi should i push it or do you prefer me to wait until every last thing is done?
20:12 and do you want unittests for it? i didnt see any for the frontend stuff so i didnt write any...
20:14 oh, noes..the rebase messed it up..
20:14 ...ok i'll have it done in 2 minutes :-)
20:15 willkg zanoi: i don't know if we have infrastructure for doing unittests in the frontend, yet.  i know it's something we've talked about at various points.
20:15 zanoi: given that you're newish, can you run git format-patch and send the files to bendk to peer review?
20:16 bendk zanoi: there's a couple classes that have frontend tests, maybe PlaybackPlaylist could possibly be one of them, since it dosen't really touch the UI
20:16 willkg bendk: is that ok?  can you look at his patches?  i figure you're the best candidate since you've been fiddling with playback code.
20:16 bendk I can review as well, no problem
20:17 willkg bendk: cool--thank you!
20:17 zanoi yay, thx. Would be nice to have it reviewed
20:18 willkg some day, i'd like to get reviewboard up and running or some similar review infrastructure for "tentative checkins".  that'd help a lot.
20:18 but, not it's not something i'm doing for this dev cycle.
20:19 zanoi yes, that would probably be good
20:19 willkg as a side note, i'm going to shut down the windows build box now and run memtest86 for a few hours.
20:21 ha!  it hung on shutdown.
20:21 how funny is that....
20:22 also, it's different than all the other times so that's interesting.
20:31 zanoi bendk: sent
20:35 bendk zanoi: cool, looking at it now
20:38 paulproteus ajonas: I'm looking at http://bugzilla.pculture.org/r[…]munity&target=1.2 and making a plan for today
20:40 There are some bugs in tiers enforcement; those I want to do today: http://bugzilla.pculture.org/s[…]_bug.cgi?id=15809 http://bugzilla.pculture.org/s[…]_bug.cgi?id=15808 for sure.
20:40 bendk zanoi: can you check that patch again?  I'm seeing lots of stuff from kaz at the begining of the diff
20:40 zanoi bendk: : yes i was wondering about that
20:40 paulproteus Finishing the notifications would be nice, too: http://bugzilla.pculture.org/s[…]_bug.cgi?id=15502 http://bugzilla.pculture.org/s[…]_bug.cgi?id=15428
20:41 http://bugzilla.pculture.org/s[…]_bug.cgi?id=14796 seems to be just a tracking bug for now; I'll read it carefully to figure out the next steps
20:41 zanoi bendk: it looked very wrong but in the end i thought that it's probably because i don't understand git well enough...
20:42 paulproteus I think that with http://bugzilla.pculture.org/s[…]_bug.cgi?id=14796 we should just close it; it tracked some stuff that is done, and then the remainder got split out. What do you think ajonas?
20:42 ajonas i don't think we should close it
20:42 but i'l llook
20:42 zanoi bendk: i'll continue my fight with git then and try to fix it
20:42 ajonas because i think it might be a good checklist
20:42 paulproteus Fair enough re: checklist
20:43 ajonas though there are things not included in there
20:43 but let's leave it for now
20:43 paulproteus Okay, so I'll just ignore it for now and focus on the ones I like more; and then toward the end we can revisit it and see if it reminds us of things we wanted to get done.
20:43 ajonas sounds good
20:43 paulproteus There's a simple (30m-1h) solution to http://bugzilla.pculture.org/s[…]_bug.cgi?id=15356 that would be good. It would be better to do a 2-4h thing of using an add-on package, but I can save that for the performance release.
20:44 ajonas is that necessary for this release?
20:44 bendk zanoi: is it one changeset, or a group of them?
20:44 ajonas i was wondering if we could bump it
20:45 paulproteus If we don't do it, it might cause some JS and CSS not to load for people, forcing them to do a shift-refresh if that occurs to them.
20:45 More likely, they'll just think, "Weird, it's broken".
20:45 ajonas *sad face* ok
20:46 let's leave it, then, but let's do it at the end
20:46 after the other tiers stuff is set
20:46 paulproteus b
20:46 zanoi bendk: well it should be two commits if that's what you mean
20:46 * paulproteus notes so on the ticket.
20:46 paulproteus That's a shame; bugs like that are fun, ajonas (-;
20:47 bendk I would do git --rebase origin/master first, then git diff -p -n2
20:47 ajonas paulproteus: a fun reward at the end of the tunnel, then?
20:47 paulproteus ajonas: Sure, I can do that (-:
20:47 bendk I know there's a way to format it even nicer for an email, but I'm not sure what that is
20:47 willkg bendk: much easier to use git-format patch.  then you get the author and commit message.
20:47 paulproteus The notifications are nice and small and provide the good feedback feeling for me, so that's good so far.
20:48 willkg bendk, zanoi: http://develop.participatorycu[…]ubmitting_a_patch
20:48 * paulproteus tries to understand http://bugzilla.pculture.org/s[…]_bug.cgi?id=15809
20:48 bendk yup, willkg has the good way as usual
20:48 zanoi bendk, willkg: i did rebase origin/master and then format-patch
20:49 paulproteus ajonas: I think http://bugzilla.pculture.org/s[…]_bug.cgi?id=12610 is specific to the blue theme. That seems high priority.
20:49 bendk that should result in a bunch of .patch files in the working directory
20:49 paulproteus + http://bugzilla.pculture.org/s[…]_bug.cgi?id=15446
20:49 ajonas paulproteus: what about the one maggie mentioned, with the saving problem?
20:49 i think that should maybe be highest priority
20:50 sorry, firefox takes awhile to load the links
20:50 paulproteus: possibly i can do 15446 with my new patching skillz?
20:50 paulproteus That would be so super awesome (-:
20:50 http://bugzilla.pculture.org/s[…]_bug.cgi?id=15884 I was thinking might be similar.
20:51 zanoi bendk: well to be exact i did git-format-patch master stdout > file.patch, so had it all in one file
20:51 bendk: why it adds other commits to that i don't know
20:51 ajonas paulproteus: k so let's leave those and i'll try it
20:51 zanoi bendk: but i'll try wills instructions
20:51 bendk -n2 maybe?
20:53 paulproteus ajonas: Which one is the maggie one? Sorry to be losing track.
20:53 ajonas: Maybe I can play with the "Assignee" so that it's clearer to me what to work on next.
20:53 Like, the stuff that I'm working on in the next 1-2 days I assign to myself, and the rest I leave unassigned.
20:53 ajonas paulproteus: no problem, let me look... the one about changes not being saved, i know you did some work on that
20:53 paulproteus How does that seem to you?
20:53 ajonas paulproteus: fine
20:53 paulproteus I'll run and do that right now
20:54 ajonas: Oh, I guess I can use the ASSIGNED status for that. I'll do that instead.
20:54 ajonas: http://bugzilla.pculture.org/s[…]_bug.cgi?id=12324 probably
20:54 ajonas yep
20:55 paulproteus I'm going to aim for Thuish on that.
20:55 ajonas is the part that's fixed pushed to the dev server?
20:55 maggie said it's causing lots of tests to fail
20:57 paulproteus Yeah -- unfortunately, the part that I fixed isn't the important part.
20:58 (The part I fixed was a bug in *our* code, but there's some other bug in how we use a third party module.)
20:58 zanoi willkg, bendk: the codingStandards instructions give me a gazillion patches, even though i did a rebase
20:59 willkg zanoi: what branch are you working in?
20:59 zanoi: and what was the specific git format-branch line you gave?
20:59 bendk also, what's the output of "git status"?
21:01 paulproteus http://blogs.gnome.org/jessevd[…]pository-browser/ has "drag and drop format-patch export", fwiw (willkg, zanoi, bendk)
21:01 zanoi: in case you're selecting the wrong set of revisions to export as patches, you can visually select them using gitg
21:04 zanoi willkg: err well i cloned from master and gave the line: git format-patch -o patches origin/master add_shuffle_repeat
21:04 paulproteus: thx! will install
21:04 bendk: git status doesn't say anything really
21:05 bendk: just lists the files that i haven't added to git
21:05 bendk mine usually says my branch is ahead of master by X commits
21:05 willkg zanoi: and that's after doing a "git fetch" and a "git rebase origin/master" ?
21:05 paulproteus ajonas: That goes for you, too -- if GitX has visual selection of the local commits to export with git format-patch, that would likely be helpful.
21:05 bendk or that both differ by X and Y commits
21:05 ajonas paulproteus: that flew over my head, but i'll keep it in mind
21:06 paulproteus ajonas: Okay, I'll also hold onto the tip so I can say it again when it makes sense.
21:06 zanoi bendk: hmm thats odd
21:07 bendk could just be different git versions
21:07 so 1) I still think you should figure out how to generate the patch, since it's definitely useful
21:07 2) I just skipped to the bottom of that file and saw the patches, so it's not totally necessary right now
21:07 overall looks great, I'll send you feedback as a reply email
21:08 zanoi bendk: well i did send you a working patch a while ago already, dunno why this one is so borked :-(
21:09 paulproteus ajonas: Okay, my plan is to go through http://bugzilla.pculture.org/r[…]munity&target=1.2 and do all the ASSIGNED ones, and then revisit the list.
21:09 ajonas paulproteus: good stuff
21:12 willkg mmm...  i'm going to check in a new column for the downloading tab list view that shows all the torrent numbers.  it's not pretty, but it fixes the problems i've had since we nixed all that information from the interface a while back.
21:15 JasonWoof left #miro-hackers
21:18 zanoi it looks all fine in gitg
21:18 mattjohnson left #miro-hackers
21:18 paulproteus ajonas: A question about free trials, as per http://bugzilla.pculture.org/s[…]_bug.cgi?id=15428 -- we only send this text when you are moving to a premium level and have a free trial available, right?
21:19 If, say, you've already been through one trial, then go down to basic, then back up to a paid tier, we'd need different text? (This is w/r/t the first of the two emails on that ticket.)
21:19 ajonas ugh
21:20 paulproteus: can we just remove the graph about the free trial when it doesn't apply
21:20 ?
21:20 paulproteus Yeah, I can do that.
21:21 ajonas paulproteus: great, ty
21:21 paulproteus Random good news: I think the laptop Lenovo is sending me arrived "today" (except I slept through it), so they'll redeliver tomorrow.
21:21 mattjohnson joined #miro-hackers
21:22 zanoi bendk: ok, weirdly enough i have it working now
21:22 i'll send it
21:24 willkg deardiary: working on torrent-related bugs.
21:25 paulproteus deardiary: Slightly refactoring tiers email sending, and adding more notifications.
21:25 bendk zanoi: heh, I already responded to the 1st email
21:26 zanoi bendk: oh, that's odd
21:26 bendk: haven't received anything
21:27 bendk did you check the spam folder?
21:27 At this point, I'd start looking out for gremlins
21:27 JasonWoof joined #miro-hackers
21:27 JasonWoof left #miro-hackers
21:27 JasonWoof joined #miro-hackers
21:28 zanoi bendk: well it should have gone to my personal mail server and i dont have a spamfilter there
21:28 bendk: i do have a greylist though, but that shouldn't kick in if you reply to my mail
21:28 bendk okay, just replied to the 2nd email
21:30 zanoi bendk: ok, i got that
21:33 paulproteus ajonas: :D Thanks for the patch!
21:33 For me, git complains about "whitespace errors" -- that's because some of your lines end in the " " (space) character, and that can make merging slightly harder down the road.
21:34 git is built to complain about those, even though I think it won't be a big deal for us.
21:34 That's my only piece of feedback except that I'm going to push your patch!
21:34 ajonas: Oh, one more thing actually.
21:36 Since it's a plain text email, it will wrap just the way that you type it out there. The p.s. has a really long line. It'd probably be good to correct that.
21:36 ajonas paulproteus: ok
21:36 i don't quite understand the whitespace problem
21:37 oh, i do
21:37 never mind
21:37 paulproteus Cool.
21:37 ajonas better to start with a space than end with one?
21:37 paulproteus It's always okay to start with a space, but it's frowned-upon to end with one.
21:37 ajonas got it
21:38 paulproteus ajonas: Other than that, my feedback is, "Yaaaaaaaaaaay! Patches!!!"
21:38 Now's a good time to "git pull --rebase" to refresh your repository's view of the world.
21:38 ajonas paulproteus: i didn't know how to change the subject
21:39 paulproteus: so i left a note on the ticket for that
21:39 paulproteus ajonas: yeah, the subject line is obscured elsewhere in Python code. It could be a template file, but for now it's not. It's in localtv/localtv/management/c​ommands/nightly_something.py
21:40 ajonas ok
21:40 if you can change this one i'll try to make changes to those in the future as well
21:40 paulproteus b
21:41 bendk zanoi: your plan sounds great, feel free to fix the bug however you want
21:42 zanoi bendk:  will do after i have finished the new epic fight with git i just started
21:42 ajonas paulproteus: so should i go ahead and edit the p.s. part now and make a new patch?
21:42 paulproteus ajonas: yeah, though preferably "git pull --rebase" first
21:42 ajonas paulproteus: yeah, did that
21:43 paulproteus Great. brb food.
21:46 DGMurdockIII left #miro-hackers
21:47 zanoi bendk: interesting, just now your first email arrived
21:47 paulproteus I seem to be much more interested in eating fruit if there is cake to eat with it.
21:47 zanoi: I wonder if your mail server does greylisting, then.
21:48 Interestingly, this holds even if it's not very good cake.
21:48 zanoi paulproteus: yes it does greylisting, but i thought it wouldnt greylist if someone replied to my mail :-/
21:48 paulproteus zanoi: The behavior you described is totally in line with greylisting.
21:49 Clearly I should get out of the programming business, and start working in Internet security. :P
21:50 zanoi paulproteus: i thought it would only temp reject if the email is from a not recognized sender
21:50 paulproteus I'll talk more about greylisting at the next break, in 20 min (-:
21:50 zanoi wikipedia seems to agree with me
21:50 heh
21:56 willkg zanoi: it really depends on how you've implemented greylisting.
21:56 zanoi: what mta and greylisting system are you using?
21:57 zanoi willkg: exim4+greylistd
21:57 willkg ok.  i used that.
21:58 zanoi willkg: did it greylist everything?
21:58 *temp reject
21:58 willkg that greylist daemon creates triples of (ip address, sender email, recipient email).  if it's seen a triple before, then it lets the email through.  otherwise it tells the sending mta to try again in 5 minutes.
21:58 it doesn't take into account any outgoing email from your server.  i'm not sure if you can configure it to do that.
21:59 a couple of years ago, i got tired of fiddling with exim's rules, so i switched to postfix.
21:59 janet_ left #miro-hackers
21:59 willkg so i think in theory, greylisting can take into account email you've sent, but that particular daemon doesn't as far as i remember.
22:00 er, where "ip address" is the ip address of the sending mta.
22:00 zanoi willkg: ok that makes sense. The mailing list subscription is to a different account and it doesn't take sent mails into account
22:00 willkg: in that way it was the first time ben sent me an email
22:00 willkg: so hows postfix?
22:00 willkg it's a ton easier to configure.
22:01 beyond that, i haven't had any troubles so i haven't fiddled with it.
22:01 zanoi willkg: well i can't imagine it getting much harder ;-)
22:02 willkg this is very much a factor of my knowledge on the systems, but i think sendmail was hardest, exim was second hardest, and postfix was nicely easy to deal with.
22:02 zanoi ah ok
22:02 willkg but, like i said, that probably says more about me than those systems.
22:03 zanoi i just remember being confused because debians config setup was different from all the tutorials on the internet
22:03 (in exim4 that is)
22:03 willkg yes.  that's also a problem.  and i found the debian docs confusing on the matter, too.
22:04 zanoi maybe i should try out postfix at some point
22:04 willkg anyhow...  i'm off for a bit.
22:04 see you all later!
22:04 willkg left #miro-hackers
22:12 paulproteus ajonas: re: latest patch: thanks, pushed!
22:13 ajonas paulproteus: i'm on a roll! though i don't know if this next one will work.
22:14 JasonWoof left #miro-hackers
22:15 ajonas paulproteus: added that one to the ticket, let me know if that's bad
22:16 paulproteus Patches on the bug tracker are what make the world go around. (-:
22:23 ajonas: Random blather of the moment: Making the the text list of account features at http://bugzilla.pculture.org/s[…]_bug.cgi?id=15428 is going to be something of a pain.
22:23 * paulproteus gets on it.
22:24 ajonas paulproteus: fwiw, i would vote for doing the limit implementation stuff first
22:24 paulproteus: if it's a big pain, we can change it to a simpler email and just say "go look at the features list" or something
22:24 paulproteus Okay!
22:25 It's, like, just 25 minutes of focused fiddling.
22:25 I'll skip it for now, though.
22:25 ajonas paulproteus: sounds good - i mean, it's nice and all, but doesn't seem critical, so let's see where we're at in a bit
22:26 paulproteus b
22:27 ajonas paulproteus: let me know if these little patches create work instead of saving time
22:27 paulproteus ajonas: They are awesome. For now they save time, rather than create work.
22:28 Sooooo awesome.
22:28 Honestly. (-:
22:28 ajonas ok, cool
22:28 paulproteus I will keep you posted in case the workflow needs changing.
22:28 ajonas great
22:31 paulproteus: when you have a chance, can you update http://bugzilla.pculture.org/s[…]_bug.cgi?id=13607 with the changes you made?
22:35 paulproteus Wow, I love automated tests.
22:37 maggie_s left #miro-hackers
22:43 morganknutson left #miro-hackers
22:44 zanoi deardiary: got a fully working shuffle repeat, but still needs some corner case love and a few more features
22:44 deardiary: one learns a new thing every day. today i learned that I could actually despise git more than I already did
22:49 paulproteus Poor zanoi.
22:51 morganknutson joined #miro-hackers
22:52 paulproteus I'm going to go to the grocery store, and then come back in a bit.
22:53 ajonas: http://bugzilla.pculture.org/s[…]_bug.cgi?id=15502 -- do we really send people *no* welcome email for signing up?
22:53 Or, let me rephrase: I can't find it and will now start looking, but let me know if I will never find it.
22:53 ajonas paulproteus: *confused*
22:53 paulproteus ajonas: eh, don't worry about it.
22:54 ajonas 16013
22:54 and there's one for free sites somewhere
22:54 paulproteus free sites is in 15502 that I linked
22:55 ajonas so does 16013 answer your question?
22:55 paulproteus Well, I have a different question.
22:55 Right now, in the 1.1.2 world, when I sign up for a site, does the system email me?
22:55 I was under the impression it already did.
22:55 ajonas paulproteus: it does
22:55 it's just not a very nice email
22:55 paulproteus Sweet. So these are just changes on that.
22:55 ajonas totally
22:55 paulproteus I'll look for the part where we send that email.
22:56 And I'm glad you're having a better time with git so far than zanoi is seeming to!
22:58 ajonas paulproteus: i've been pleasantly surprised.
22:59 DGMurdockIII joined #miro-hackers
23:00 zanoi ajonas: pleasantly surprised by git? :-)
23:00 ajonas zanoi: well, by not breaking anything yet that i can tell, by being able ot use it so far
23:00 zanoi: but i suspect i am trying to do much much smaller things. :)
23:00 * zanoi is chocked
23:00 zanoi oops
23:00 shoked i mean! :-)
23:01 *shocked
23:01 getting tired..
23:01 ajonas: but you can't tell me you find the commands intuitive? :-)
23:02 ajonas zanoi: i used a couple cheat sheets
23:02 http://www.kernel.org/pub/soft[…]ocs/everyday.html
23:02 http://hoth.entp.com/output/gi[…]or_designers.html
23:03 zanoi ajonas: thx, i'll save those
23:04 ajonas paulproteus: did you see http://bugzilla.pculture.org/s[…]_bug.cgi?id=16091 ?
23:04 zanoi ajonas: but it's still wrong if one has to look up every second command in a cheat sheet :-(
23:06 paulproteus ajonas: Just skimmed it
23:06 I have to run for now; will be back later tonight
23:06 ajonas paulproteus: cool
23:06 zanoi with other VCSs i can just type what is most intuitive with git I have to look up every command and in the end it's still borked
23:11 oh btw...I had completely missed the launch of miro community
23:12 so I tried it out recent for the first time
23:13 and it completely blew me away
23:14 ajonas zanoi: yay!
23:14 zanoi in my head i had been thinking about creating something like that, just not as good
23:14 but you had already done it..
23:15 ajonas zanoi: are you using it for anything in particular?
23:16 zanoi i am planning to in the future
23:16 i am retracing all the interesting lectures i have watched in the past years and adding them to my mirocommunity account
23:17 ajonas zanoi: neat
23:17 zanoi also if i ever find the time i'd love to try to my own instance
23:18 preferably on my sheevaplug if it can handle it :-)
23:19 ajonas zanoi: cool
23:19 zanoi though that "find the time" thing might mean it will never happen
23:19 ajonas paulproteus: when you get back, http://bugzilla.pculture.org/s[…]_bug.cgi?id=16096 is making life pretty sad for me as a curator, so if you can bounce that up the queue i'd appreciate it, if you can reproduce it
23:19 zanoi: fair
23:41 ajonas left #miro-hackers
23:59 maggie_s joined #miro-hackers

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