[07:04] <mandel> morning!
[07:29] <karni> Morning!
[08:34] <Laney> Hey, is there any way I can connect a shared folder on U1 with a differently named folder locally?
[08:35] <Laney> I have my music on U1 as Music, and on my PC the same stuff is in /srv/music; would like to keep the two in sync.
[08:44] <JamesTait> Good morning all! :)
[08:46] <JamesTait> Laney, sounds like you need someone familiar with ubuntuone-client, which is not me. :(
[08:49] <JamesTait> Laney, I'm trying to track down someone who can help you.
[08:49] <Laney> JamesTait: Ta
[08:56] <rye> Laney: Ubuntu One client at the moment does not allow adding the folders outside $HOME directory and it is not following symlinks so I am afraid it won't be possible to synchronize /srv/music
[08:56] <Laney> ah, sad
[08:57] <rye> Laney: When you say you have "music on U1 as Music", what do you mean?
[08:57] <Laney> that's the folder name
[08:59] <JamesTait> Laney, so you set up ~/Music as a synchronised folder?
[09:00] <Laney> JamesTait: No, I once synced it from ~/Music to U1, but now it's a symlink to a different drive
[09:01]  * JamesTait nods
[09:49] <mandel> Laney, so there is an issue with adding support for things outside $HOME because it could mean that is in a diff physical volume
[09:49] <mandel> Laney, that adds lots of problems, for example, ensure that when you unmount it we understand the right thing and not that you deleted the share
[09:50] <mandel> Laney, and we don't follow symlinks for the same reason
[10:04] <Chipaca> Laney: hold it, you added ~/Music/, then replaced it with a symlink?
[10:04] <gatox> good morning
[10:04] <Chipaca> Laney: tell me more
[10:04] <Laney> Chipaca: It's a new installation
[10:04] <Chipaca> mandel: we don't follow symlinks, but the top dir of a udf might be different
[10:05] <Laney> (I got an SSD and kept music on the HDD)
[10:05] <Chipaca> Laney: question, will it *always* be mounted at /srv/music?
[10:05] <Laney> yes
[10:05] <Chipaca> Laney: keep in mind that if ever the machine comes up with it unmounted and the u1 syncdaemon sees that, it'll say "oh, he's deleted all the files while i was away", and sync the deletion everywhere
[10:06] <Chipaca> Laney: so i have to ask again, will it **always** be there? :)
[10:06] <Laney> Chipaca: well, as always as it's possible to be (until the hard drives both break at once)
[10:06] <Laney> it's mounted from fstab
[10:06] <Chipaca> ah
[10:07] <Chipaca> question, is there a reason you're not mounting it on ~/Music directly?
[10:07] <Laney> and RAID1
[10:07] <Laney> yes, because there's other things on the drive that aren't music
[10:07] <Chipaca> question, is there a reason you're not bind-mounting it to ~/Music?
[10:07] <Laney> it's just a big partition with some top level folders for data
[10:08] <Chipaca> you can bind-mount from fstab also
[10:08] <Laney> I suppose I could bind mount, yes
[10:10] <Chipaca> Laney: otherwise, as i understand it, if you trick syncdaemon into thinking the toplevel folder exists already (say, bind-mount it once), then enable syncing of the folder from syncdaemon, then *quit syncdaemon* (maybe enter single-user mode just to be super safe), unmount it, and symlink it, syncdaemon should, as i understand it, not notice nor care
[10:10] <Chipaca> Laney: here be dragons, and i haven't tested this myself
[10:10] <Chipaca> the "don't follow symlink" rule is for things inside synced folders, not for synced folders themselves
[10:10] <Laney> Chipaca: I can probably bind-mount, if that will work
[10:11] <Chipaca> it does feel like the safer option to me
[10:11] <Laney> let's try
[10:11] <Chipaca> less likely to accidentally delete or move a mountpoint than a symlink
[10:15] <Laney> Chipaca: Is it still known/expected to get timeouts on the streaming player for large collections, btw?
[10:15] <Laney> 504 Gateway Time-out
[10:16] <Chipaca> Laney: hold that thought
[10:16] <Chipaca> beuno: ping. ^
[10:16] <Laney> beuno gave me a special link last time, but I (a) forgot it and (b) thought it was obsolete (re: the mail to warthogs a month or so ago)
[10:17] <Chipaca> Laney: when we went live we nuked the bootstrapping, but i thought we'd recreated it for everybody
[10:17]  * Laney nods
[10:18] <Chipaca> Laney: apparently we haven't, so beuno will sprinkle his magic dust. And soon (soon!) things will be Magic and Just Work As They Should.
[10:19] <Laney> bind-mounting seems good btw :-)
[10:20] <Chipaca> k
[10:20] <Chipaca> if it messes things up and you want to go symlink, i'd say try with a small throwaway synced folder first
[10:20] <Laney> ya
[10:20] <Chipaca> and let me know how it goes :)
=== eu is now known as Guest50686
[12:51] <ralsina> good morning!
[12:54] <gatox> ralsina, hi
[12:54] <ralsina> hola gatox
[12:59] <beuno> Laney, remind me what your sso email is?
[12:59] <Laney> beuno: iain@orangesquash.org.uk - but it looks already fixed?
[13:00]  * beuno triple checks
[13:00]  * Laney can play songs now anyway
[13:00] <beuno> right, it is
[13:00] <Laney> :-)
[13:01] <beuno> I shoould overlseep more often
[13:01] <Laney> I wish I could click the column headers to reorder search results ;)
[13:01] <karni> hi ralsina o/
[13:02] <ralsina> hi karni!
[13:13] <gatox> ralsina, my spider sense says that you want to do a code review :P https://code.launchpad.net/~diegosarmentero/ubuntuone-control-panel/new-share-design/+merge/131397
[13:13] <ralsina> gatox: sure!
[13:13] <gatox> ralsina, awesome.... my spider sense  is working \o/
[13:14] <ralsina> gatox: hehe
[13:15] <ralsina> gatox: I could ruin spider man for you by mentioning it's a very obvious analogy for puberty ;-)
[13:15] <gatox> mmmm didn't understand that part.... and i'm not sure if i want
[13:15] <gatox> jejej
[13:17] <ralsina> gatox: you really don't :-)
[13:17] <gatox> jejeje
[13:34] <dobey> lol
=== m_conley_away is now known as m_conley
[14:16] <mandel> ralsina, new super bug #1071327 ;)
[14:16] <ubot5> Launchpad bug 1071327 in Nux "StaticCairoText makes NUX get into an infinte loop" [Undecided,New] https://launchpad.net/bugs/1071327
[14:17] <ralsina> mandel: yay
[14:17] <gatox> wow
[14:20] <dobey> hmm
[14:22] <karni> o_O
[14:33] <mandel> ralsina, we are getting there :)
[14:33] <ralsina> mandel: I wonder where there is
[14:33] <mandel> ralsina, though, that blocks us a little with the payment, I found a bug with log labels, which I tried to fix using setMaxwidth, and the booom!
[14:35] <ralsina> mandel: well, an infinite loop makes sense if cairo can't actually deliver a text layout that's smaller than that
[14:35] <ralsina> mandel: I have seen similar errors in typesetting
[14:37] <mandel> ralsina, yes.. but the funny thing is that it can
[14:37] <mandel> ralsina, so, the real problem is that you cannot set the width of the layout AND the width of the text
[14:37] <mandel> ralsina, we have a work around which is to set the width of all the elements in the layout, not ideal but if we are late will get use there on time
[14:44] <mmcc> hi folks - looks like my home internet may have fixed itself. I guess it just took a nap.
[14:45] <mmcc> briancurtin, could you look at https://code.launchpad.net/~mikemc/ubuntuone-windows-installer/calling-mr-scripty/+merge/130182 when you get a chance? It's that tweak to let u1trial work on windows without a manual hack to add comtypes
[14:45] <briancurtin> mmcc: will do
[14:47] <dobey> hey mmcc
[14:48] <dobey> mmcc: any more luck with my branch/pyflakes btw?
[14:49] <mmcc> dobey, yeah - did you get my email? A quick one line fix in buildout.cfg makes it work
[14:49] <mmcc> except now it doesn't complain about 80 char lines, is that intentional (it confused me)
[14:50] <gatox> briancurtin, dobey if you can please: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/fix-tests-jenkins/+merge/131429
[14:51] <briancurtin> gatox: nice, i'll have a look
[14:52] <dobey> mmcc: what do you mean it doesn't complain? pyflakes never did; pep8 complains about line length though
[14:53] <dobey> gatox: that just moves the method to a different class? and for some reason changes the argument to "pattern" from "name" but the docstring isn't updated with the argument name
[14:54] <mmcc> dobey, oh, hm, right. I wasn't running pep8 when I was testing it. I guess my emacs error-highlighting is broken then
[14:55] <mmcc> so never mind that part. the diff in the email does work, in any case
[14:55] <mmcc> brb
[14:55] <gatox> dobey, docstring updated
[14:55] <gatox> dobey, the problem was that the  method was defined in the wrong place
[14:56] <dobey> gatox: how did the tests pass in tarmac and locally for people?
[14:56] <gatox> dobey, because everyone was running the tests on linux.... and this was in the windows part
[14:57] <dobey> gatox: then how does this fix the tests in nightlies (which are obviously on linux)?
[14:58] <gatox> dobey, this fix the tests on windows as the bug says..... the nightlies part, that is in the comment....... i wasn't able to reproduce it
[14:59] <ralsina> team, mumble in 1'
[15:00] <dobey> oh crap
[15:01] <mandel> ralsina, I can't make it :)(
[15:01] <ralsina> mandel: ack
[15:01] <mandel> that is a sad face, I promise!
[15:01] <ralsina> gatox, dobey, mvo, briancurtin, mmcc
[15:01] <mvo> me
[15:01] <briancurtin> on my way to mumble, not starting up nicely for some reason
[15:02] <mvo> ups, team call, hold on a sec
[15:03] <briancurtin> i'm going to reboot...mumble doesn't want to work
[15:03] <mmcc> d'oh, be right there
[15:03] <ralsina> COME ON :-)
[15:04] <ralsina> dobey?
[15:08] <briancurtin> reconfiguring mumble...
[15:10] <briancurtin> anyone else having issues connecting to the server? i'm getting "a referral was returned from the server" message on startup
[15:11] <mmcc> briancurtin, no issues here
[15:11] <briancurtin> haha...i wonder if it has to do with the certs i manually removed from my system to test what i was working on
[15:14] <briancurtin> yep, certificate issue...
[15:24] <dobey> mmcc: ah, actually, i need to make a new pyflakes i guess. i see why the args for main is there, and it makes no sens to me
[15:25] <mmcc> dobey, ok, huh.
[15:26] <dobey> it's a patch, which isn't from upstream, and which seems to have no rationale in the debian/changelog for pyflakes. so i'll remove it and rebuild the tarball
[15:33] <dobey> doh, i need to add all my other accounts to evolution so i can send mail for them :-/
[15:43] <karni> Started "normally"/earlier today, I will EOD now. Need to run. I'm travelling tomorrow on a train, will send out an e-mail to ubunet
[15:43] <karni> Have a great weekend, all! And have great holiday, ralsina :)
[15:44] <mvo> bye karni
[15:45] <karni> o/
[15:47] <dobey> bye karni
[15:52]  * gatox lunch
=== gatox is now known as gatox_lunch
[15:58] <mmcc> brb, have to watch the kid for a couple minutes
[16:01] <dobey> mmcc: uploaded pyflakes tarball without that weird patch. if you could try my branch again to verify it works, when you return, that would be awesome.
[16:05] <dobey> ok, i'm off to get lunch. bbiab
[16:09] <ralsina> Off for lunch & banks, wil be back in (hopefully) 90'
[16:11] <mandel> ralsina, where you in a plain recently, right? http://bit.ly/XmGa0z
[16:12] <mmcc> not back yet, but had a sec - dobey your new package works (without my buildout.cfg tweak.)
[16:24] <mmcc> I could still use a couple reviews for u1-client: https://code.launchpad.net/~mikemc/ubuntuone-client/add-downloads-to-menudata/+merge/131231 and https://code.launchpad.net/~mikemc/ubuntuone-client/no-multi-add-recents-fix-1069953/+merge/131233
=== gatox_lunch is now known as gatox
[16:47] <gatox> back
=== Chipaca is now known as ChipAway
[17:09] <gatox> mmcc, we can exchange reviews if you want
[17:10]  * gatox really needs reviews jeje
[17:13] <mmcc> gatox I'm reviewing your cp files search branch now
[17:13] <gatox> awesome!
[17:14] <gatox> mmcc, i'll take the menuddata one
[17:17] <gatox> mmcc, this add the download data to the sync-menu everywhere?
[17:18] <mmcc> gatox - what do you mean by everywhere?
[17:19] <gatox> mmcc, nothing, i see it does..... also, this can be tested on linux with d-feet
[17:19] <gatox> consuming the dbus service from there
[17:19] <mmcc> gatox - looks like we missed something in the u1-client ipc for your new search_files call to work on mac & win: http://paste.ubuntu.com/1305374/
[17:20] <mmcc> this IPC code, man, I just don't know
[17:20] <gatox> mmcc, already fixed in this one: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/fix-tests-jenkins/+merge/131429 if you want to review it
[17:21] <mmcc> gatox, great! looking now
[17:25] <dobey> sigh, email
[17:26] <dobey> mmcc: care to approve my branch then if it works for you now? :)
[17:27] <mmcc> btw gatox, I noticed that the u1client search-filter branch still had this commented line in it: " #basename = os.path.basename(p)" in filesystem_manager.py line 1374. since fix-tests-jenkins is a cleanup branch should we fix that in there?
[17:27] <mmcc> dobey, oh, yeah sure
[17:27] <gatox> mmcc, ok
[17:29] <gatox> mmcc, done
[17:29] <mmcc> great, thanks gatox
[17:30] <gatox> oops!! the departure time is not the one that i thought..... i will need to leave one hour earlier
[17:30]  * gatox almost misses his flight
[17:31] <briancurtin> gatox: i owe you a review from earlier - i'll make sure to get to it
[17:31] <gatox> briancurtin, thx
[17:31] <mmcc> gatox, so you have to go now? have a safe flight. I'm +1 on both these, fyi
[17:32] <gatox> mmcc, i'll be here for a while....... but yep..... in around 30 min or so
[17:32] <gatox> mmcc, awesome for the +1 :D
[17:32] <gatox> mmcc, i'm running the tests for your branch.... code seems ok
[17:32] <mmcc> although the bug title and branch title is confusing :P - "files that were shared during the time controlpanel was open" - but we're searching files that aren't shared, right?
[17:34] <briancurtin> mmcc: i think you were also seeing the "file sync starting..." showing up forever on windows?
[17:35] <mmcc> briancurtin, yes I was, maybe…
[17:36] <briancurtin> i'm seeing it right now, KeyError inside tritcask that looks like it hung up the whole SD
[17:36] <briancurtin> fun times
[17:36] <mmcc> briancurtin I think in my case it was the issue with the log files being owned so I can't rotate them
[17:36] <briancurtin> ah yeah
[17:36] <mmcc> I hacked around that by setting the log file size to BIGNUM
[17:36] <mmcc> because I think you mentioned it was only an issue when running from source
[17:37] <gatox> mmcc, +1 to this one: https://code.launchpad.net/~mikemc/ubuntuone-client/add-downloads-to-menudata/+merge/131231
[17:37] <mmcc> gatox great, thanks!
[17:38] <mmcc> gatox, the other one is really short… I'll understand if you have to leave tho
[17:39] <gatox> mmcc, give me a couple of minutes.... and i'll try to look at it
[17:40] <dobey> grr, wtf evolution
[17:41] <briancurtin> anyone knowledgeable in tritcask? wondering what a proper behavior would be when receiving this KeyError - https://pastebin.canonical.com/77268/ - log it, return None, make sure everything else up the chain will handle None?
[17:42] <gatox> briancurtin, you can ask verterok
[17:42] <briancurtin> verterok: any opinion on the above paste? i'm seeing this on startup on Windows
[17:45] <briancurtin> i think this is causing me to be stuck in local rescan
[17:47] <verterok> briancurtin: looking
[17:48] <dobey> are there branches about to land in u1client?
[17:48] <verterok> briancurtin: looks like the metadata for a share is missing
[17:48] <verterok> briancurtin: isn't a tritcask error perse
[17:49] <verterok> briancurtin: local rescan is failing because it's unable to find the medata for a share/volume
[17:53] <mmcc> dobey: yes, a few.
[17:55] <briancurtin> verterok: interesting. i'll further into it
[17:55] <mmcc> dobey well, maybe not approved and waiting to land but there are a few in line - why?
[17:56] <dobey> mmcc: because my cp branch will require a new build of nightlies to be installed, in order to land; and given that nightlies have failed to build since gatox's share-files branch, there's a bit of a problem :)
[17:58] <mmcc> aha. well, his branch to fix that should be able to land soon… I just approved it
[17:58] <mmcc> IIRC it does need one more
[18:01] <dobey> well i don't think it fixes the test failure in nightlies
[18:01] <mmcc> oh, ok. what's that one?
[18:02] <dobey> https://bugs.launchpad.net/ubuntuone-client/+bug/1070991/comments/1
[18:02] <ubot5> Ubuntu bug 1070991 in Ubuntu One Client "Tests broken in windows cause of the search filter branch" [Undecided,In progress]
=== yofel_ is now known as yofel
[18:04] <mmcc> yeah, huh. those seem like different bugs
[18:04] <dobey> or not
[18:05] <dobey> test isolation issues can seem that way, and totally not be :)
[18:06] <mmcc> well, I remember thinking when I reviewed that branch that I wasn't sure all those path name tests would pass on windows, but I thought a 'windows review' was pending from someone else. I should've just checked myself :(
[18:09] <mmcc> so I'm doing that now
[18:10] <dobey> man, amazon picks some odd 'genre' tags for their mp3s sometimes it seems
[18:11] <mmcc> ugh, every ipc test is timing out on windows
[18:14] <gatox> mmcc, +1 to the other one.......
[18:14] <mmcc> gatox, great - thanks!
[18:14] <gatox> and nowwwww i need to run to the airport!! byeeeee.... see you on monday people
[18:14] <mmcc> and have a good trip!
[18:14] <briancurtin> bye gatox
[18:14] <gatox> thx.... bye!
[18:20] <mmcc> dobey yeah, I guess that branch might fix the nightlies issue after all. maybe our test dbus bus instance is ignoring that error and just returning []?
[18:20] <dobey> not sure
[18:21] <dobey> if it still fails i'll open another bug and i guess look into it myself, since he's gone til monday now
[18:36] <dobey> brb, gotta run real quick.
[19:08] <dobey> bah. it still fails :(
[19:21] <briancurtin> back in a bit - late lunch
[19:42]  * briancurtin back
[19:48] <briancurtin> ralsina: i figured out the cert thing, just need to clean it up to actually be usable (it's hardcoding some paths just so i could get it going).
[19:48] <ralsina> briancurtin: cool!
[19:50] <dobey> nice
[20:07] <dobey> bah
[20:07] <dobey> it's almost like this test is somehow talking to an actual instance of ubuntuone-syncdaemon, rather than the fake service in the tests; but only in the nightlies builds
[20:07] <dobey> sigh
[20:39] <briancurtin> ralsina: since you are going to be gone, what are the guidelines we need to follow for the windows release? is there anything more needed than this cert stuff, working tests, and a sign off from QA, or are there other things we are waiting on?
[20:40] <ralsina> briancurtin: that is enough
[20:40] <ralsina> briancurtin: in an ideal world, it would have diego's improved shares tab but that's not going into stable-4
[20:40] <ralsina> briancurtin: so we may have to do a 4.1 or whatever in a few weeks
[20:41] <briancurtin> ralsina: alright. i'm currently seeing an issue where local rescan fails and hangs SD because of some metadata mismatch. im going to try the cert branch in a clean VM to see that it's truly working
[20:41] <ralsina> briancurtin: ack
[20:42] <dobey> ok, i am about to propose a branch that just skips this test, because i can't tell what the heck is going on; not sure how to debug further; and gatox is gone til monday
[20:43] <dobey> and i want to get the pyflakes switch done already
[20:44] <mmcc> dobey it's just one test? I thought it was a bunch that were failing…?
[20:45] <dobey> no, just the one
[20:46] <ralsina> dobey: I'll rubberstamp it
[20:46] <ralsina> dobey: as long as you file a bug about it ;-)
[20:47] <dobey> https://code.launchpad.net/~dobey/ubuntuone-client/skip-search-files-test/+merge/131484
[20:47] <dobey> yes i already filed the bug. was going to fix it, but then saw it was a bit too weird to fix quickly
[20:48] <ralsina> dobey: maybe the nightlies will sleep it off over the weekend
[20:48] <ralsina> dobey: our build slaves may just be tired
[20:49] <dobey> nope
[20:50] <dobey> it's definitely a problem; just not sure whate exactly, as it's quite odd. it is however, also quite odd that *only* that test fails in this way
[20:51] <dobey> and doh, i totally just forgot to import the skipTest call
[20:53] <dobey> and actually, now i spot another issue in that test, which should have been caught in review :-/
[20:55] <ralsina> I have been doing crappy reviews for at least a week :-(
[20:55]  * ralsina slaps himself into shape... oh, wait, vacations start in 5 minutes.
[20:56] <dobey> paths with hardcoded / separators in them
[20:57] <dobey> well, feel free to rubberstamp https://code.launchpad.net/~dobey/ubuntuone-client/skip-search-files-test/+merge/131484 too :)
[20:57] <mmcc> dobey, I saw those paths too but they don't cause a problem on windows,so …
[20:57] <dobey> mmcc: well, the test was failing on windows, so… :)
[20:58] <dobey> and still failing apparently :-/
[20:59] <mmcc> I'm confused. I saw no failures due to path issues, only IPC timeouts…
[20:59] <mmcc> (on windows)
[21:00] <mmcc> i wonder if I was testing the right test_tools.py
[21:00] <mmcc> I did selective testing because the timeouts were screwing up the whole build
[21:00] <mmcc> and of course there are two test_tools.py
[21:00]  * mmcc goes back to windows
[21:01] <mmcc> oh, before ralsina leaves - would anyone complain if I used mock (not mocker) in the new cocoa-menu project? I'm going to have to mock out a lot of Cocoa and it was starting to get big
[21:01] <ralsina> mmcc: I can never get those two straight
[21:02] <ralsina> mmcc: yes to the one that is going to be in the stdlib in 3.x
[21:02] <ralsina> mmcc: the one done by our friend and coworker mfoord, IIRC
[21:02] <dobey> mmcc: yeah it's fine to use python-mock there
[21:02] <mmcc> mock is the stdlib one
[21:02] <mmcc> great, that's going to be much nicer
[21:03] <ralsina> and no to the one done by another friend and coworker, that we used before, and that drived me nuts with useless failures whenever I moved a line
[21:03] <dobey> mocker
[21:04]  * ralsina starts writing mockest
[21:05] <ralsina> dobey: STAMP on skip-search-files-test
[21:08] <dobey> yay
[21:12] <ralsina> EOW for me, people
[21:12] <dobey> bye ralsina
[21:12] <ralsina> I will be around for a little bit tomorrow very early, then just email, then back on thursdays, then at the sprint
[21:12] <ralsina> if you REALLY need me, I will leave a phone number somewhere as soon as I have one
[21:13]  * ralsina goes pack
[21:13] <briancurtin> enjoy your time off!
[21:13] <ralsina> thanks briancurtin!
[21:37] <dobey> mmcc: can you review https://code.launchpad.net/~dobey/ubuntu-sso-client/lint-roller/+merge/131016 now that fixed pyflakes bits are in for buildout?
[21:38] <mmcc> dobey yeah sure.
[21:38] <briancurtin> "yeah sure, or yes sir"
[21:39]  * mmcc doesn't get it
[21:39] <briancurtin> from Super Troopers
[21:39] <briancurtin> and mmcc i still owe that review for mr.scripty. if i dont get to it by EOD i'll have it first thing tomorrow
[21:39] <mmcc> o rite. I really need to re-watch super troopers again. It's been too long
[21:40] <mmcc> no big hurry on mrscripty
[21:51] <dobey> ok, need to roll. later all
[21:52] <dobey> mmcc: approved your two client branches btw
[21:52] <dobey> ok, really bye :)
[21:53] <mmcc> thanks!
[21:53] <mmcc> and bye
[22:13] <ralsina> dobey: silly question... what do you think of this monitor http://www.ebay.com/itm/ACHIEVA-Shimian-QH270-Lite-Quad-HD-FREE-Voltage-Monitor-2560x1440-16-9-D-Sub-NEW-/280928105274?pt=Computer_Monitors&hash=item41689eb33a
[22:49] <mmcc> ok, time to go for me. so far, liking mock a bunch. might come back tonight, or just start early tomorrow