| rye | i know what's missing! | 10:51 |
|---|---|---|
| rye | progress bar of file upload on the web ui! | 10:51 |
| rye | content-length for form upload is sent, need a background request to update the progress. | 10:51 |
| === teknico is now known as teknico_away | ||
| duanedesign | hello rye | 12:26 |
| rye | duanedesign, hello :) | 12:26 |
| rye | duanedesign, have you used gedit? | 12:54 |
| duanedesign | rye: yes | 12:58 |
| rye | duanedesign, ok, have you used it on remote filesystems? I mean i am debugging gedit file save and timestamp update and it looks like there is something happening behind the scenes, but strace looks legitimate... | 12:59 |
| duanedesign | rye: hmm. | 13:04 |
| duanedesign | rye: i am working on some files on my duanedesign@people.ubuntu.com/ right now | 13:13 |
| * rye discovered bootchart | 13:14 | |
| === teknico_away is now known as teknico | ||
| duanedesign | that came in handy the other day. A guy came into #ubutnu-beginners and half his processes werent starting. Was able to use that to find the offending process that was causing the problem | 13:15 |
| Pretto | to share a folder to other user, does he need to have an ubuntu one account? | 14:04 |
| beuno | Pretto, he will need to in order to access the files | 14:04 |
| beuno | but not to offer the share | 14:04 |
| beuno | he can create an account after you share with him | 14:04 |
| Pretto | beuno: thank you | 14:05 |
| mattgriffin | rmcbride: ping | 14:07 |
| rmcbride_ | mattgriffin: hi | 14:08 |
| mattgriffin | rmcbride: hi rick. are you a moderator of this channel? can you change the topic? | 14:08 |
| rmcbride_ | mattgriffin: I can change the topic sure | 14:08 |
| rmcbride_ | (from the machien I'm logged on with out the _) | 14:09 |
| * rye gedit&u1 make me angry, went to get some foods | 14:11 | |
| rodrigo_ | aquarius: why is https://bugs.launchpad.net/ubuntuone-servers/+bug/400153 marked as incomplete? | 14:40 |
| ubottu | Error: This bug is private | 14:40 |
| * aquarius looks | 14:40 | |
| aquarius | rodrigo_, because we haven't done it yet. :) | 14:41 |
| rodrigo_ | aquarius, ah, I thought incomplete is for bugs missing information, right? | 14:41 |
| rodrigo_ | not not-done-yet ones? | 14:41 |
| aquarius | rodrigo_, the original intent of that bug was that you'd be able to talk to couch entirely over D-Bus, and not talk http at all. | 14:41 |
| aquarius | it's "incomplete" because it's missing information, specifically whether we think it's a good idea or not. | 14:41 |
| aquarius | I am inclined to invalid it. | 14:41 |
| aquarius | ooh, is it bug day? | 14:42 |
| rodrigo_ | as you wish | 14:42 |
| rodrigo_ | yes, bug day | 14:42 |
| rodrigo_ | aquarius, so, should I mark it as invalid then? | 14:42 |
| aquarius | rodrigo_, actually, wontfix, I think. It's a reasonable request, we're just not going to do it that way. :) | 14:43 |
| rodrigo_ | ok | 14:43 |
| aquarius | blimey, we must have done all the "easy" bugs now. All of these are quite hard to decide how important they are :) | 14:44 |
| rodrigo_ | :) | 14:45 |
| rodrigo_ | aquarius, another one, what is this https://bugs.launchpad.net/ubuntuone-servers/+bug/400315 for? | 14:46 |
| ubottu | Error: This bug is private | 14:47 |
| aquarius | ha. | 14:47 |
| aquarius | invalid. | 14:47 |
| rodrigo_ | ok | 14:47 |
| aquarius | the idea of control channels has gone away, pretty much. | 14:47 |
| aquarius | rodrigo_, I'll trade you those two for https://bugs.edge.launchpad.net/evolution-couchdb/+bug/504282 because I don't understand what's causing the problem, so I don't know how important it is? | 14:48 |
| ubottu | Ubuntu bug 504282 in evolution-couchdb "Evolution CouchDB Error" [Undecided,Incomplete] | 14:48 |
| rodrigo_ | aquarius, ok | 14:48 |
| rodrigo_ | aquarius, I've already marked them as won't fix/invalid, but yeah, have a look at them and add whatever you want | 14:48 |
| * rodrigo_ looks at the evo-couchdb bugs | 14:48 | |
| rodrigo_ | aquarius, hmm, is the 'The name org.freedesktop.Avahi was not provided by any .service files' a cause for desktopcouch to not start? | 14:51 |
| aquarius | rodrigo_, it might be. DC looks for avahi to publish its pairing record, if it's paired with anything. | 14:51 |
| aquarius | why would someone not have avahi? | 14:51 |
| rodrigo_ | not sure, but the evo-couchdb logs show it not being able to contact dc | 14:52 |
| aquarius | that could be the problem. I don't know how to resolve that, though -- I can't think of any good reasons why avahi wouldn't be there? It's part of the platform! | 14:53 |
| rodrigo_ | I'll tell the user to install avahi, if it's not installed | 14:53 |
| === CardinalFang__ is now known as CardinalFang | ||
| duanedesign | rmcbride_: I was coming up with a list of test to run on nightlies. It was mentioned you have some automated acceptance tests that you use? | 14:59 |
| rmcbride_ | duanedesign: sort of | 15:00 |
| mattgriffin | something's odd so i can't update the topic... here's info about how you can help with Bug Day: https://wiki.ubuntu.com/UbuntuOne/BugDays | 15:00 |
| rmcbride_ | duanedesign: I have tests that can be run that create test data | 15:00 |
| * rodrigo_ reboots | 15:00 | |
| rmcbride_ | duanedesign: currently the verification is manual, and the test sets take a long time to sync | 15:00 |
| rmcbride_ | duanedesign: it's really rather evil test data (max POSIX FS depth, etc) | 15:00 |
| rmcbride_ | duanedesign: it's available at lp:ubuntuone-testing | 15:01 |
| rmcbride_ | duanedesign: There are plans to greatly enhance both the tests as they are, to create a nice checkbox-based runner for people to be able to run the suite or parts of it. | 15:01 |
| rmcbride_ | duanedesign: and complete automation in VMs as well. | 15:02 |
| CardinalFang | Editing my grub config remotely. I told you I was hardcore. | 15:02 |
| aquarius | CardinalFang, that sounds like the quick path to a scragged unbootable machine, to me -- you're braver than I am :) | 15:04 |
| duanedesign | rmcbride_: I had an informal list i use. As i added on to it i realized it wouldnt be much more effort to come up with a list like the testcases on the QA site | 15:04 |
| rmcbride_ | duanedesign: the stuff I have is structurally based on what's on the QA site. I have yet to port the formerly internal wiki pages to that location | 15:05 |
| rmcbride_ | duanedesign: to the test set that I have, I also typically check that problems in the changelog as fixed are truely fixed | 15:05 |
| CardinalFang | I don't really need that machine today. I just don't want it to burst into flames. | 15:06 |
| rmcbride_ | duanedesign: we've held off on nightlies for the last couple of weeks due to some protocol capabilities changes that would break clients built from current trunk. | 15:06 |
| rmcbride_ | duanedesign: hopefully that will change this week. | 15:06 |
| rmcbride_ | the package builds will be automated and REALLY nightly once we resume building | 15:06 |
| duanedesign | rmcbride_: nice. I'll take a look at ubuntuone-testing. | 15:07 |
| mattgriffin | rodrigo_: ping | 15:08 |
| rodrigo_ | mattgriffin, pong | 15:08 |
| rmcbride_ | duanedesign: cool :) it can use some polish, but using doctests makes it easy to parse anyway | 15:08 |
| mattgriffin | rodrigo_: can you take a look at this bug? https://bugs.launchpad.net/ubuntuone-servers/+bug/515380 | 15:08 |
| ubottu | Ubuntu bug 515380 in ubuntuone-servers "Error U1 in web intrface for contacts" [Undecided,New] | 15:08 |
| mattgriffin | rodrigo_: seen it before? | 15:08 |
| rodrigo_ | mattgriffin, ok | 15:08 |
| rodrigo_ | mattgriffin, teknico might know better | 15:09 |
| rodrigo_ | teknico, ^^ | 15:09 |
| mattgriffin | rodrigo_: ok. i'll ping him | 15:09 |
| mattgriffin | rodrigo_: thanks! | 15:09 |
| rodrigo_ | you're welcome :-) | 15:09 |
| teknico | mattgriffin, rodrigo_, looking at it | 15:10 |
| mattgriffin | teknico: thanks | 15:12 |
| teknico | mattgriffin, that bug is a duplicate of #506584 | 15:14 |
| mattgriffin | teknico: cook. thanks! | 15:14 |
| teknico | mattgriffin, sorry, I can't cook to save my life ;-) | 15:14 |
| mattgriffin | :) | 15:15 |
| mattgriffin | haha... cool :) | 15:15 |
| rodrigo_ | hmm, we don't do standups anymore? my DONE list is getting long :-) | 15:19 |
| dobey | rodrigo_: we don't do it on tuesday, since tuesday is bug day | 15:19 |
| rodrigo_ | yeah, but yesterday, and Friday, there wasn't a standup, right? | 15:20 |
| rodrigo_ | or did I miss it? | 15:20 |
| dobey | i don't know | 15:20 |
| dobey | friday at that time i would have been sleeping still, being in -0800 | 15:21 |
| dobey | don't know about yesterday | 15:21 |
| * rtgz is now rye | 15:45 | |
| === rye changed the topic of #ubuntuone to: Help with bugs: https://wiki.ubuntu.com/UbuntuOne/BugDays | https://one.ubuntu.com | https://launchpad.net/ubuntuone | Please honk if you want a music store | ||
| rye | facundobatista, bug 426930, syncdaemon log seems to be missing any upload directives... the client just stays in START_WORKING_ON_METADATA_WITH_CONTQ | 15:59 |
| ubottu | Launchpad bug 426930 in ubuntuone-client "Files not uploading" [High,Incomplete] https://launchpad.net/bugs/426930 | 15:59 |
| === jamalta-afk is now known as jamalta | ||
| facundobatista | rye, let me see (btw, welcomed) | 16:02 |
| verterok | rye, facundobatista: bug 517361 | 16:06 |
| ubottu | Bug 517361 on http://launchpad.net/bugs/517361 is private | 16:06 |
| rye | hm | 16:07 |
| rye | ok, how do I make my syncdaemon hang this way?... | 16:07 |
| verterok | rye: no idea :( | 16:08 |
| rye | jblount, bug #506601 - erm... in case avahi is not enabled, then desktopcouch replication silently stops working. Is it really "Low" Importance? | 16:11 |
| ubottu | Launchpad bug 506601 in ubuntuone-client "Pairing and replication does not work if Avahi is down" [Low,Triaged] https://launchpad.net/bugs/506601 | 16:11 |
| * rye is still playing with gedit... | 16:11 | |
| rye | so far gedit wins | 16:11 |
| jblount | rye: Feel free to adjust, but that's what I was thinking. | 16:12 |
| facundobatista | verterok, why is this happening? is new? | 16:12 |
| verterok | facundobatista: I don't think it's new | 16:13 |
| verterok | facundobatista: but I couldn't find older bugs, probably marked as dupe of a AQ bug | 16:13 |
| facundobatista | verterok, why logs don't tell the free space the user have? | 16:13 |
| verterok | facundobatista: no idea, that info isn't logged? :( | 16:14 |
| facundobatista | verterok, don't know | 16:14 |
| verterok | facundobatista: in the case of aquarius, the free space isn't a problem... ~2GB of 50GB | 16:14 |
| aquarius | ? | 16:15 |
| rye | verterok, i believe 517361 can be made public with private log removed and 426930 can be made as duplicate of the first one, just to tie them together | 16:15 |
| facundobatista | verterok, AQ is not logging if the Queue is blocked by something | 16:16 |
| verterok | aquarius: the hung ActionQueue in your syncdaemon :) | 16:16 |
| facundobatista | verterok, in any case, don't know why it could hang | 16:16 |
| aquarius | verterok, ah right, yeah :) | 16:16 |
| verterok | facundobatista: bad AQ! :) | 16:17 |
| verterok | rye: yes, as the logs don't have any useful info | 16:17 |
| rye | verterok, removing logs from 517361 and making it public.. | 16:29 |
| verterok | rye: ack, thanks! :) | 16:30 |
| mandel | teknico, ping | 18:00 |
| teknico | mandel, pong | 18:01 |
| mandel | teknico, mind giving me a hand with my python? | 18:01 |
| mandel | teknico, is about desktopcouch ;) | 18:01 |
| teknico | mandel, ehm, I sure hope there's no double-entendre in there ;-) | 18:01 |
| teknico | mandel, jokes aside, sure | 18:02 |
| mandel | teknico, hehe good one :D | 18:03 |
| mandel | teknico, the implementation of remove is missing from the MergeableList and I wanted to implement it, the way I see it is that looping through the keys and removing the first match will be close to the performance of O(n) plus the time it takes to get the keys... | 18:04 |
| mandel | but I have a problem with the comparison since I cannot the a cmp(value_to_remove, self._data[key]) because it could be a list and all lists are mergeable list | 18:05 |
| mandel | ups, I mean "I cannot do a simple cmp" | 18:05 |
| teknico | mandel, sorry, in a call now, you'll be better off asking thisfred anyway, I guess | 18:06 |
| mandel | is there a better way or do I have to implement the __cmp__ of the MergeableList to handle lists, other mergeable list et.. | 18:06 |
| mandel | teknico, ok, np :P | 18:06 |
| mandel | thisfred, ping | 18:06 |
| thisfred | mandel: pong | 18:07 |
| mandel | thisfred, can you help me with the above? | 18:07 |
| thisfred | mandel: reading through it now | 18:07 |
| thisfred | mandel: ah | 18:09 |
| thisfred | the values in a mergeable list will not be simple values usually, so you'd be comparing dictionaries mostly | 18:10 |
| thisfred | aquarius: any thoughts on this? To be honest I never thought of removing by value. It's not a list feature I often use anyway | 18:12 |
| mandel | thisfred, indeed, but the data from the user does not have to, what I mean is that we could have remove([2,3,4]) where the list is converted to a MergeableList, that means that either I implement a __cmp__ in the mergeable list that can work with lists or tuples and | 18:12 |
| thisfred | mandel: yeah. That's doable: a mergeable list can behave like a list already, so you can just loop through the values in order and compare them. | 18:13 |
| thisfred | Efficient it won't be, but I doubt we'll be seeing monstrous mergeable lists in desktopcouch records. | 18:13 |
| mandel | thisfed, the remove is needed 'cause the onyl way to get the keys is by looking in _data | 18:14 |
| thisfred | Though it never hurts to not make the unusual impossible | 18:14 |
| thisfred | mandel: huh, you don't need the keys: | 18:14 |
| thisfred | doesn't pop(x) work? | 18:14 |
| thisfred | if not, then we need to fix that | 18:14 |
| mandel | AttributeError: 'MergeableList' object has no attribute 'pop | 18:15 |
| mandel | I was going to file a bug but I wanted to have a patch first | 18:16 |
| mandel | and pop will have the same issue for performance since pop does not take the value in a dict and popitem will not pop the item you want but one at random | 18:17 |
| mandel | my initial idea was to implement a BiHashMap but I have problems with the cmp too since I cannot be using a dict as a key | 18:20 |
| dobey | hmm | 18:20 |
| mandel | dicts are not hashable... | 18:20 |
| rye | bug | 18:27 |
| rye | uploaded a file with cyrillic name via web interface, tried to download via web, the file is said to be named as '????.flv' | 18:28 |
| mandel | sorry machine gave up on me, did I miss anything? | 18:30 |
| mandel | thisfred, I was wondering why are the list made mergeable list so early? can the be just simple list and modify the json serialization instead? | 18:32 |
| thisfred | mandel: doesn't pop(x) work? | 18:34 |
| mandel | thisfred, no: AttributeError: 'MergeableList' object has no attribute 'pop' | 18:35 |
| thisfred | mandel: we want to be able to do comparisons/merging/conflict resolution in the clients, rather than in couchdb, which provides no ways to do this | 18:35 |
| thisfred | mandel: ok, then we should definitely fix that | 18:36 |
| rye | hmmmmmmmm | 18:36 |
| rye | vim zig.txt; ls --full-time zzz.txt | 18:36 |
| rye | -rw-r--r-- 1 rtg rtg 7 2010-02-09 20:35:46.000000000 +0200 zzz.txt | 18:36 |
| rye | ubuntuone updates file... | 18:36 |
| rye | ls --full-time zzz.txt | 18:36 |
| rye | -rw-r--r-- 1 rtg rtg 7 2010-02-09 20:35:52.000000000 +0200 zzz.txt | 18:36 |
| mandel | thisfred, certainly, but implementing pop is going to be as hard as to implement remove | 18:37 |
| thisfred | mandel: well, it's by position, so the comparison of the value doesn't need to happen. | 18:38 |
| mandel | thisfread, true, pop will not need the cmp | 18:38 |
| rye | ubuntuone-nautilus ??? | 18:38 |
| rye | grrr | 18:39 |
| dobey | rye: huh? | 18:39 |
| thisfred | mandel: hang on, we do implement __delitem__ | 18:39 |
| rye | dobey, no, it's my debug info on files that get timestamp updated... checking what version i am actually running | 18:39 |
| thisfred | mandel: so implementing pop is easy | 18:40 |
| dobey | ok | 18:40 |
| mandel | thisfread, yes, but with a key, which I do not know unless I loop though the list.. kinda lame | 18:41 |
| thisfred | def pop(self, index): | 18:41 |
| thisfred | result = self[index] | 18:41 |
| thisfred | del self[index] | 18:41 |
| thisfred | return result | 18:41 |
| thisfred | mandel: you never need the keys | 18:42 |
| thisfred | mandel: everything works by numerical index | 18:42 |
| thisfred | right now you can already do del my_mergeable_list[3] | 18:42 |
| thisfred | which is awful since we're mixing dict and list syntax | 18:43 |
| mandel | thisfred, but I need the index of what I'm removing... which then makes useless the need of calling get)uuid_for index cause I already did a loop around | 18:43 |
| mandel | what I mean is that if I have a list, and I do not know the position of the data I want to be able to do remove, which should be O(n) but it will not be so any app that works assuming that it has a list because it put a list will brake... | 18:44 |
| mandel | thisfred, so far i have work around this because I know the internal representation, but it certainly not a nice way to do it | 18:45 |
| thisfred | mandel: agreed, I think we need support for normal lists back | 18:45 |
| thisfred | and make mergeable list an explicit feature | 18:46 |
| rye | yup | 18:46 |
| mandel | thisfred, lets get CandinalFang involved ;) | 18:46 |
| rye | nautilus plugin, it is utiming() the file | 18:47 |
| rye | :) | 18:47 |
| thisfred | mandel: indeed, since I currently don't have any time to work on this :) | 18:47 |
| mandel | thisfred, I was going to do it 'cause i'm cleaning my code and this looks ugly :P | 18:47 |
| mandel | thisfred, if you give me some feedback I'll give it a go | 18:48 |
| thisfred | mandel: note though that in most cases lists are iterated over or accessed by index. Maybe that's just my style of using python, though | 18:48 |
| mandel | well, doing a for to find what you want and remove it is the same as calling remove | 18:49 |
| thisfred | mandel: that I can certainly do. We need some way to signify in the JSON that something is a mergeable list, rather than a dictionary. An "_is_mergeable_list": true property would be fine with me | 18:49 |
| mandel | thisfred, and maybe the remove implementation is more efficient, that would depend on CPython | 18:49 |
| thisfred | mandel: yeah, I hardly ever remove stuff from lists, I think, but then that's kind of essential to an app that stores lists I realize | 18:51 |
| mandel | thisfred, give me some time to think while I have a coffee and I'll be back to ask some more (and with some code) | 18:51 |
| thisfred | mandel: then, instead of moving all lists in python to mergeable lists, we should provide API to generate them, and leave regular lists as regular lists | 18:52 |
| thisfred | mandel: great, I have to go and walk the dog now anyway | 18:52 |
| rye | duanedesign, grrr, posted the reponse in bug 501493 but said that that was me who wrote #2... | 18:54 |
| ubottu | Launchpad bug 501493 in ubuntuone-client "modification time changes after upload" [Undecided,Confirmed] https://launchpad.net/bugs/501493 | 18:54 |
| rye | duanedesign, sorry about that | 18:56 |
| rye | ok, gedit done, this is definitely bug 491777 | 18:57 |
| ubottu | Launchpad bug 491777 in ubuntuone-client "Ubuntuone-Nautilus should not update file modification time to refresh the emblems" [Medium,Triaged] https://launchpad.net/bugs/491777 | 18:57 |
| rye | he he | 18:59 |
| rye | :) | 18:59 |
| rye | the irony is that because the fix was not applied in full, the files do not get utime()d on simple metadata query, they are only utime()d on upload... | 19:01 |
| dobey | rye: you're testing the stable branch or something? | 19:11 |
| rye | dobey, yes, actually it is 1.0.3-0ubuntu1 | 19:12 |
| rye | dobey, this is all fixed in trunk | 19:13 |
| dobey | yes i know | 19:13 |
| dobey | hrmm | 19:20 |
| rye | guys, could you please pastebin me the output of tracepath -n files.one.ubuntu.com | 19:20 |
| dobey | 13: no reply | 19:21 |
| dobey | 14: no reply | 19:21 |
| dobey | 15: no reply | 19:21 |
| dobey | 16: no reply | 19:21 |
| dobey | 17: no reply | 19:21 |
| dobey | i don't think it's going to get to it :) | 19:21 |
| * dobey wonders what happened to xtraceroute | 19:23 | |
| aquarius | 16: no reply | 19:23 |
| aquarius | 17: no reply | 19:23 |
| aquarius | 18: no reply | 19:23 |
| aquarius | 19: no reply | 19:23 |
| verterok | 17: no reply 18: no reply 19: no reply 20: no reply 21: no reply | 19:24 |
| rye | nobody replies :) | 19:25 |
| mandel | 1: 192.168.1.10 0.109ms pmtu 1500 | 19:27 |
| mandel | 1: 192.168.1.1 2.271ms | 19:27 |
| mandel | 1: 192.168.1.1 1.946ms | 19:27 |
| mandel | 2: 192.168.1.1 1.980ms pmtu 1488 | 19:27 |
| mandel | 2: 81.241.62.1 23.548ms | 19:27 |
| mandel | 3: 80.201.237.102 26.759ms | 19:27 |
| mandel | 4: 194.78.0.140 24.435ms | 19:27 |
| mandel | 5: 80.84.21.94 25.027ms asymm 8 | 19:27 |
| mandel | 6: 94.102.162.126 24.990ms asymm 7 | 19:27 |
| mandel | 7: 94.102.162.18 35.302ms | 19:27 |
| mandel | 8: 212.72.42.5 48.526ms asymm 9 | 19:27 |
| mandel | 9: 4.68.23.190 35.807ms | 19:27 |
| mandel | 10: 4.69.140.21 35.701ms | 19:27 |
| mandel | 11: 4.69.137.54 112.785ms asymm 10 | 19:27 |
| mandel | 12: 4.69.134.154 115.934ms asymm 11 | 19:27 |
| mandel | 13: 4.68.17.80 111.264ms | 19:28 |
| mandel | 14: 4.79.20.22 115.864ms asymm 13 | 19:28 |
| mandel | 15: 72.21.199.32 114.917ms asymm 14 16: 72.21.222.139 113.140ms asymm 15 | 19:28 |
| mandel | buahaha the Spaniard can :D | 19:28 |
| * aquarius waits for mandel to get autokicked for flooding ;) | 19:28 | |
| mandel | :P | 19:28 |
| aquarius | bah, it works for you? | 19:28 |
| rye | mandel, wow | 19:28 |
| aquarius | bet he's got "127.0.0.1 files.one.ubuntu.com" in /etc/hosts or something :) | 19:29 |
| rye | 174.129.230.205 is still not reached :) | 19:29 |
| mandel | aquarius, hehe trust me I dont play much for ubuntu so for me is a mystery :P | 19:30 |
| rye | mandel, ok, you've most likely got a DSL / PPPoE connection, right? | 19:30 |
| mandel | rye, yes | 19:38 |
| dobey | uhm | 19:56 |
| dobey | mandel: that last IP you pasted isn't files.one.ubuntu.com though :) | 19:56 |
| dobey | in fact, it looks like they're probably the same host where the route drops off for me | 19:57 |
| mandel | dobey, I've got no idea, to be honest I've never look much at ubuntu one, just did what rye asked for ;) | 19:58 |
| dobey | sure | 19:58 |
| rye | if anyone wonders what I am trying to do - question https://answers.edge.launchpad.net/ubuntuone-client/+question/100387 - looks like the guy has firewalled himself from anything ICMP, that's why uploads do not work | 19:59 |
| dobey | well i can download files | 20:02 |
| dobey | even though tracepath doesn't finish :) | 20:02 |
| dobey | and i uploaded a file | 20:03 |
| dobey | so i doubt that's the issue | 20:03 |
| rye | dobey, tracepath was needed to find out mtu value. Though it won't help much, since the actual tracepath will need to be done from the server itself | 20:03 |
| dobey | but weird | 20:03 |
| rye | dobey, the guy has blocked all incoming ICMP requests, therefore, when his router says "hey, i can't use 1500 as mtu, use 1460 / 1488 (as mandel has)", his machine does not receive that and no retransmit happens. | 20:04 |
| rye | this is still a theory, but his last comment clearly states that tracepath could not even get anything reasonable | 20:05 |
| * rye had a nice talk with his ISP a month ago or so about MTU and all around that :) | 20:06 | |
| dobey | oh i'm not saying he didn't disable ICMP requests | 20:06 |
| dobey | but i would suspect that a lot more internet would be failing if blocking ICMP requests made everything else not work | 20:06 |
| rye | dobey, btw, files.one.ubuntu.com seems to be a stealth server :) | 20:06 |
| rye | dobey, nope, only uploads :) | 20:07 |
| dobey | rye: lots of web servers block ICMP | 20:07 |
| dobey | so you can't ping flood them | 20:07 |
| dobey | but then again, they probably only block some requests | 20:07 |
| rye | dobey, the downloads might be handled by his ISP, the server sends response, ISP says "he's got 1460 only, retransmit" (ICMP 4:3), server retransmits and download happens | 20:07 |
| dobey | since hopefully professional sysadmins are smart enough to do it right | 20:08 |
| rye | dobey, yes, this is a major problem with a lot of internet sites, sysadmins tend to block everything, even useful ICMPs :) | 20:08 |
| rye | erm | 20:25 |
| rye | hover the mouse over "Sharing", "Upload a File", "New Folder" items in web ui - look at the tooltip ("title") that appears. | 20:25 |
| Apache_ | hi | 20:48 |
| dobey | hi Apache_ | 20:58 |
| Apache_ | dobey: hi, I got some questions re desktop-couch and replication not happening.... is this the right place? | 21:00 |
| dobey | Apache_: yes, you are likely experiencing an odd issue we just discovered yesterday :) | 21:03 |
| Apache_ | well I've only been messing around with it for a day or so nad followed the example on http://arstechnica.com/open-source/guides/2009/12/code-tutorial-make-your-application-sync-with-ubuntu-one.ars/1 | 21:05 |
| Apache_ | but it never replicates to my other machine | 21:05 |
| Apache_ | I also checked on Freedesktop.org at the trouble shooting section, but I don't have a replication.log | 21:06 |
| dobey | Apache_: do you have couchdb running, but without desktopcouch-service also running? | 21:06 |
| Apache_ | in fact no /log/ directory either | 21:06 |
| Apache_ | not sure, my code successfully created a test database locally | 21:07 |
| Apache_ | and ubuntu-one file syncing is working fine | 21:08 |
| rye | Apache_, try restarting desktopcouch, /usr/lib/desktopcouch/desktopcouch-stop and then start it with /usr/lib/desktopcouch/desktopcouch-service | 21:11 |
| Apache_ | is there a way for me to check to see if the desktopcouc-service is running | 21:11 |
| rye | Apache_, bug 519028 | 21:11 |
| ubottu | Launchpad bug 519028 in desktopcouch "desktopcouch replication and org.desktopcouch.CouchDB.service are NOT started on desktopcouch.records calls" [High,Confirmed] https://launchpad.net/bugs/519028 | 21:11 |
| dobey | run "ps afx|grep couch" in a terminal | 21:11 |
| rye | ps aux| grep [d]esktopcouch-service | 21:12 |
| Apache_ | when I restarted desktopcouch-service it popped up a window asking for application access to keyring, don't remember seeing that before | 21:12 |
| rye | Apache_, eeexactly! | 21:13 |
| Apache_ | ahhh | 21:13 |
| Apache_ | :) | 21:13 |
| rye | I was promised that this is going to be fixed in next stable release update of desktopcouch, this issue has been hanging around for quite a long time w/o being noticed :( | 21:14 |
| Apache_ | reading the log now | 21:16 |
| Apache_ | is this normal :) | 21:16 |
| Apache_ | 2010-02-09 21:13:09,881 WARNING Can't reach service ubuntuone. No JSON object could be decoded | 21:16 |
| rye | Apache_, hm, is there anything else in the log? Do not post the contents of the log that say about oauth, this is confidential info you do not want to share :) | 21:21 |
| Apache_ | the bit immediately before that is | 21:23 |
| Apache_ | 2010-02-09 21:13:09,881 WARNING Can't reach service ubuntuone. No JSON object could be decoded | 21:23 |
| Apache_ | 2010-02-09 21:13:09,661 INFO No o.u.c key. Maybe there's uo.c key? | 21:23 |
| Apache_ | 2010-02-09 21:13:09,660 DEBUG Looking up prefix for service 'ubuntuone' | 21:23 |
| rye | Apache_, and that's it? I mean you don't have any replication actually going, i.e. no message for that, right? | 21:27 |
| Apache_ | 2010-02-09 21:13:00,172 DEBUG found 0 self_identity records | 21:29 |
| Apache_ | 2010-02-09 21:13:00,211 DEBUG set new self-identity value: 'c1863e49-10ee-4472-be98-87ec7d1111d8' | 21:29 |
| Apache_ | 2010-02-09 21:13:00,242 INFO starting advertising _couchdb_location._tcp on port 35017 | 21:29 |
| Apache_ | 2010-02-09 21:13:00,259 DEBUG started replicating | 21:29 |
| Apache_ | 2010-02-09 21:13:00,308 DEBUG replication of discovered hosts finished | 21:29 |
| Apache_ | 2010-02-09 21:13:00,342 DEBUG static pairings are [('8e2f36f5-bbed-477e-ba01-86274d51dfa4', 'ubuntuone', True, True)] | 21:29 |
| Apache_ | 2010-02-09 21:13:00,342 DEBUG Looking up prefix for service 'ubuntuone' | 21:29 |
| Apache_ | 2010-02-09 21:13:09,660 DEBUG Looking up prefix for service 'ubuntuone' | 21:29 |
| Apache_ | 2010-02-09 21:13:09,661 INFO No o.u.c key. Maybe there's uo.c key? | 21:29 |
| Apache_ | 2010-02-09 21:13:09,881 WARNING Can't reach service ubuntuone. No JSON object could be decoded | 21:29 |
| Apache_ | 2010-02-09 21:13:09,882 DEBUG finished replicating | 21:29 |
| Apache_ | 2010-02-09 21:15:01,859 INFO ending advertising _couchdb_location._tcp on port 35017 | 21:29 |
| Apache_ | thats the log file | 21:29 |
| rye | not good | 21:34 |
| rye | Apache_, can you access futon? CouchDB web interface? | 21:35 |
| rye | Apache_, the link to the html was given on desktopcouch startup | 21:35 |
| Apache_ | yeah, can see my test db that was created from the code and the 2 records | 21:35 |
| rye | is there 'management' database? | 21:36 |
| Apache_ | yep | 21:36 |
| rye | Apache_, what does /_utils/database.html?management/_design/ubuntu_one_pair_record/_view/ubuntu_one_pair_record return | 21:38 |
| rye | Apache_, by the way, can you please start seahorse and see whether there is a record for UbuntuOne token for https://ubuntuone.com ? | 21:41 |
| Apache_ | brings back a row | 21:41 |
| Apache_ | rye: seahorse, ok you got me there | 21:41 |
| rye | Apache_, huh? | 21:42 |
| Apache_ | rye: when you say start seahorse, whats seahorse? | 21:42 |
| rye | Apache_, ah, this is "Applications/Accessories/Passwords and Encryption Keys" | 21:43 |
| Apache_ | rye: yeah there is a record for https://ubuntuone.com | 21:44 |
| rye | Apache_, is desktopcouch service allowed to access that key? Doubleclick the key entry and switch to Applications tab | 21:46 |
| rye | hm, no, this does not make sense, it says "No o.u.c key. Maybe there's uo.c key?" only when keyring does not present it with a key | 21:48 |
| Apache_ | checked the applications tab, and desktopcouch serveice has Read, Write and Delete all checked | 21:49 |
| Apache_ | there are two other keys in seahorse, both the same description "Desktop Couch user authentication" | 21:52 |
| Apache_ | first one has "desktopcouch servce" and "ubuntuone-login" as applications and second one only has "ubuntu-one" | 21:54 |
| Apache_ | if that helps any? | 21:54 |
| rye | Apache_, this is ok, that is required by applications to know the credentials for couchdb acces | 21:54 |
| Apache_ | ok | 21:55 |
| rye | Apache_, ok, lets do it the dumb way, could you please shut down and start desktopcouch again? | 21:55 |
| Apache_ | ok | 21:56 |
| rye | Apache_, and, what is the name of the keyring the key got put into? was that login or default? | 21:56 |
| Apache_ | got this Browse your desktop CouchDB at file:///home/simon/.local/share/desktop-couch/couchdb.html and then nothing | 21:57 |
| rye | Apache_, anything new in replication log? | 21:58 |
| Apache_ | should I be getting my shell prompt back.... not getting anything at the moment | 21:58 |
| rye | Apache_, no, it will grab the terminal until you kill it | 21:59 |
| Apache_ | hold on, loads more in the log file now.... gonna go and check my other machine | 22:01 |
| rye | weird | 22:01 |
| rye | so first it starts, finds no tokens in the keyring, then the token magically appears and it finds it on next restart? This looks weird | 22:01 |
| Apache_ | do I have run desktopcouch in a terminal window like this always? | 22:01 |
| Apache_ | or should it normally start up on boot | 22:02 |
| rye | Apache_, the applet starts desktopcouch, but it starts it via CouchDatabase('name') call. desktopcouch starts couchdb directly, w/o starting the service that will take care of replication. | 22:03 |
| Apache_ | or should it normally start up on boot | 22:04 |
| Apache_ | ok, so nothing has appeared on the other machine yet | 22:05 |
| rye | Apache_, is other machine running desktopcouch-service? | 22:06 |
| rye | Apache_, http://paste.ubuntu.com/372764/ | 22:06 |
| Apache_ | ahh, let me check :) | 22:06 |
| rye | Apache_, this will start desktopcouch-service via dbus call, if it is not started already | 22:07 |
| Apache_ | yes! | 22:09 |
| Apache_ | so database has replicated to the other machine | 22:09 |
| Apache_ | rye: that script you pointed me at, I shoud add that to my python script? | 22:11 |
| rye | Apache_, actually, I'd add it to the startup applications, so that couchdb is always properly started | 22:12 |
| Apache_ | ok, cheers, I will do that... really appreciate all your help this evening, gotta run.... early start, but if I can do anything to help | 22:14 |
| rye | It won't hurt but it will be eventually fixed in desktopcouch itself | 22:14 |
| Apache_ | thanks again rye! | 22:15 |
| rye | Apache_, you are very welcome! | 22:15 |
| Apache_ | bye | 22:15 |
| rye | mission accomplished | 22:15 |
| rye | no | 22:18 |
| rye | 2010-02-09 21:13:09,661 INFO No o.u.c key. Maybe there's uo.c key? | 22:18 |
| rye | this is weird, added to TODO for tomorrow | 22:18 |
| fagan | aquarius: im having trouble viewing my desktopcouch records in lucid | 22:27 |
| fagan | its asking me for a username and password to access the page | 22:27 |
| aquarius | fagan, OK. We've fixed that bug in trunk, but trunk isn't released into lucid yet. See the troubleshooting page in the documentation for how to work arounf it | 22:28 |
| fagan | aquarius: link? | 22:28 |
| aquarius | http://www.freedesktop.org/wiki/Specifications/desktopcouch/Documentation | 22:29 |
| fagan | thanks aquarius | 22:29 |
| aquarius | which is the documentation index, which you may find both useful and interesting | 22:29 |
| aquarius | and then try "troubleshooting" from there. | 22:29 |
| rye | bye everyone, it is night again here and looks like I need to log out :) | 22:43 |
| fagan | aquarius: restarting desktopcouch worked | 22:43 |
| fagan | Are we going to have any empathy contact syncing in lucid? | 22:43 |
| aquarius | fagan, cool. As I say, we've worked out what the problem is, and CardinalFang fixed it, but that fix hasn't made it into Lucid yet. | 22:43 |
| aquarius | fagan, we're not working on Empathy contact sync for Lucid. If someone else does that'd be great, though! | 22:44 |
| fagan | aquarius: give me 15 mins ill get something working :) | 22:44 |
| dobey | well empathy should poke the address book, so syncing the contacts should work that way. and pretty much all the protocol servers store info on the server | 22:56 |
| * dobey really goes aaway now | 22:57 | |
| === jamalta is now known as jamalta-afk | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!