[01:52] is there anyone here thats using FTP to upload to launchpad since it was "fixed" [01:52] i've been using sftp to upload to lp fine [01:53] havent tested using ftp though [01:53] but afaict it works [01:53] i can test that for ya if you want [01:55] well I used to use ftp then it failed on last byte with large files tehen was told that ftp had some issues told to use sftp which is slow and slower and now apparently ftp has been fixed but I can see any fifference [01:56] i have smaller files, but if sftp is broken i'll test [01:56] s/sftp/ftp/ [01:56] something like a nvidia driver will never upload via ftp if its 93231KB it uploads 93230 and hangs [01:57] smaller files dont seems to have that problem say less that 2 meg [01:57] well i just pushed a package that i built in another ppa [01:57] and it has >= 3MB of data [01:57] unless i pushed the wrong package [01:57] thats too small [01:57] regardless, my personal recommendation is use sftp [01:57] because its more secure [01:57] is slowwwwwww [01:57] because its more secure... :P [01:58] meh Im not uploading governement secerts [01:58] :P [02:01] I guess I should be grateful something works :/ [02:01] *shrugs* [02:02] do you have experience packaging stuff from source? [02:03] some [02:03] why [02:04] and define "packaging from source" [02:04] because if you mean package the source, and then build it and install it, well... [02:04] that's a different story [02:04] but i have the general idea down (havent ever had a successful test yet) [02:05] i mean debianizing new packages [02:08] i mainly looking for a partner in crime to do a modify a minimla install to make it leaner than a regular minimla install [02:08] you did check that it was okay to do that first, right? :P [02:09] because usually if you arent affiliated with the project... you shouldnt go modifying installation packages much... [02:09] yes its fine [02:09] i've been yelled at for that... :P [02:09] its GPL they cant yell [02:09] oh that rhymes [02:10] heh [02:10] whoever yellled at you was wrong [02:10] actually [02:10] they were correct [02:10] their software was copywritten [02:10] wy its GPL isnt it [02:10] and i modified their installation scripts [02:10] without permission [02:10] ah well [02:10] (this was when i was newbie) [02:10] although since then... [02:11] i proposed the changes to them correctly, and they were like "Wow, that's not a bad idea" [02:11] [02:11] yes, that's not a word. [02:11] well [02:12] its GPL so thers no issue with that except its beyond my skills [03:44] why does this show last rev as 660 when original repo has last rev as 1183? https://code.launchpad.net/~renatosilva/purple-plugin-pack/trunk [03:45] what is showned there is only internal tomlauchpad [03:45] the actual revision is right [03:46] No, it's just that bzr and hg and svn revnos have different meanings. [03:47] and thats the same I said only more confusing [03:47] what Iif I merge from it? should I say "merge from rev 1183" or 660? [03:47] youll get 1183 [03:47] X3lectric: It's nothing to do with Launchpad. [03:47] It's a Bazaar thing. [03:47] potatoe [03:48] * RenatoSilva confused [03:48] ok where I said lauchpad replace with bazaar [03:48] should I say merge from 660, 1183 or 5bb4c56f954f? [03:52] i dont speak bazaar but hwta bazaar I know does is crete its own revision even though if you were to create a recipy or compile you actually get what revison is from the real repo [03:54] X3lectric: So, say the svn repository has revision number 68. Someone then requests an import of that, so Launchpad creates a Bazaar branch, imports the code from revision 68 and commits it. Now the SVN repo is revision 68, and the Bazaar repo is revision 1. [03:55] something like that [03:55] no [03:55] otherwise it would not be 660 but 1 [03:55] Bazaar revnos are per-branch; Subversion are per-the-big-svn-repository. [03:56] its a eample [03:56] RenatoSilva: It was an example, I wasn't using the real numbers. [03:56] * RenatoSilva more confused [03:56] its simple [03:56] sorry, should I say merge from 660, 1183 or 5bb4c56f954f? [03:57] X3lectric: should I say merge from 660, 1183 or 5bb4c56f954f? [03:57] if 660, how to map to 1183 [03:57] how to map them all [03:57] if your working with bazaar use 66- it should align ok with real revision [03:57] I thought it would just match the revisions [03:57] X3lectric: '66-'???? [03:58] typo [03:58] oh [03:58] X3lectric: what if I want to find in the original repo, what 660 is? [03:58] http://bazaar.launchpad.net/~renatosilva/purple-plugin-pack/trunk/files [03:59] http://bazaar.launchpad.net/~renatosilva/purple-plugin-pack/trunk/revision/660?start_revid=660 [03:59] essentially its [03:59] X3lectric: ctrl+f 1183 => no result [03:59] 1 [03:59] 1 [03:59] EXTRA_DIST = \ [03:59] 2 [03:59] 2 [03:59] glib_compat.h \ [03:59] 3 [03:59] 3 [03:59] gtk_template.c \ [03:59] 4 [03:59] libjson-glib-1.0.dll \ [03:59] 5 [03:59] 4 [03:59] purple_template.c \ [04:00] ??? [04:00] !pastebin | X3lectric [04:00] X3lectric: For posting multi-line texts into the channel, please use http://paste.ubuntu.com | To post !screenshots use http://imagebin.org/?page=add | !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic. [04:00] thx micah [04:00] maybe the 12 first digits of rev ID, then lookup in original repo the 1183? [04:01] how to find 1183 from 660? [04:01] just use 660 [04:02] X3lectric: how to find 1183 from 660? [04:02] petend 660 is actually saying 1183 [04:02] *pretend [04:02] X3lectric: how to find 1183 from 660? [04:02] X3lectric: how to find Y from X? [04:02] X3lectric: x is bzr, y is hg [04:02] there is no X [04:03] X3lectric: x = 660 at the moment [04:03] oh lord [04:03] X3lectric: y = 1183 [04:04] X3lectric: HOW TO FIND the correspondent original revision in hg repo BASED ON the revision shown in LP? EXAMPLE: 660 => 1183 [04:04] which part of bazar doesnt use same revison tagging dont you get [04:04] X3lectric: I have noted that 660's rev ID contains the original hg rev ID, but it's an annoying hack [04:04] what bazar shows as 660 is ACTUALLY 1183 [04:04] my God! [04:05] Meu Deus [04:05] PQP [04:05] whats PQP [04:05] X3lectric: I have a dirty solution for my problem, unfortunately for you [04:06] X3lectric: what's Meu Deus [04:06] MEu Deus= My God in Portuguese [04:06] X3lectric: good for you [04:06] either me or X3lectric is a kid [04:07] i presume your POrtuguese with a name like that [04:07] anyone else, is that hack the only thing I have? [04:08] https://www.guifications.org/projects/purple-plugin-pack/repository/revisions [04:09] RenatoSilva: So you're whole problem is that the two repositories which both use entirely different version control tools have different revision numbers? [04:09] s/you're/your/ [04:10] Mercurial revision numbers are not meant to be used for identification. They are entirely local to a repository; unlike bzr, even if the history is identical between two repositories the revision numbers can differ. [04:10] You should use a Mercurial changeset ID. Which is the hash-like thing. [04:11] "Revision numbers referring to changesets are very likely to be different in another copy of a repository. Do not use them to talk about changesets with other people. Use the changeset ID instead." [04:11] wgrant: so I should never bzr commit -m "merge from hg repo xyz, revision 123"? I should rather use hg rev IDs like here? http://bazaar.launchpad.net/~renatosilva/purple-plugin-pack/irchelper/revision/2 [04:12] http://wiki.bazaar.canonical.com/BzrRevisionSpec [04:13] With bzr-svn you can say -rsvn:1234. I'm not sure if bzr-hg has something similar for changeset IDs. [04:13] wgrant: "Upstream revision 87484177eba0" [04:13] http://stackoverflow.com/questions/5413602/monotonically-increasing-bazaar-trunk-revision-numbers [04:15] wgrant: the problem is that in the web interface for the original repo, I can't easily find the changeset by typing 87484177eba0 in a searchbox for example (seems a bug since the form has a changeset checkbox). That is, I'm stuck with commit comments and page browsing :( [04:16] wgrant: ftp uploads for large file like e.g nvidia drivers still fail at 1 byte form complete filesize [04:17] wgrant: say upload is 97600KB it uploads 97599KB and hangs stops uploading [04:17] RenatoSilva: Works for me. [04:17] with small files its ok [04:17] X3lectric: Ubuntu developers frequently upload files several times larger without trouble. [04:18] well it doesnt work here I thought it was the ppa upload path but its not [04:20] and I cant get anyone that can do anything about it to pay attentiaon [04:20] sftp is too slow [04:21] It works fine for just about everyone else. [04:21] Both FTP and SFTP. [04:21] sftp works fine but its slowwww [04:21] not everyone has 10mbit upload speeds [04:22] Huh? [04:22] SFTP should not increase the transfer size significantly. [04:22] wgrant: this works for you??? https://www.guifications.org/search/index/purple-plugin-pack?changesets=1&q=87484177eba0 [04:23] here in UK the max uplaod speed is 1.5MB/s if you have a 100MB/s download speed [04:23] RenatoSilva: I used the real hg web UI. http://hg.guifications.org/purple-plugin-pack/log?rev=87484177eba0 [04:23] X3lectric: Yes, and I have 256Kbps upstream. [04:24] X3lectric: SFTP does not increase transfer size by any significant volume. [04:24] wgrant: yeah that works but I think they're about to disable it :( [04:24] That would be rather silly. [04:25] SFTP addstoomuch overhead with security its 10x slower than FTP [04:25] No, it's not. [04:25] Unless you are using a Z80 or something. [04:26] wgrant: yeah would be silly since the new one doesn't work [04:26] RenatoSilva: Exactly. [04:26] wgrant: but thanks for helping [04:27] wgrant: so your telling me that SFTP is not slower than FTP [04:27] X3lectric: Not by 10x. [04:27] and the main overhead is CPU, not bandwidth. [04:27] wll I dont wnat to be pedantic [04:28] lol CPU is not a problem [04:28] Then you have no problem. [04:28] yes I do [04:28] even with failures [04:29] the files that fail take 10 minutes to uplaod befor it hangs [04:29] with sftp it twice the time sometimes 30 minutes [04:29] So SFTP has connection issues too? [04:29] Have you considered that your connection may be problematic? [04:29] Nobody else has reported problems like this. [04:30] Im not nobody else [04:30] Pardon? [04:30] jst because nobody else reports it doesnt mean its not true [04:31] No, but given that it works for thousands of other people, it sounds like a good idea to not rule out your connection as an issue. [04:31] I dont enjoy comming here trying to get to the bottom of this and be told that I have to use SFTP because nobody else has reported any issue [04:33] Since Im CCNA and CCNIE qualified Im gonna pretend I didnt hear that [04:33] .... CCNIE? [04:33] I'm gonna pretend I didn't hear that. [04:35] For the few other people that have had issues with FTP, SFTP has worked fine. This is unsurprising, given that SFTP is a slightly less insane protocol. [04:35] I have lots of evidence that Launchpad's FTP and SFTP work well for a lot of people. [04:35] I have no evidence that your connection works. [04:36] So I'm going to suspect your connection first, particularly if SFTP doesn't work either. [04:36] ok SFTP works but ITS slow [04:36] I never said sftp didnt work ok [04:36] It's also suspect that you claim to hold a qualification without knowing its name, but I admit that may have been a typo. [04:36] it is a typo [04:37] considering I havent slept for 2 days [04:37] so [04:38] Im saying that SFTP work but its much slower, FTP on the other hand fails to upload at last byte of a large transfer [04:39] Have you tried from an entirely independent Internet connection? [04:39] yes [04:40] And the behaviour is identical? [04:41] like twins [04:44] I get the expected 25KB/s over SFTP here. Let me try from a faster connection. [04:44] sftp uploads every last byte and doesnt fail, despite it being slower by a significant factorthan the FTP which is significantly faster but only works with smaller files here [04:46] Hmm, only 4MB/s. [04:46] Slower than it should be, but possibly network congestion. [04:46] But still, not "slow". [04:46] That was uploading a 100MB file to ppa.launchpad.net with SFTP. [04:47] So it can handle 32Mbps with no trouble, and if that's slow for you then, well, I am quite envious. [04:49] Your server may be able to handle 32Mbps upload with SFTP but I dont have thatsort of speed available to me [04:50] And I'm able to upload at 200Kbps from home, despite my upstream being 256Kbps. [04:50] So the overhead is minimal (considering I'm also downloading stuff at a few Mbps, so there's a number of acks there) [04:50] Not sure what possessed Australian ISPs to over 20000/256 connections. [04:50] s/over/offer/ [04:51] Yes, Australian ISPs _are_ possessed. [04:52] thanks all [04:52] My upload is 512 from home but when Im uplaoding to launchpad via sftp it maxes out at variable 80 to 90kbps [04:53] Unless we are dealing with a terrible TCP windowing issue or similar, that doesn't really make sense :( [04:53] well UK ISP are insufrerable too considering there is only one "fiber-coax" provider in the whole country [04:53] for e.g [04:54] We have two HFC providers. [04:54] But they only cover small portions of the country. [04:55] my home connection is 10Mbps down down and 512Kbps uplaod while in constarst they offer a 100MBps download and a measly 2.5Mbps upload [04:56] the price for that 100Mbit connection is daylight robbery [04:56] evn 30 Mbit is impossibly expensive [04:58] well anyway the only way for you to test this failure over FTP is to use same files I use which is nvidia-current drivers to my ppa [04:59] ...Have you tried uploading with a different FTP client? [04:59] because I dont use/upload any other files that are large enough to trigger the hang at last byte [04:59] ? [04:59] ? what? [04:59] peng this is using ubuntu dput [05:00] Oh. I shouldn't admit this in #launchpad, but I don't know the Debuntu tools. :P [05:01] the rpoblem is using FTP via.dput.cf [05:03] Are those files available somewhere? [05:04] what bugs me is that when I starting uploading to ppa's using FTP all was well one day thsi very same error ocurred, I was told that Launchpad was replacinf its FTP servers, now 6 months later nothing has changed form the error which hangs tranfer at last byte of file [05:04] what files? [05:05] the ones that fail? [05:05] acually saying they fail is incorrect the trasmission hangs at last byte [05:06] the file sthta hang via FTP are nvidia-graphics-drivers and can be found in https://launchpad.net/~x3lectric/+archive/nvidia-vdpau [05:08] Let's see. [05:11] this is my dput http://pastebin.com/tg3KUFij [05:12] Uploaded fine over FTP from the 4MB/s connection. [05:13] ok can you loook at my dput and tell me whats wrong? [05:13] sftp dput that never fails http://pastebin.com/mL5Qng6Q [05:13] The default one doesn't have passive_ftp = 1 [05:14] ftp dput that fails dput http://pastebin.com/tg3KUFij [05:14] mmmm [05:14] It's clearly using passive anyway, but perhaps forcing it does something bad. [05:15] ok Ill remove it clearly oversight [05:16] Added that, still works fine. [05:16] mmmm [05:17] wait you just upload nvidia drivers twice in 3 minutes [05:17] via FTP [05:18] Not from my home connection. [05:19] would the OS be a factor? [05:19] I really hope not. But all my tests have been from Ubuntu 10.04 LTS and Ubuntu 11.04. [05:20] well my dev box is still karmic [05:20] Erm. [05:20] Karmic is unsupported. [05:20] You probably want to upgrade :) [05:21] I havent the patiente [05:21] You also lack any kind of security. [05:22] how so [05:22] Karmic no longer receives security updates. [05:24] right [05:25] wait [05:26] I agree that karmic may not be ideal security wise but if your going to compromise karmic the extra seconds to compromise any other OS is trivial [05:27] It's far easier to compromise something if it's not got known vulnerabilities patched. [05:28] its trivial either way nothing is secure 100% it may take a few extra minutes to find one [05:33] i guess Ill have to test this with another OS and create a new development environment [05:34] to ruleout the FTP issue not being os associated [05:35] anyway thx for testing [05:35] Hmm. === Mkaysi_ is now known as Mkaysi === nyuszika7h is now known as I_had_gun === I_had_gun is now known as nyuszika7h [20:44] can someone delete a bug for me? [20:46] there is no delete facility for bugs. They can be orphaned - made private with no subscribers - but thats a bug that we will be fixing to make sure there is always someone from the project that can access the bug. [20:46] whats up ? [20:55] I reported the bug upstream, the patch is not simple, no one is interested in LP. It's just junk [21:12] RenatoSilva: what bug? [21:14] wgrant: ping re powerpc buildd [21:16] RenatoSilva: That's hardly grounds to delete it [21:17] lifeless: bug 682780 [21:17] Launchpad bug 682780 in purple-plugin-pack (Ubuntu) "Password dialog for IRC Helper" [Undecided,New] https://launchpad.net/bugs/682780 [21:18] RenatoSilva: you should talk to debfx about that in #ubuntu-motu [21:19] RenatoSilva: its entirely appropriate to have a bug open in Ubuntu while the defect is still present in Ubuntu [21:19] I reported the bug upstream -- there is the right place, the patch is not simple -- unlikely any package maintainer want to manage it, no one is interested in LP -- since months, no one really. [21:19] lifeless: technically it's not a defect, it's a feature [21:20] *if* anyone get ever interested, the report will be in upstream [21:21] RenatoSilva: the line between defect and feature is different to different people. [21:21] RenatoSilva: not using a keyrign is a defect [21:26] :'( [21:39] hoices in gui are kepunder add vdeisource or add source [21:50] nite nite im gonna see if I can close my eyes for an hour or two and try to relax and meditate [21:54] Anyone else manage to parse that -2 comment ? === yofel_ is now known as yofel [23:11] thanks all anyway === med_out is now known as medberry [23:49] hi [23:50] i try to edit description of a ppa and get an error like this: No REFERER Header [23:50] but it worked be4 [23:51] dont matter which paa i try hm [23:51] im using icecat5 [23:51] *ppa [23:55] DNS777: Your browser is configured to not send a Referer header.