[00:07] how does the power button work for servers? does it work thro systemd/logind.conf ? or thro acpid ? [00:08] the powerbtn.sh seems to check for logind and exit when acpid is used [00:08] and uncommenting powerkey in logind.conf does not help either === thomi_ is now known as thomi === Ursinha-afk is now known as Ursinha [00:44] Hi I have installed ubuntu 14.04 on my system. I would like anytime dhcp server starts on my system remove with new client lease file so it can try to get new IP address. Is this possible? thx [01:10] hi guys any idea on my ubuntu server - i went to cd /mnt and i created a directory im already as root user but i get this ----> [root@elastix mnt]# mkdir testcalls --- mkdir: cannot create directory `testcalls': No such file or directory [01:13] ruben23: try 'cd `pwd`' then re-try your mkdir [01:13] ruben23: that can happen if you deleted the /mnt directory out from underneath your process [01:15] sarnold: what i will do is cd..? and pwd..? [01:16] ruben23: or just cd /mnt :) your choice [01:16] [root@elastix mnt]# cd /mnt [01:16] [root@elastix mnt]# mkdir testcalls [01:16] mkdir: cannot create directory `testcalls': No such file or directory [01:17] yikes. no idea! sorry [01:31] hi I want to remove a debian package using apt-get remove command but I do not want the package dependencies to be removed, Is this possible? thx [01:43] zartoosh: use dpkg --remove [01:57] rbasak, thx [02:02] This may seem like a basic question however when I scour the internet for the answer there isn't anything straight forward I need to figure out how to make a group in apache === mfisch` is now known as mfisch === mfisch is now known as Guest32427 === CripperZ- is now known as cripperz [05:11] Kawaiola: what as a 'group in apache'? [05:39] hello can someone help me with this issue? http://paste.ubuntu.com/7812741/ [05:39] I'm running ubuntu-server 14.04 [05:41] tortib: you don't want to do that; it or more less destroys your system :( https://bugs.launchpad.net/ubuntu/+source/dbus/+bug/1072518 [05:41] Launchpad bug 1072518 in ifupdown "Restart networking crashes dbus and the desktop manager" [Critical,Fix released] [05:41] sarnold, oh? [05:41] but I added some IP addresses to the /etc/network/interfaces config file and I want to configure them on the interface now...what's the 'proper' way of doing that on ubuntu? [05:41] tortib: yeah. without that new guard in place, it made for some extremely upset bug reports. (One from me, haha) [05:42] tortib: 'ifup ifacename' [05:42] oh [05:42] okay [05:42] tortib: or maybe 'ifup -a' [05:42] That didn't do it. [05:43] hrm, can you pastebin your interfaces file? [05:44] I would rather not [05:44] fairenough [05:45] you could strace -f -o /tmp/networking ifup iface and look for errors in the strace [05:46] ifup: interface eth0 already configured [05:46] ifdown eth0; ifup eth0 ? [06:01] sarnold, so I guess I had the config wrong :) [06:02] how do I add an address? I have iface eth0 inet6 static and then below i have address 2400:6180:0000:0000:0000:aaaa:bbbb then below that I have netmask 64 [06:04] sarnold, I have that set in /etc/network/interfaces [06:04] :x [06:06] tortib: hrm, the manpage says "Address (colon delimited/netmask) required" -- maybe try 2400:6180:0000:0000:0000:aaaa:bbbb/64 ? [06:06] I already have a static ip address set in there and it's the format I described [06:06] gateway is also specified that way [06:08] tortib: what's "2600"? I don't see that class on the wikipedia ipv6 address page [06:08] (I haven't fought the ipv6 fight yet myself..) [06:08] 2400 you mean? [06:08] It's digital oceans IPs [06:10] sigh, yeah, that's what I meant. hrm, I'm surprised it looked so unusual.. [06:20] sarnold, I'm trying to add this to my /etc/network/interfaces config [06:20] http://paste.ubuntu.com/7812857/ [06:21] it's a tunnel for hurricane electric [06:21] but everytime I try ifup eth0 it says the interface is already configured [06:22] tortib: how about 'ifup he-ipv6'? [06:22] tortib: is the 'ip' command available? [06:22] yes [06:22] oh it's working now [06:23] woot! nice :) [06:24] * tortib does a dance [06:25] but i'm still not able to use ifup eth0 [06:25] so I can't assign any ipv6 ips from my /48 to the interface to route traffic :( [06:27] tortib: you need to bring it down bevfore you can bring it up; did ifdown eth0 ; ifup eth0 work? [06:28] sec [06:29] oh jeeeze, I just figured out 'tortib' :) must be bed time :) [06:32] well [06:32] i can't add any ips to the config [06:32] it doesn't start the interface for some reason [06:33] tortib: drat. maybe you're only allowed one addr family per interface in that file. if you want to add more IPs to an interface, you have to use the 'up' or 'post-up' commands to run ip addr add ... dev ... manually [06:33] tortib: maybe you have t o do the same if you want both ipv6 and ipv4 addresses on an interface :/ [06:35] http://askubuntu.com/questions/313877/how-do-i-add-an-additional-ip-address-to-etc-network-interfaces [06:36] funny; the foo:0 'aliased' interfaces haven't been necessary for over a decade but the interfaces file never adapted.. [06:37] yeah [06:37] either way I guess, whatever you like more :) [06:38] bedtime ;) good luck tortib [06:38] i'm just going to use if.up and create a script to add the IPs to the server that way [06:38] thanks sarnold nice talking to you [06:38] nice talkin to you too :) have fun === cripperz is now known as N0DE` [07:29] I'm looking for an Ubuntu 14.04 Rescue image, that I can boot from via PXE/Netboot. Or something similar. I'm using PXE to install OS, but I also want to have a simple OS that I can create and startup my self. === disposab1e is now known as disposable [08:37] messing around with preeseeding however it's hang up - any way to see what's frong [08:46] pds: hit ctrl+alt+f# until you find the one with the log [08:46] hmmm you don't have way to install software in /bin/sh [08:46] or look around for the log file in a termina [08:48] try to run the following in a postinst [08:48] sudo apt-get install apt-add-repository [08:48] sudo apt-add-repository ppa:rquillo/ansible [08:48] sudo apt-get update [08:48] sudo apt-get install ansible [08:48] [08:49] don't use sudo [08:49] and you should use the in-target thing [08:50] and are you aware that it is one long logical line, so you have to use \ at the end of the line, and ; between commands? [08:50] and you should test your commands manually too ... I bet the repo one will ask interactively about accepting a key or some such [08:51] any workaround ? [08:52] hmmm will mess a bit around :) [08:52] like always :) [08:52] pds: Does the script run as root? [08:52] a postinst so probably the user that has be created during the install [08:53] preseed stuff runs as root [08:53] it is running in the installer environment [08:54] Ah, yes, then drop the sudo. It will probably break on that asking for a password or something. [08:55] preseeding = lots and lots of testing :) [11:55] HI I have a server which gets its Ip address from dhcp. Uing ubuntu 12.04, I could move the disk on my system to another system by removing /etc/udev/rules.d/70-persistnet-network (not sure about name). System would get new IP address from dhcp server. Now I have moved to ubuntu 14.04. I do not see that file is getting created all the time. Also if I move the disk to another system, the new system still tries to boot with old I [11:55] P address and it does not try to get new IP from dhcp server? thx === jibel_ is now known as jibel [12:43] Hi guys and gals! Anyone willing to help me with a port/firewall issue? I want to change my nginx configuration such that the server listens on a different port other than 80, say 82. But as soon as I do that it becomes unavailable, so the firewall thing is my first guess, but I don't seem to be able to fix it... :( [12:46] Not exactly knowing what I'm doing -- I didn't set the server up myself and I'm not very familiar with Ubuntu --, but I've added a rule in iptables (iptables -A net2fw -p tcp --dport 82 -j ACCEPT) and in ufw service (ufw allow 82) but that doesn't seem to help. When I try to access the server I get a timeout. [12:50] Could someone help me with an upgrade from 10.04 server? Here's the issue I'm having: http://pastebin.com/GHtY9mcm [12:54] Dave404: you need to go via 12.04. YOu can't upgrade from 10.04 to anything else. [12:54] rbasak: erk. Erm, how can I do that without losing my configuration? [12:55] Dave404: carefully? [12:55] Dave404: back up first. [12:55] Dave404: you will lose some configuration, because configuration necessarily changes with newer releases of various packages. [12:55] Dave404: you have to adjust everything. [12:55] I was mostly worried about having to configure my mail server again [12:56] I'm amazed I managed it the first time [12:56] You will most likely have to. [12:56] Stuff changes. Upstreams deprecate old configuration directives. Etc. [12:57] It's more a concern because when I installed mail server documentation apparently required a degree in compsci to be understood [12:57] :) [12:57] The best thing to do is record what you do to configure a server, and make that automatic. Then work on a new installation on a new release and update that configuration until it works. Then switch over. [12:57] rsyslog pisses me off so much for that :( [12:57] Apache and so forth are a doddle these days [12:57] You'll be upgrading from Apache 2.2 (at most) to 2.4. Various configuration directives have changed. [12:57] Apache I'm not worried about [12:57] Apache is fairly user friendly [12:57] Well, you see my point. [12:57] what mailserver do you use? [12:58] I'm honestly not sure [12:58] cause nothing has really changed in postfix [12:58] and only alittle in dovecot [12:58] I recall Dovecot being one component [12:58] atleast those programs are very good about backwards config compatability [12:58] I remember it not being clear what the different parts were for only that they needed to be configured [12:58] well the issue with email is, it's a huge stack [12:59] lots of software doing stuff [12:59] I was just surprised that we (that is to say geeks worldwide) hadn't manage to make it a little easier to install by now [13:00] One can have an instance of Apache up and running in minutes [13:00] On the plus side the documentation looks to have improved [13:00] Sorry to be a negative nitwit [13:01] dave404, it's not suppost to be easy :) [13:01] if it was, every spammer would have one :) [13:01] and that is what everyone running a mail system is attempting to avoid [13:01] That seems like throwing the baby out with the bathwater [13:01] why? [13:02] if you really want your emails to be read [13:02] you will follow what the receiver requires, or pay someone to do it for you [13:02] I don't think I follow but it sounds confrontational [13:03] email is :) [13:03] due to spam/viruses, no server will accept the fact you are who you say you are, till you prove yourself [13:03] and doing that proof requires a lot of work [13:03] I have some plans to make all of this better in a fairly revolutionary way, but I haven't had any time to work on it recently :-( [13:04] rbasak, including dns, spf, dkim, dmarc? [13:04] I'm not sure how that affects setup to the point of making it so difficult. That sounds like the sort of thing that affects the underlying mechanics rather than the end-user's setting up of it [13:04] patdk-wk: yep! [13:05] end user? the end user is simple [13:05] the issue with end users, is email programs are still written from the 80's [13:05] I recall the documentation alternating between aggravatingly vague and impenetrably deep [13:05] why do they default to port 25? 25 was done away with in 2005 [13:06] So it would say "You need these packages. Oh and this one." without explaining what the purpose of any of them was in an approachable way [13:07] Then further down it would assume one understood how the parts fit together and the explanation could skip to properly in-depth things [13:07] From what I can see at least some of that has been remedied [13:08] Okay, so PostFix exchanges email between the server and the outside world. Cool. [13:08] Then Dovecot makes it accessible remotely - serving up email via POP3 or IMAP [13:08] Sorry, rubber ducking [13:11] :) [13:11] oh, I think I know what you want [13:11] a generic all in one, smtp server tutorial [13:11] I'm not much of a fan of those, probably why I'm coming off that way [13:16] Something like that would be nice, yes. I find it difficult to get into things properly if I'm just stumbling in the dark. I need a general grasp of the big picture before I can appreciate where the details fit in :( [13:20] well, what I would agree more with, and I have seen a few but kindof rare [13:20] is more a guide to what is needed [13:20] and pointers to documentation for them [13:20] like, you will need postfix, dovecot, likely too amavisd-new, clamav, spamassassin [13:20] and optionally, policyd, postsrs, ... [13:22] It might also be worth noting it's a tiny mail server in terms of users. There's like 3 accounts [13:22] but like most all tutorials, you get, this is what I did, you have no clue if I did it correctly, or if I documented this correctly, or what any of this means, just follow it :) [13:22] the amount of people doesn't matter [13:22] it comes down to, how successful you want your emails to get out [13:22] how much spam you don't want to get in [13:22] In all honesty I've never received any spam on these accounts [13:22] and how much protection you want from if an account gets compromised or the website is hacked [13:23] In the entire time I've run them [13:23] then you must not submit those email accounts to many likes [13:23] the worst amount of spam I got, was after I had to sign up for microsoft partner :( [13:24] They're not signed up to anything beyond PayPal, I think [13:24] second worse, and this involed phone calls also, was hitting the inc500 [13:24] now, my wife gets all kinds of crud [13:24] but she puts her email in everything [13:25] and signes up for every single *offer* [13:26] I use my personal gmail account for stuff like that and let them handle spam [13:26] so atleast the spam part you don't have to worry about then [13:26] postsrs you only need to worry about if you relay/forward emails [13:27] policyd protects you from a compromised account from sending too much spam, or from your website doing it [13:27] It's a low traffic mail server mostly for handling some super-simple business stuff [13:27] the first time someone gets a virus, and a spammer gets the account password, it will no longer be low volume :) [13:28] No Windows users :) [13:28] not sure what that has to do with it [13:29] I'm not sure how to respond to that [13:29] your claiming, if you run windows, there is no way your password can be compromised? [13:30] don't run [13:30] I didn't claim that. [13:30] ok, not sure what, no windows users, has to do with anything then [13:30] You mentioned viruses. [13:30] yes [13:31] those exist on macos, linux, ... [13:31] Not exactly a common problem though compared to the way they are in the world of Windows [13:31] no? [13:31] every single java issue is also a linux issue, unless your not using oracle java, then it might not have that security issue :) [13:32] every flash issue is also an issue, likely more so now, without anymore updates [13:32] if you never run java/flash/javascript your likely safer [13:33] In all honesty I'm more concerned about issues with WordPress than I am about viruses on my desktop [13:33] wordpress has endless issues :) [13:33] you can limit it's issues somewhat by using apparmor [13:34] Yes, I know my *nix desktops aren't impenetrable fortresses but they also don't roll over at the first auto-downloading .exe file [13:36] So I'm not super concerned about someone pinching my password. I'm more concerned about issues somewhere else in the system allowing access without said password [13:37] Hi all! I'm sorry, I posted this before (half an hour ago) from a webchat but was disconnected (and didn't receive an answer afaikt). [13:37] Can someone help me with configuring a Ubuntu 14.04 server such that nginx listens on port 82 instead of 80? [13:38] Here's the thing: Nginx works when set to 80 but requests time out when I change the port to 82. I added a iptables rule [iptables -A net2fw -p tcp --dport 82 -j ACCEPT] and also [ufw allow 82] but I'm probably missing something... [13:39] strange [13:39] ufw? and shorewall? [13:39] and you didn't use shorewall to do it? [13:39] are you sure nginx is listening on port 82? [13:40] you verified via netstat? [13:40] patdk-wk: viruses exist on virtually all platforms, but given the volume of such things on windows compared to anything else, you can say viruses virtually do not exist on anything but windows [13:40] royk, depends [13:40] sudo netstat -nap |grep 'nginx' [13:40] tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN 25639/nginx [13:40] tcp 0 0 0.0.0.0:82 0.0.0.0:* LISTEN 25639/nginx [13:41] I find a virus every single day on linux [13:41] I only find them about once a year on windows [13:41] but then my windows machines have no internet access [13:41] but the linux ones do [13:41] then it's not remotely comparable [13:41] I'm not familiar with the server (didn't set it up myself). How do I check shorewall? [13:42] is there an /etc/shorewall folder? [13:42] oh, yes. [13:42] edit the rules file in there [13:42] ACCEPT net $FW tcp 82 [13:42] if you want net only [13:43] depends on what all else is connected to that machine [13:43] then do a, shorewall restart [13:44] No way!! [13:44] it works [13:44] Thank you so much patdk-wk! [13:44] I knew it was something trivial. [13:45] well, if you used iptables correctly, that would have worked (for awhile) [13:45] but adding an accept rule, AFTER the return rule, is not too useful [13:47] heh, funny email [13:47] Remember before you call "Reboot" - "Reboot" - "Reboot" [13:48] Thanks for pointing that out! Makes sense. I'm still just a bloody noob after all... :) [13:51] So using ufw (which is enabled on that server apparently), doesn't automatically take care of shorewall, is that right? [13:52] those are two totally different firewall tools [13:52] and how do you know ufw is enabled? and not just *installed* [13:53] ufw status_ [13:53] ? [13:53] !ufw | RoyK [13:54] RoyK: Ubuntu, like any other Linux distribution, has built-in firewall capabilities. The firewall is managed using the 'ufw' command - see https://help.ubuntu.com/community/UFW | GUI frontends such as Gufw also exist. | An alternative to ufw is the 'iptables' command - See https://help.ubuntu.com/community/IptablesHowTo [13:54] Chris_hubu: I know... [13:54] eh [13:55] cfhowlett: I know [13:55] heh [13:55] cfhowlett: I was just trying to suggest using ufw status to check if it was enabled :P [13:56] RoyK sorry, I came in mid-thread. proceed. [14:18] zul, coreycb, matsubara, lutostag, gnuoy, gaughen, kickinz1, beisner, rharper: o/ [14:18] Again, for my own future reference, here's the code I used: grep -rl "gupfiihzra" . | xargs sed -i 's/<\?php \$gupfiihzra.*\$cmhvuwnohn-1\; \?\>//g' [14:18] o/ [14:18] o/ [14:18] rbasak: \o [14:18] \o [14:18] Everyone else: my team are going to do some virtual sprinting for the next couple of hours to get a bunch of server packages merged and sponsored [14:18] \o [14:18] o/ [14:19] o/ [14:19] helo [14:19] Feel free to watch, or join in, etc. There are sponsors here to try and help and get packages and other server fixes landed. [14:20] http://reqorts.qa.ubuntu.com/reports/ubuntu-server/merges.html is a list of ~ubuntu-server subscribed packages that may need a merge [14:20] https://merges.ubuntu.com/ has the full list [14:21] https://wiki.ubuntu.com/UbuntuDevelopment/Merging is a helpful guide [14:22] Does everyone want to pick a package that looks interesting to you, and see if you can either update it, or find out why that would be inappropriate? [14:22] * kickinz1 reading [14:22] Let us know what you're looking at to save duplicate work [14:23] rbasak: I'll try vgabios [14:24] rharper: sounds good! [14:24] grab-merge says it's maintained in revision control -- need to investigate [14:25] It may be telling you about for Debian, rather than Ubuntu. [14:25] ah [14:25] rbasak, if it shows one of our team members names for last uploader, should we ask first? [14:25] some of these look very openstacky [14:26] gaughen: yes. The general locking mechanism to avoid duplicate work is to check with the last uploader. [14:26] * lutostag takes a look at autofs [14:26] sorry for the delay: service ufw status says "running" -- that's why I assumed it was active. [14:27] lutostag: great, thanks! [14:27] nagios-nrpe for me! [14:27] I don't particularly have a plan for how this session might work btw, we'll just see how it goes I guess. [14:28] I'm happy to go through decisions about each package merge if you want to have that discussion here. Or if you know what you're doing, just go ahead and prepare the merge. [14:28] gnuoy: thank you! [14:29] okay, rbasak I pick numactl [14:29] * kickinz1 takes alsa-utils [14:29] rbasak, I have no idea what I'm doing. [14:29] rbasak: i'll give samba merge a spin. [14:29] (why is alsa-utils, in server list ?) [14:30] beisner: great! Note that I know I can't upload samba, so we may need zul to sponsor that. [14:30] gaughen: no problem - I'll go through it here. That'd be a good example of the process I think. [14:30] rbasak: *cough* core-dev *cough* [14:30] zul, hahahahahah! [14:30] beisner: also I happen to know that samba FTBFS in Utopic right now. I discovered that yesterday. Maybe it's fixed in Debian now though. [14:30] rbasak, do you have a blog post yet? ;) [14:30] (I haven't had time to look yet!) [14:31] beisner, he needs a wiki page! [14:31] kickinz1: that's a good question. [14:31] beisner, with info about him like that he likes american football, etc, etc [14:31] kickinz1: http://people.canonical.com/~ubuntu-archive/germinate-output/ubuntu.utopic/ is a good place to look [14:31] rbasak, I take another one? [14:31] rbasak: also things like keystoneclient, heatclient, etc should probably be blacklisted [14:32] kickinz1: sure. It's certainly a low priority for server anyway. [14:32] zul: yes - I have a plan to implement some sort of blacklisting for the merge report, but there are a bunch of things above it on my TODO list :( [14:32] rbasak: i so dont know what that is like [14:33] gaughen: OK so let's look at numactl [14:33] okay [14:34] ok I'll look at keepalived [14:34] I see that it was last uploaded by dannf, sponsored by seb128, from that report, and it looks like Debian have moved from 2.0.9~rc5-1 to 2.0.9-1 since. [14:34] kickinz1: great! [14:35] gaughen: so the first place that I go to next is http://packages.qa.debian.org/numactl. This is the Debian PTS, which tells me all about this package in Debian. [14:35] Another important page is https://launchpad.net/ubuntu/+source/numactl, which is the equivalent Ubuntu page. [14:36] From the Debian PTS page, I follow a link to the changelog, on the right under links [14:36] Here I can see what Debian have change since Ubuntu diverged. [14:36] let me find the versions again [14:36] I better write them down [14:37] I'm interested in changelog entries that follow 2.0.9~rc5-1, since the 2.0.9~rc5-1ubuntu2 in the Ubuntu version in Utopic tells me that that's the point where Ubuntu diverged. [14:37] I'll take python-flake8 [14:37] matsubara: great - thank you! [14:37] grab-merge is dangerous! [14:37] rbasak, oh cool that doesn't look too bad [14:37] just one changelog entry [14:37] gaughen: right. Just one change. [14:38] (well, one Debian upload, with three distinct changes logged) [14:38] gaughen: next, I want to know what Ubuntu has changed, so that we can assess what we need to upload to update Ubuntu [14:39] rbasak: so, the file with conflicts was debian control, so looks like a simple change in the Maintainers list [14:39] So to do that I click on 2.0.9~rc5-1ubuntu2 under the Utopic section in https://launchpad.net/ubuntu/+source/numactl [14:39] rbasak, what does "* Upgrade standards to 3.9.5 (no change)" mean? [14:40] rbasak: should I update the XSBC-Original-Maintainer value to match the Debian Maintainer value ? Debian QEMU Team vs Debian QA Group ? [14:41] rharper: that's from grab-merge for vgabios, right? What's the rest of the previous Ubuntu delta look like - is it still applicable now? [14:41] gaughen: debian/control must define the Debian policy version that the package complies to. When Debian updates policy, then the package declares compliance with an old policy. [14:41] aaah okay [14:41] rbasak: yes -- I think so http://paste.ubuntu.com/7814707/ [14:42] gaughen: so the Debian maintainer must check the policy changelog, make sure the package is compliaant with latest policy, and bump the policy version the package declares compliance with. [14:42] gaughen: in this case, he's done that, and found that he's not needed to make any changes. So he's just changed the declaration in debian/control, and logged the change like that. [14:43] cool [14:43] gaughen: in Ubuntu, we don't usually want to introduce a delta just to meet Debian policy, as that's just tedious to maintain. So we leave it to Debian (or file bugs and patches with Debian) [14:44] gaughen: the exception is for packages which aren't derived from the Debian package (eg. Juju). There, we originate the package so we update in the same way to bring packages up to compliance. [14:44] rharper: yes - XSBC-Original-Maintainer should match the Maintainer field from the Debian package our diverged package is based on [14:45] rharper: is that the entire diff? [14:45] rbasak: thats the only change , yes [14:45] the REPORT file shows only debian/control [14:45] as a file with conflicts [14:45] * rbasak takes a look [14:46] lol, grabmerging samba advises: 'It looks like this package is maintained in revision control ... You almost certainly don't want to continue without investigating.' [14:47] ha! [14:48] * beisner give samba back to the floor. [14:48] * beisner gives samba back to the floor. [14:49] rbasak, I'm in a similar situation as rharper, there's a conflict in debian/control in the maintainer field: http://paste.ubuntu.com/7814740/ [14:50] rharper, matsubara: so there's a tool called update-maintainer [14:51] The background is https://wiki.ubuntu.com/DebianMaintainerField [14:51] This is why, if in Ubuntu we change the package (ie. diverge from Debian), we rename the field. [14:52] What the tool does is trivial, so you can fix it up by hand. But matsubara's conflict looks a bit confusing to me because the original fields were multiline. It may be easiest to select the Debian side of the conflict, and then to just run update-maintainer to fix it for Ubuntu again. [14:53] rbasak: so, once I've updated the file with conflicts, what's next? [14:55] rharper: looking [14:56] rharper: so you've fixed the conflict manually, but you still have something that merge-o-matic thinks might be right. But this assumes that it's fine to merge Ubuntu changes with Debian changes. [14:56] rharper: it may be the case that the changes apply but don't make logical sense, or break the package. [14:56] ok [14:56] rbasak, Cool, I left the debian part and ran update-maintainer and now I got this debian/control file: http://paste.ubuntu.com/7814762/ and am ready for the next step [14:57] rharper: debian/source/format says "3.0 (quilt)", so this is a modern best-practice package that uses the quilt system. [14:57] rbasak: I think they make sense; the ubuntu changes add new screen modes; the upstream changes in debian are unreleated, so I think the ubuntu change still make sense [14:57] rharper: so the next thing I did was to run "quilt push -a", and this failed. Do you get the same thing? [14:57] rbasak: where are you running that? in the 5ubuntu1 dir ? [14:58] rharper: I also followed the bug link in the Ubuntu delta, which linked to a Debian bug, which is apparently fixed in Debian. So I think the changes maybe do conflict. [14:58] rharper: yes - in vgabios/vgabios-0.7a-5ubuntu1/ [14:58] ah [14:58] % quilt push -a [14:58] No series file found [14:58] rharper: ah, sorry. [14:58] rharper: http://wiki.debian.org/UsingQuilt [14:59] export QUILT_PATCHES=debian/patches [14:59] export QUILT_REFRESH_ARGS="-p ab --no-timestamps --no-index" [14:59] ah, right [14:59] rharper: those two lines should set you up quickly for hnow [14:59] new box doesn't have those settings [14:59] yeah, I get failed push now [15:00] rharper, matsubara: so the general task here is to examine the previous Ubuntu delta carefully, the new Debian changes carefully, and figure out how to apply the Ubuntu delta on top of latest Debian package. [15:00] In git terms, this is a rebase. [15:00] right [15:00] Debian may have picked up changes that Ubuntu made, for example, but in a different way. [15:01] merge-o-matic tries to do it automatically, but this isn't necessarily safe. You still have to check that it makes logical sense. [15:02] When you're done, you can build a source package (I use "debuild -us -uc -nc -S"). Then examine diffs using debdiff against the old and new .dsc files. [15:03] rbasak, ah ok, even though the REPORT says the only conflict is in debian/control, whatever the tool did merge automatically and didn't flag as a conflict needs some eyeballing just to make sure? [15:03] A sponsor usually wants to see the diff between the new debian package and the new proposed ubuntu package, and the diff between the old ubuntu package and the new ubuntu package. I usually want to see the diff between the old debian package and the old ubuntu package as well. [15:03] matsubara: right - you go tit [15:03] you got it [15:05] I find merge-o-matic useful only for trivial merges, because I want to check every aspect of the old Ubuntu diff manually to make sure that whatever the reason it was there before will still work in the new upload. [15:05] I know other Ubuntu devs who are happy to examine the whole diff that way though. [15:06] * hallyn_ does like to look at full debdiffs [15:06] So what you do exactly is up to you - provided that you're confident at the end of your process that the whole previous Ubuntu delta is accounted for in some way [15:06] rbasak: in this case, it appears that some of the ubuntu patches have already been applied; should they be dropped if all of the contents of the patch are included in the newer package ? [15:06] yup, that's the idea case [15:06] right [15:06] then note that in the changelog [15:07] s/idea/ideal/ [15:07] ok and what's the quilt way to drop the patch? [15:07] If all changes can be dropped, then we can sync the package directly from Debian and drop our delta. That's the most ideal. [15:07] rbasak: right [15:07]